The historical background of FMD vaccine production, the Valléé-Schmidt-Waldmann concept, the production of antigen in bovine tongue epithelium (Frenkel culture) and improvements achieved through the use of monolayer and suspension cultures of cell lines are described. Key elements of modern vaccine production are discussed such as the production environment, bio-safety concepts, rules for good manufacturing practice, requirements for culture medium, cells, virus strains, inactivation of viral antigen, and successive concentration and purification of the antigen. Storage of the concentrated antigen at ultra-low temperatures creates greater flexibility for the producer. In addition, it allows national and international organisations, the opportunity to establish vaccine banks for emergency vaccination. For the latter purpose it is even more important that antigens are purified. The purification of FMD viral antigens – including the removal of nonstructural proteins (NSP) – provides a system that can distinguish between the immune responses of vaccinated animals from those of animals infected with live FMD virus. Consequently, the combined use of purified vaccine and anti-NSP tests essentially provides a "marker" system. Possibilities for the world-wide control and eradication of FMD by vaccination, including the institution of regional vaccine banks, are discussed. Modern (qualified) FMD vaccines perform very well both for regular vaccination programs and for the control of outbreaks, and over the past 15 years there are no well-documented cases where cattle vaccinated with an approved vaccine have caused new outbreaks. Therefore, whether there continues to be justification for sanctions on trade, when controlling single (limited) outbreaks by (ring-) vaccination.

(51 int. Cl. .......... C12P21/06; C12N 11/06; G01N 33/549; CO7K 17/06 52 U.S. Cl. ........................ 435/68.1; 435/41; 435/91.1; 435/101; 435/180; 435/181; 435/240.23; 435/240.24; 436/531; 436/532; 530/412; 530/415; 530/417; 530/815; 530/816; 530/402; 525/383 (58) Field of Search ................................... 435/815, 174, 435/177, 176, 180, 181, 182, 41, 68.1, 91.1. 101, 240.23, 240.24; 436/531, 532; 530/412,415, 417, 815, 816, 402; 525/383

For FMD vaccine production, inactivation of the FMD virus is the most critical step. Formerly, from 1940 onwards, the virus was inactivated with formaldehyde. This inactivation was relatively slow, about 0.2 - 0.3 log 10 per hour. Because formaldehyde not only reacts with the virus produced but with many other components in the medium, such as proteins and amino acids, its concentration can become rate-limiting and inactivation plots may show tailing-off, resulting in residual infectivity. Many of the bad stories of post-vaccination outbreaks date back to the use of formaldehyde-inactivated vaccines (e.g. the outbreaks in France in 1981 and in Eastern Germany causing the Danish outbreak in 1982). Much faster and safer inactivation was obtained with aziridines and in the 1980s binary ethyleneimine (BEI) was introduced in practically all vaccine production laboratories. If inactivation plots are made of every production batch, as is now required by the European Pharmacopoeia, and these plots show proper inactivation rates, vaccines can considered to be completely safe. Under optimal conditions, inactivation rates are in the range of 0.5 - 1.0 log 10 per hour. In general, the inactivation takes 40-48 hours,which will guarantee complete inactivation of all virus particles in a batch. Since formaldehyde (FA), the 'classical' inactivating agent, inactivates at a rate of 0.3 logs per hour only, a significant contribution of FA to the inactivation of BEI can hardly be expected. However, here it is shown that FA added during the BEI-inactivation process strongly augments inactivation rates with a hundred to thousand-times (to 2.5-3.5 logs per hour). This will enable inactivation during a working day or just overnight with even higher safety levels of the vaccines. Also, it is known that formaldehyde cross-links viral proteins which will stabilise the antigen. The short inactivation times will limit proteolytic destruction of 146 S antigen and increase antigen yields. It is expected that by the cross-linking activity of FA the stability of the antigen (and of vaccines) and the endurance of the immune response will be favourably influenced.

The historical background of foot and mouth disease (FMD) vaccine production is briefly described. Improvements achieved through the use of monolayer and suspension cultures are outlined. Elements that are crucial in the production of modern vaccines are discussed, such as inactivation of viral antigen, successive concentration and purification of the antigen and the final formulation of the vaccine. Storage of concentrated antigen at ultra-low temperatures creates greater flexibility for the producer and has also enabled national and international organisations to establish vaccine banks. The purification of FMD viral antigens, including the removal of non-structural proteins (NSPs), enables the immune responses of vaccinated animals to be distinguished from the responses of animals infected with live FMD virus. Consequently, the combined use of purified vaccine and tests for the detection of antibodies against NSPs essentially provides a marker system to distinguish between vaccinated animals that subsequently become infected and those that have not. Bearing in mind the good record of modern vaccines in the control of outbreaks and the possibility of screening vaccinated herds for carriers, the author proposes that the OIE reconsider the differences between the requirements for regaining export status following the use of stamping-out as opposed to vaccination in outbreak situations.

The current status of foot-and-mouth disease (FMD) vaccine production is reviewed. The production of antigen in bovine tongue epithelium (Frenkel culture) is described and improvements in monolayer and suspension cultures of cell lines are outlined. Inactivation of viral antigen and safety tests are discussed. A 'minimum safety level' is recommended: at the end of the inactivation process, antigen batches of any size should contain less than one virus particle. After inactivation the antigen can be formulated into a vaccine or purified and concentrated for storage at ultra-low temperatures in a vaccine bank. Vaccines prepared with the adjuvants Al(OH)3 and saponin are compared with (double) oil emulsion vaccines. Because oil vaccines can protect both cattle and pigs and induce long-term protection, they are most suitable for use in ring vaccinations. A new generation of vaccines, based on constructed modified-live viruses or (bio-) synthetic peptides, is briefly reviewed.

The Chinese strain of hog cholera virus (HCV) was adapted to suspension cultures of the established swine kidney cell line SK6. The strain designated "Cedipest", is produced on the basis of a seedlot system. The masterseed virus was identified in vitro and in vivo, and was found free from extraneous pig pathogenic viruses by repeated animal inoculation followed by appropriate serological tests. A distinct and reproducible relationship was ascertained between infectivity in vitro and protection. Pigs inoculated with 400-600 TCID50 of the Cedipest strain proved fully protected against challenge with greater than 100 pig LD50 of a virulent strain of HCV at 7 days and at 6 month post vaccination.

A set of monoclonal antibodies was used to isolate nonneutralizable foot-and-mouth disease virus variants, and the RNAs of the variants were sequenced. Cross-neutralization studies and mapping of the amino acid changes indicated two major antigenic sites. The first site was trypsin sensitive and included the VP1 140 to 160 sequence. The second site was trypsin insensitive and included mainly VP3 residues. Two minor sites were located near VP1 169 and on the C terminus of VP1. Comparison with poliovirus type 1 and human rhinovirus 14 showed a similarity in the immunogenicity of comparable sites on the viruses.

Using polyclonal sera raised against foot-and-mouth disease virus in susceptible animals, evidence was obtained for the existence of at least one further important antigenic site in addition to the neutralizing site on VP1 140-160.

Summary The paper reviews developments in foot‐and‐mouth disease research with an emphasis on the Netherlands and research developments at the Central Veterinary Institute.

Variants of type A10 FMDV were isolated by passage of virus in BHK-cells in the presence of a neutralizing anti-peptide serum or monoclonal antibodies. These variants which were no longer neutralized by the particular anti-peptide serum or monoclonal antibody were easily obtained from (crude) virus populations ("cattle" virus and BHK-adapted virus). The rapidity of isolation (in two or three passages) suggested that these variants are already present in normal virus populations. All (plaque purified) variants isolated so far seem to be different: A panel of 20 monoclonal antibodies and an anti-peptide serum showed different neutralization patterns for all isolates and parent virus. Electrofocusing patterns of many variants were found to be different showing changed charges for VP2 as well as for VP1. Thus in FMDV both VP1 and VP2 are probably involved in antigenic sites. Normally the variants escape our attention because in neutralization assays only a limited quantity of infective units are used, representing only the "top of the iceberg". In classical inactivated virus vaccines many of these variants will be represented and therefore can be expected to be "primed" immunogenetically. This will not be the case with peptide vaccines or in case of recombinant DNA products, where only one virus clone is represented. In addition, and probably more important, inactivated virus vaccines will raise antibodies against completely independent epitopes that each have a limited chance to be changed in the variants present in the challenge virus population. Thus if peptide vaccines can be composed in such a way that antibodies are raised against completely different antigenic sites the chance of break through of variants will be strongly limited and it is expected that the efficient protection of inactivated virus vaccines can be approached.

In a previous study we challenged the generally accepted opinion that inactivation of FMDV by formaldehyde (FA) is an (unsafe) non-linear process. Our data showed that under proper conditions inactivation will be linear without "tailing-off". For more than forty years two other fixed beliefs existed with respect to FMD vaccine preparation: Virus must first be adsorbed to Al(OH)3-gel before being inactivated. Concentrations of formaldehyde are critical and must be within a very narrow range. Until recently the prescription of adsorption of the virus prior to inactivation made proper control of inactivation kinetics impossible. However, by eluting the virus from the gel by caesium chloride density centrifugation, the kinetics of inactivation can be studied. For adsorbed and non-adsorbed virus, identical inactivation curves were found. Antigenicity was found to be of identical "quality" for adsorbed and non-adsorbed virus. The second dogma was challenged by inactivating non-adsorbed virus with FA-concentrations of up to 6 times the one originally prescribed. In parallel inactivation was performed with acetyl-ethylene-immine (AEI). The antigen was purified after inactivation. Antigen yields after purification were systematically lower at high FA-concentrations, however, antigenicity seemed only slightly changed by the treatments: In an immunosorbent (ELISA) assay using a panel of 22 monoclonal antibodies (McAb) only few McAb's reacted differently with FA-treated antigen if compared with AEI-treated or fresh (A10) FMDV. The FA-treatment induced decreases, as well as increases in reactivities. The AEI-treated virus reacted almost identically to non-treated 146S particles.(ABSTRACT TRUNCATED AT 250 WORDS)

Both whole virus particles and isolated VP1 of foot-and-mouth disease virus type O1 induce neutralizing antibodies. Results obtained with pigs vaccinated with either isolated VP1 or intact particles and subsequently challenged show that neutralizing activity induced by intact virus correlates well with protection in pigs, whereas neutralizing activity induced by isolated VP1 confers little or no protection. Further evidence suggests that the epitope responsible for the induction of neutralizing antibodies by VP1 is located at the C-terminal end of the protein between residues 200 and 210.

All overlapping hexapeptides of the outer structural protein VP1 of type O1, type A10, and type C1 were reacted with the appropriate anti-virus, anti-viral subunit and anti-VP1 sera. The results suggest that anti-virus sera may contain activities against viral subunit and VP1 as well as against virus. Furthermore the antigenic peptides associated with the intact virion of all three serotypes are found at similar locations on their respective VP1s, and produced neutralizing activities when used for vaccination. The results further offer an explanation for the often observed cross-reactions between serotypes, especially at the level of the viral subunit and VP1. The reliability of predictions of useful peptides from hydrophilicity profiles and secondary structure predictions is questioned. Predictions based on variation profiles appear to be more useful.

Antisera were raised against the chemically synthesized peptide corresponding to each epitope of three foot-and-mouth disease virus strains. Peptide synthesis was further used to determine which amino acid residues in each epitope are important for the specificity of antisera raised against the whole virus. The specificity of the antibody paratope for its epitope was shown to depend on structure as well as sequence. Anti-virus sera demonstrated a greater specificity for the homologous peptide than did the anti-peptide sera. Two of the three peptides were able to induce neutralizing antibodies against the homologous virus. The specificities of the antibodies present in the anti-peptide sera were also inferred from the reactions of each with related sets of peptides. The cross-reactions observed for the anti-peptide sera were readily explained in terms of the antibody specificities determined to be present. The findings also suggest that the diversity of antibodies raised against small peptides is limited and is determined by the immune system. A similar limited response to the native protein was observed, which may account for the high frequency with which anti-peptide sera react with the native homologous protein.

The inactivation of foot-and-mouth disease virus by formaldehyde was studied under different conditions, both as free virus and (as in routine vaccine production) after adsorption of the virus to aluminium hydroxide gel (alhydrogel). In the latter case infectivity was monitored after elution of the virus from the gel by isopycnic ultracentrifugation of the virusalhydrogel mixture in CsCl. By this method good virus recoveries were obtained. Adsorption of the virus to alhydrogel (without formaldehyde) did not reduce infectivity significantly. Both adsorbed and non-absorbed virus lost infectivity at a rate of about one log10 per day (at pH 8.5, 25° C—no formaldehyde).