Vaccine 19 (2001) 2107 2117 www.elsevier.

com/locate/vaccine

The suitability of the emergency foot-and-mouth disease antigens held by the International Vaccine Bank within a global context
P.V. Barnett *, A.R. Samuel, R.J. Statham
Institute for Animal Health, Pirbright Laboratory, Ash Road, Pirbright, Woking, Surrey GU24 0NF, UK Received 17 May 2000; received in revised form 28 September 2000; accepted 13 October 2000

Abstract The International Vaccine Bank (IVB) based at the Institute of Animal Health (IAH) in Pirbright, United Kingdom (UK), routinely monitors the suitability of the currently held strains of foot-and-mouth disease (FMD) vaccine virus, in anticipation that vaccine may be required to control FMD outbreaks that pose a threat to member countries. Using primarily the two-dimensional micro-neutralisation test (VNT), bovine polyclonal sera raised against each of the seven current emergency antigens were utilised to measure the relationship of IVB stocks to selected eld isolates. The O serotypes, Manisa and Lausanne, exhibited adequate levels of cross-protection against most of the type O eld isolates examined. A22 Iraq 24/64 showed the broadest spectrum of reactivity against the type A eld isolates examined and was supplemented by A15 Thailand 1/60. Some type Asia1 eld isolates, particularly those from South East Asia, showed antigenic difference to the Asia1 India 8/79 vaccine strain by VNT, but in-vivo testing in the guinea pig model indicated this to be insignicant. The only C serotype representative, C1 Oberbayern, may be one of the least antigenically diverse of the current portfolio of bank antigens. Comparison of the serological and sequence data shows that despite signicant genetic variation between the eld isolates examined the antigens held by the IVB should still prove efcacious in the eld. 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Foot-and-mouth disease viruses; Vaccine; Antigenic relationships

1. Introduction Foot-and-mouth disease virus (FMDV) is exotic to the member countries of the IVB, but the sporadic occurrence of disease in Eastern Europe and in the countries of trading partners poses a continuing threat. Traditionally, control in the UK relied upon surveillance, rapid diagnosis, slaughter and movement restrictions. Future outbreaks may require a different approach, including the use of emergency vaccination, due, in part, to the problem of disposing of large numbers of slaughtered animals, together with the associated welfare and environmental issues. To meet this possible eventuality, the International Vaccine Bank (IVB), established in 1985, maintains a bank of concentrated FMDV antigens stored over liquid nitrogen. These can be rapidly formulated into vaccine as re-

* Corresponding author. Tel.: + 44-1483-232441; fax: + 44-1483232448.

quired by either the UK authority (The Ministry of Agriculture, Fisheries and Food (MAFF)), or any of the other IVB member countries (Australia, New Zealand, Ireland, Norway, Sweden, Finland and the associate member, Malta). The facility operates according to the principles of Good Manufacturing Practice (GMP) and currently holds licences from the UK Department of Health and the UK Veterinary Medicines Directorate (VMD), for the production of FMD vaccines. Presently the IVB portfolio consists of 0.5 million doses of each of seven strains of highly potent inactivated FMD antigens, specically O1 Lausanne, O1 Manisa, A22 Iraq, A24 Cruzeiro, A15 Thailand, C1 Oberbayern and Asia1 India, each strain having been selected because of its wide antigenic spectrum. However, the ever-changing global epidemiology requires the continued monitoring of contemporary viruses. Should an outbreak threaten any one of the member countries, it would therefore be possible to advise concerning the appropriateness of the currently held antigens or the need to acquire further strain(s).

0264-410X/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 0 ) 0 0 3 9 9 - 6

2108

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

Little has changed in terms of the world distribution of foot-and-mouth disease since the mid 60s, and readers are referred to the recent detailed review by Kitching [1]. FMD is still endemic in many areas of the world including Africa, Asia, the Middle East, India, and South America. Prophylactic vaccination against FMD ceased in Europe during 1990 1991 to allow the implementation of a harmonised policy and with it greater movement of livestock and their products within the single market (Directive 90/423/EEC). Since that time sporadic outbreaks in countries in or bordering southern Europe have maintained the threat of large outbreaks particularly in the high density domestic livestock of mainland Europe and the consequential risk to the UK [3]. FMD remains endemic in Asiatic Turkey, and there have also been viruses of FMD serotype O in Bulgaria (1991, 1993 and 1996), Italy (1993), and Greece (1994 and 1996), and outbreaks due to serotype A in the Former Yugoslav Republic of Macedonia, and Albania (1996). Since the breakdown of the Soviet Union into Independent States, FMD has spread north from Iran, Afghanistan and Turkey particularly into the transCaucasian countries of Azerbaijan, Armenia and Georgia and into Kazakhstan and neighbouring countries constituting a signicant threat to the borders of Europe. One of the FMD viruses isolated from the outbreaks in Turkey, which had spread from neighbouring Iran, was genetically and antigenically distinct and prompted the development of a new vaccine from the strain A Iran 96. This vaccine was subsequently used in Thrace to create a buffer zone and protect south-east Europe. Genetically and antigenically distinct viruses do occasionally arise emphasising the importance of surveillance. Remarkably, three years later another antigenically new strain of type A was identied in Iran, termed A Iran 99, and this has also spread to Turkey, again threatening Europe. In eastern Asia, FMD was reintroduced into Malaysia and there have been outbreaks due to virus types O, A and Asia1; serotype O virus was introduced into Taiwan Province of China (POC) in 1997 which resulted in around 3.8 million pigs being slaughtered and more than 20 million doses of vaccine being used to control the disease. In 1999, FMD, due to type O, was declared in Algeria. The disease rapidly spread both east and west affecting Morocco and Tunisia. Although less extensive than the previous epidemic of 1989 92 [4], the number of outbreaks still reached 173 (160 in Algeria, 11 in Morocco and two in Tunisia) involving almost exclusively cattle [5]. Sequence studies and comparison to recent isolates from West Africa suggested that the virus was exotic to North Africa but related to strains in West Africa and that it entered Algeria via illegal movement of animals [5].

The latter part of 1999 saw the occurrence of outbreaks of Asia 1 FMDV in Malaysia and Turkey. Outbreaks caused by this serotype have been frequent in Malaysia during the last decade. However, the outbreak of Asia1 in Turkey is of particular concern to the IVB. To examine the antigenic relationships between isolates from FMDV epidemics, of particular concern to IVB member countries, and the currently stored vaccine strains we have primarily applied the well-established two-dimensional virus neutralisation assay [6]. The epidemiological and genetic relationships of these isolates have been examined by the analysis of nucleotide sequence data using phylogenetic techniques. The work described here summarises these studies and evaluates the relevance of the current portfolio of emergency vaccine strains.

2. Methods

2.1. Viruses
All the strains of FMDV were obtained from the OIE/FAO World Reference Laboratory for FMD and were grown on either baby hamster kidney (BHK-21) cells or swine kidney cells (IB-RS2). The virus stocks used for the micro-neutralisation tests were stored as claried tissue culture harvest material at 20C in 50% glycerol.

2.2. Bo6ine polyclonal serum

Reference sera were derived from IVB cattle potency tests using 21 day post-vaccination samples taken from individually immunized animals receiving a half bovine dose of antigen formulated as an aqueous aluminium hydroxide/saponin vaccine. For each antigen, a pool of sera from eight individual animals was used in the serological tests.

2.3. Oligonucleotide primers

Oligonucleotide primer pNK72 labelled with a Cy5 amidite uorescent dye for use with the ALFexpress automated sequencer was purchased from Pharmacia Biotech. The sequences of primers used in this study are detailed in Table 1.

2.4. RT -PCR
Reverse transcription-polymerase chain reaction (RT-PCR) was performed using the primer set L463F and NK61 essentially as described by Knowles and Samuel [7].

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

2109

2.5. Cycle sequencing

A fmol DNA sequencing kit (Promega, UK) which utilises the method described by Murray [8] was used according to the manufacturers protocol with the following amendments. Approximately eighty femto-moles of cDNA template was used in the reactions and 1.5 pmol of the Cy5 amidite labelled pNK72 primer. The reactions were heated to 94C for 2 min and then subjected to 30 cycles of the following programme on a thermal heating block (Hybaid Ominigene, Hybaid UK): 94C for 1 min, 55C for 1 min and 72C for 1.5 min. The reactions were terminated by adding 4 ml of Cy5 sequencing stop solution (Pharmacia Biotech. Sweden) and cooled to 4C. The reactions were heated to 95C for 3 min prior to loading on a ALF express DNA sequencer (Pharmacia Biotech, Sweden).

2.7. Neutralisation assay

A two-dimensional micro-neutralisation technique similar to that described by [6] was used. Briey, doubling dilutions of antibody were reacted with 0.5 log10 dilutions of virus for 1 h at room temperature, BHK-21 or IB-RS2 cells were added as indicators of residual infectivity and the test incubated at 37C for 3 days prior to xing and staining. Antibody titres were calculated from regression data as the log10 reciprocal antibody dilution required for 50% neutralisation of 100 tissue culture infective units of virus (log10 SN50/ 100 TCID50). The antigenic relationship of viruses based on their neutralisation by antibodies is given by the ratio: r = neutralisation antibody titre against the heterologous virus/neutralisation antibody titre against the homologous virus. The signicance of differences in the values of r obtained by the polyclonal antiserum was evaluated according to the criteria of [10]. Generally, serological relationships between two viruses in the range r = 0.3 1.0 are indicative of reasonable levels of cross protection whereas values less than 0.3 indicate very dissimilar strains and the need to acquire or develop a new vaccine strain.

2.6. Analysis of sequencing data

The software AM V3.01 (Pharmacia Biotech, Sweden) was used to process the data which was then exported as an ASCII text le onto an IBM compatible personal computer where alignments were made manually in Wordperfect 6.1 before analysis using the software EpiSeq v2.0 suite of computer programs written by N.J. Knowles (IAH, Pirbright). All pairwise comparisons were performed by giving each base substitution equal statistical weight (ambiguities were ignored). Phylogenetic trees were constructed using the computer programme NEIGHBOR and the unpaired group mean averaging (UPGMA) method. Dendrograms were plotted with the program DRAWGRAM. These programmes were from the PHYLIP 3.5c phylogeny package ([9]). The UPGMA method constructs a tree by successive (agglomerative) clustering using an average-linkage method of clustering. UPGMA assumes a clock but the branch lengths are not optimized by the least squares criterion. This makes the method very fast and thus able to handle large data sets.

2.8. In 6i6o analysis

The potency (PD50) of the IVB strain Asia India 8/79 against eld isolates Asia Thailand 4/95 and Asia Nepal 28/90 was determined in guinea pigs according to the procedure described by Barnett and Statham [11] and compared with that observed against the homologous strain.

3. Results

3.1. Serological relationships 3.1.1. O1 Manisa and O1 Lausanne 6ersus O type eld isolates According to the two-dimensional test neutralisation results, adequate levels of cross-protection should be attained by the high potency O1 Manisa vaccine (PD50 ] 112) against the eld isolates examined, including those from Greece in 1994, Philippines in 1995 and the more recent Taiwan POC and Vietnam outbreaks (Table 2). A comparable study with O1 Lausanne (Table 2) also indicates that adequate levels of cross-protection should be attained from this vaccine strain against most of the viruses examined. However, direct comparison of the serological results and r values, indicate that the Lausanne vaccine strain is unlikely to protect against certain isolates, such as O Philippines 11/94 (r = 0.23**) or O Algeria 2/99 (r = 0.16**).

Table 1 Oligonucleotide primers used for the RT-PCR molecular epidemiology sequencing studies Primer Virus Sense Gene Primer sequence (5%3%) GACATGTCCTCC TGCATCTG GAAGGGCCCAGG GTTGGACTC ACCTCCRACGGG TGGTACGC

NK61 NK72 L463F

Universal Universal O universal

Negative Negative Positive

P2B P2A L

2110

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

Table 2 Serological relationships (r ) between recent O FMDV eld isolates determined in 2D microneutralization tests using IVB bovine 21 day post vaccinal O1 Manisa (lot 0163) and O1 Lausanne (lot 6.0.142C) seraa Virus Bovine antisera O1 Manisa O1 Lausanne O1 Manisa O Bulgaria 1/93 O Cambodia 3/92 O Cambodia 1/98 O Greece 4/94 O Greece 38/94 O Hong Kong 1/92 O Hong Kong 4/95 O Hong Kong 5/95 O Italy 1/93 O Morocco 1/91 O Philippines 2/95 O Philippines 9/95 O Taiwan 10/97 O Thailand 1/92 O Turkey 19/91 O Vietnam 7/97 O Algeria 2/99 O Philippines 1/99 1.00 \1.00, 0.93 \1.00 \1.00 \1.00 0.16, 0.12 0.81 \1.00 0.85 \1.00 0.35 \1.00 0.66 0.95 O1 Lausanne 1.00 0.5 0.59 0.45 \1.00, 0.79 \1.00 0.21* 0.69 0.54 0.05** 0.89 0.28* 0.45 0.43 0.15 0.56 0.16 0.28

Turkish isolates such as Turkey 8/99 and 10/99 with r values of 0.11* and 0.17*, respectively (Table 3). These viruses, together with those from epidemics from the end of the last decade and earlier part of this decade, which also exhibited a dissimilarity to the vaccine strain (data not shown), warranted further investigation. Due to the historical high potency of the Asia1 India vaccine (PD50 = 61), it was assumed this strain would still be efcacious against these distantly related viruses. To examine this, guinea pig adapted viruses of two serologically dissimilar eld viruses, namely Nepal 28/90 and Thailand 4/95 were tested by challenge against the IVB Asia1 India vaccine strain (Table 4). As indicated, the vaccine strain held by the IVB afforded acceptable potency gures against these selected strains, suggesting the vaccine could be expected to provide immunity against these eld isolates and that a further Asia 1 strain was not required by the IVB.

Table 3 Serological relationships (r ) between recent Asia1 FMDV eld isolates determined in 2D microneutralization tests using IVB bovine 21 day post vaccinal Asia1 India 8/79 (735 lot L244) seraa Virus Bovine antisera Asia1 India 8/79 Asia1 India 8/79 Asia Cambodia 2/91 Asia India 5/89 Asia Israel 3/89 Asia Malaysia 11/92 Asia Nepal 28/90 Asia Saudi 13/92 Asia Thailand 10/94 Asia Thailand 4/95 Asia Malaysia 9/99 Asia Turkey 8/99 Asia Turkey 10/99 1.00 0.32, 0.27 0.34, 0.29 0.17**, 0.23 0.43, 0.26 0.16**, 0.31 0.63, 0.78 0.22, 0.15* 0.24, 0.13* 0.43 0.11* 0.17*

, indicative of reasonable levels of cross protection; , indicative of very dissimilar strains. Bold data represents a repeated series of tests. The (r ) values shown represent the neutralization titre against heterologous virus/neutralization titre against homologous virus. * r signicantly less than 1.0 (P = 0.05). ** r signicantly less than 1.0 (P = 0.01).
a

3.1.2. Asia1 India 6ersus Asia type eld isolates The IVB vaccine strain Asia1 India 8/79 may not induce protection against some of the Asia1 eld viruses, in particular the isolates from South East Asia, here represented by Thailand 10/94 and 4/95, r values of 0.15** and 0.13**, respectively, and the more recent

a
, indicative of reasonable levels of cross protection; , indicative of very dissimilar strains. Bold data represents repeated tests using new high titre pool of bovine polyclonal sera from small scale cattle potency test (18/8/97). The (r ) values shown represent the neutralization titre against heterologous virus/neutralization titre against homologous virus. * r signicantly less than 1.0 (P = 0.05). ** r signicantly less than 1.0 (P = 0.01)

P.V. Barnett et al. / Vaccine 19 (2001) 21072117 Table 4 Guinea pig potency test results following vaccination with Asia1 India 8/79 and challenge with heterologous Asia1 eld isolates Challenge viruses Vaccine dilution 1/1 Asia India 8/79 Asia Thailand 4/95 Asia Nepal 28/90

2111

Potency* 1/9 5/5 4/5 3/5 1/27 4/5 3/5 4 /5 1/81 3/5 3/5 3/5 Controls 0/2 0/2 0/2 72.5 37.5 46.7

1/3 5/5 4/5 5/5

5/5 5/5 5/5

Number of animals protected/ number of animals per group. * PD50 value as determined by the Karber method.

3.1.3. A15 Thailand, A22 Iraq and A24 Cruzeiro 6ersus A type eld isolates The A22 Iraq vaccine strain held by the IVB was shown serologically to have a wide antigenic spectrum when studied against a number of diverse eld isolates (Table 5). The only exceptions being the more recent isolates, Brazil 2/95, Malaysia 38/95 and Turkey 1/98 with r values of 0.15**, 0.13** and 0.08**, respectively. The serological cross-specicity indicated by the A15 Thailand strain generally appeared to meet the needs of the member countries (Table 5), particularly in respect to the isolates Brazil 2/95 and Malaysia 38/95, which were antigenically distant from A22 Iraq. The only major concern were the new variants of type A FMD viruses originating from Iran in 1996 which spread to Turkey in 1997 and 1998, and the more recent Turkey and Iran 1999 isolates. These were not antigenically covered well by the current IVB type A vaccine strains (micro-neutralization r values of 0.08*, 0.13* and 0.10* for A Turkey 1/98, and 0.13*, 0.21 and 0.25 for A Iran 22/99, against A22 Iraq, A24 Cruzeiro and A15 Thailand, respectively). Another A variant was also isolated in Malaysia and Thailand in 1997, however, samples proved difcult to grow in tissue culture and were therefore not analysed by the virus neutralisation assay. An alternative test, the liquid phase blocking ELISA [12] indicated that none of the IVB vaccine strains were likely to confer protection (data not shown) and that another commercially available vaccine strain, A Iran 6/94, may be more applicable. A more recent acquisition to the IVBs portfolio, A24 Cruzeiro, which replaced the rst batch of antigen of the same strain acquired by the Vaccine Bank in 1985, displayed a minimal relationship to all the eld isolates examined (Table 5). This antigen is currently considered to be one of the least important vaccine strains, and was not examined further, since the A22 Iraq and A15 Thailand strains appear to offer adequate cover. 3.1.4. C1 Oberbayern 6ersus C type eld isolates C1 Oberbayern may be one of the least antigenically diverse of all the current vaccine strains held by the

IVB. All four isolates examined by VNT appear to be serologically dissimilar to the vaccine strain (Table 6). However, the number studied reect the Cs received by the World Reference Laboratory in the last few years. These same eld isolates were examined by liquid phase blocking ELISA (data not shown) and, in particular, conrmed the dissimilarity of Bangladesh 1/92 and Bhutan 2/91. Therefore, like the Asia1 India study, there is a need to examine the efcacy of this vaccine strain and three of these isolates namely, Bangladesh 1/92, Nepal 1/94 and Philippines 4/90 will undergo in-vivo protection studies.

3.2. Sequence analysis 3.2.1. O1 serotypes Compared to O1 Manisa (Fig. 1), the O Greece 4/94 virus demonstrated a 9% difference in the sequence between nucleotides 475 639 (representing amino acids 158 213 of VP1) and similar levels of difference were observed with Bulgaria 1/93, Morocco 1/91, Italy 1/93 and Turkey 19/91. The differences in sequences exhibited by the Taiwan, Philippine and Vietnam isolates were considerably greater (18%). However, it should be noted that although a small percentage of nucleotide differences are an indication of the similarity of an isolate to the vaccine strain, a high percentage of sequence difference does not necessarily equate to an inappropriateness of the vaccine strain being compared. A case in point is the O1 Manisa strain, still effective at protecting against the Taiwanese outbreak despite an 18% difference in the compared sequences. Sequence comparison of the Algerian 2/99 with the more serologically related O1 strain Manisa recorded a 15% difference between nucleotides 475 639 of VP1. In many cases the sequence relationships between the same selected eld isolates and O1 Lausanne [2] (Fig. 1) showed even greater differences ( ] 18%) than those observed with O1 Manisa. 3.2.2. Asia1 serotypes Sequence comparison of Asia isolates, such as Thailand 10/94 and 4/95 and the vaccine strain Asia1 India also revealed nucleotide differences of around 14%

2112

P.V. Barnett et al. / Vaccine 19 (2001) 21072117 Table 6 Serological relationships (r ) between recent C FMDV eld isolates determined in 2D microneutralization tests using IVB bovine 21 day post vaccinal C1 Oberbayern (batch 320) seraa Virus Bovine antisera C1 Oberbayern C1 Oberbayern C Bangladesh 1/92 C Bhutan 2/91 C Nepal 1/94 C Philippines 4/90 1.00 0.13** 0.20** 0.16** 0.21**

(Fig. 2). In contrast, the sequence examined from Saudi 13/92 was identical to the Asia1 vaccine strain and it is therefore a probability that this outbreak was caused by the vaccine. The three years between the occurrence of the Indian strain, which has been used as a vaccine, and the 1992 outbreak in Saudi Arabia could have been expected to show, at a conservative estimate, 3% difference in nucleotide homology. The recent Malaysian isolate (May 9/99) belongs to a Far Eastern genetic lineage. However, the Turkish isolates (Tur 8/99 and 10/99) show close sequence homology to strains isoTable 5 Serological relationships (r ) between recent A FMDV eld isolates determined in 2D microneutralization tests using IVB bovine 21 day post vaccinal A22 Iraq 24/64 (lot 661), A24 Cruzeiro (lot 4219) and A15 Thailand 1/60 (lot 4220) seraa Virus Bovine antisera A22 Iraq 24/64 A22 Iraq 24/64 A24 Cruzeiro A15 Thailand 1/60 A Albania 1/96 A Brazil 3/93 A Brazil 2/95 A Iran 6/94 A Malaysia 38/95 A Saudi 47/93 A Saudi 16/95 A Thailand 8/88 A Thailand 19/88 A Turkey 1/92 A Turkey 3/92 A Turkey 1/98 A Zambia 90 A Iran 22/99 1.00 0.20 0.11* 0.15* 0.37 0.13* 0.55 0.25 0.70 0.48 0.47 0.28 0.08** 0.29 0.13* A24 Cruzeiro A15 Thailand 1/60 1.00 0.26* 0.32 0.36 0.37 0.13* 0.12** 0.22 0.18* 0.10* 0.05** 0.25

, indicative of reasonable levels of cross protection; , indicative of very dissimilar strains. The (r ) values shown represent the neutralization titre against heterologous virus/neutralization titre against homologous virus. ** r signicantly less than 1.0 (P = 0.01).
a

1.00 0.11* 0.12* 0.14* 0.15* 0.12** 0.09* 0.10* 0.14* 0.12** 0.13* 0.11* 0.21

lated in the Indian sub-continent, and to those found in Iran ( # 2.5%) during 1999.

3.2.3. A serotypes Sequence comparison of all these eld isolates with A24 Cruzeiro [21], A22 Iraq or A15 Thailand showed differences in many cases of \ 14% (Fig. 3), and even

a
, indicative of reasonable levels of cross protection; , indicative of very dissimilar strains. The (r ) values shown represent the neutralization titre against heterologous virus/neutralization titre against homologous virus. * r signicantly less than 1.0 (P = 0.05). ** r signicantly less than 1.0 (P = 0.01).

Fig. 1. Dendrogram depicting genetic relationships between selected eld isolates of FMD virus serotype O and IVB antigen strains O1 Lausanne and O1 Manisa (bold).

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

2113

Fig. 2. Dendrogram depicting genetic relationships between selected eld isolates of FMD virus serotype Asia1 and the IVB antigen strain Asia1 India (bold).

as high as 24% for one isolate, Zambia/90. The closest sequence comparisons, 8% against A22 Iraq, were the isolates from Albania (1/96) and Saudi Arabia (16/95) which suggests that both of these outbreaks were caused by a Middle East strain of type A. The new variant of FMD virus type A, rst isolated from samples originating from Iran in 1996, was conrmed to be genotypically similar to isolates from later outbreaks in Turkey in 1997 and 1998. However, the more recent A isolates found in Iran and Turkey in 1999, which are now appearing in Iraq, are also antigenically and genotypically unique and very different from the 1996 variant. Current evidence suggests that both these variants are co-existing in the susceptible population and like the A Iran 96 the development of another A vaccine strain to control this outbreak is conceivable. The inadequacies of current vaccines and the need to develop further A vaccine strains is a good indication of the problems still caused by this serotype.

3.2.4. C serotypes Sequencing showed that all four isolates in the panel exhibited 18% or greater nucleotide differences in the region examined on comparison with C1 Oberbayern vaccine strain (Fig. 4). Distinct differences between the Asian isolates (Bangladesh 1/92, Bhutan 2/91 and C Nepal 1/94) and the Far East isolate C Philippines 4/90 were also evident.

Fig. 3. Dendrogram depicting genetic relationships between selected eld isolates of FMD virus serotype A and IVB antigen strains A24 Cruzeiro, A15 Thailand and A22 Iraq (bold).

2114

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

Fig. 4. Dendrogram depicting genetic relationships between selected eld isolates of FMD virus serotype C and the IVB antigen strain C1 Oberbayern (bold).

4. Discussion Emergency vaccination can play an important role in the control of FMD outbreaks in countries normally free of FMD. This was demonstrated in 1966 in Sweden when the policy was used as an adjunct to stamping out to control a single outbreak and eradicate the causal virus. The storage of conventional formulated FMD vaccines in a strategic reserve for such eventualities is expensive as vaccine must be replaced every 18 months due to limited shelf-life. An alternative to this costly practice is the well-established method of indenite storage of concentrated, inactivated, FMD antigen at ultra-low temperatures over liquid nitrogen. Cryogenically stored antigen can be rapidly formulated into vaccine when required. This concept is not unique to the IVB, and the recent establishment of a European Community FMD antigen reserve and other examples of individual countries establishing their own FMD reserves, indicate the increasing popularity of this approach. However, the IVB is unique in that it also houses its own dedicated manufacturing facility for the emergency formulation of FMD vaccine, which can be despatched within 23 days of receiving a request. Other advantages of this system include choice of adjuvant during formulation, and regular quality assurance. FMD virus is antigenically very diverse, existing as seven major serotypes with numerous sub-types, and strains selected for incorporation into vaccines must be antigenically similar to current eld strains of FMDV, to ensure an optimal immune response and protection. Arguably the most important advantage of storing concentrated antigens is the versatility to choose the most applicable vaccine strain to deal with an epidemic rather than be reliant on the strain/s already incorporated in a pre-formulated vaccine. Vigilant global surveillance of FMD viruses is of extreme importance

in order to advise on the protection afforded by a given vaccine strain, the necessity for additional antigen strains, and to maintain the ability to respond rapidly, should an infection escalate and threaten any member country of the IVB. The seven strains currently held by the IVB were accepted following the demonstration of a high potency value in cattle (greater than 10 PD50) when formulated with Al(OH)3 and saponin. Such high potency vaccine is essential where the primary aim of the emergency programme is to prevent the spread of FMDV as rapidly as possible. In this respect, the authors have reported on the rate of development of protection in cattle, sheep and pigs following vaccination with some of these emergency vaccines and demonstrated protection against natural airborne infection by pig challenge within 4 days of immunization [13 16]. All the antigens produce vaccines of extremely high potency and offer protection against a broad range of eld isolates. In this study, the authors have monitored the applicability of all the IVB antigen strains against current and past virus outbreaks by in vitro and, where necessary, in vivo methods. Serological results indicate that adequate levels of cross-protection should be attained against O isolates using O1 Manisa or O1 Lausanne and against A isolates using either A22 Iraq or the more recently acquired A15 Thailand. The only important exceptions being the new type A variants isolated from Iran and Turkey in 1996 and 1999 and those from Thailand and Malaysia in 1997. Indeed, none of the available vaccine strains were likely to confer protection against the Iran and Turkey strains which created the need for the commercial sector to develop new A type vaccine strains. Acquisition of strains such as A Iran 96 by the IVB is a matter being considered by the Technical Committee. In light of the risk to South East Europe

P.V. Barnett et al. / Vaccine 19 (2001) 21072117

2115

and thereby the European Union, as outlined at the General Session of the European Commission for the Control of Foot-and-Mouth Disease in FAO Rome, this may be acquired at some future date. Both A22 Iraq and A15 Thailand compensate for the limited antigenic spectrum displayed by the A24 Cruzeiro vaccine strain against the eld isolates studied. The A24 Cruzeiro virus, originating from Brazil in 1955, is a well established and historically important antigen but is considered to be of less relevance today. Improved cross-specicity was observed from use of bivalent A22/A24 vaccine tested against two unrelated strains, Brazil 3/93 and Malaysia 38/95 (unpublished results). In some instances, therefore, the mixing of two or more vaccine strains has a synergistic effect and improves efcacy against antigenically unrelated isolates. According to micro-neutralisation results, Asia1 India vaccine may not confer protection against some of the Asia1 eld isolates, particularly those from South East Asia. Repeat testing of this panel of viruses using a new pool of bovine Asia1 anti-sera substantiated some of the previous results, highlighting Thailand 10/94 and 4/95 in particular as being serologically different. However, the r values for many of the other isolates were also low, i.e. B 0.3, and therefore could also be considered as dissimilar. The appropriation of an additional Asia1 vaccine strain, specically Asia1 Shamir, was considered, and further in vivo tests were carried out using guinea pig adapted isolates. The results of these challenge experiments appeared to indicate Asia1 India vaccine would provide adequate protection against the isolates tested, namely Thailand 4/95 and Nepal 28/90. A similar formulation of Asia1 Shamir vaccine was also compared against these two isolates and the resulting elevated PD50 values supported the wide spectrum of serological protection observed (data not shown). The latter part of 1999 saw the occurrence of outbreaks of Asia 1 FMDV in Malaysia and Turkey. Outbreaks caused by this serotype have been frequent in Malaysia during the last decade. However, the outbreak of Asia1 in Turkey is of particular concern to the IVB. Although the serological results indicated that Asia1 India 8/79 should be effective against the Malaysian isolate its appropriateness for the Turkish isolates was in doubt. For this reason the candidate vaccine strain, Asia1 Shamir, was also examined serologically against these Turkish isolates and showed that it would be a better choice for eld application. Therefore, despite the indication from in vivo studies, that Asia1 India should provide the adequate protection required by the member countries, more condence could be placed on the Asia1 Shamir strain, and therefore this remains a strong candidate for in-

clusion in the IVB portfolio, particularly if future outbreak strains become increasingly dissimilar to the current IVB vaccine strain. The only type C representative in the banks reserve, C1 Oberbayern, is possibly the least antigenically diverse of the current portfolio of IVB antigens. Less importance has been placed on this particular serotype since it has been absent throughout the world in recent years as judged from the isolates received by the WRL. This is reected in the small panel of four viruses examined by both neutralisation and liquid phase blocking ELISA (data not shown) which showed this antigen to have narrow specicity. In a similar analysis to that involving the Asia1 India vaccine strain, the authors are now examining the efcacy of C1 Oberbayern by in vivo challenge against selected eld isolates. Although the classical serotypes have been shown to have good correlation with genetic groupings by phylogenetic analysis, and to some extent also the subtypes [17], there is no direct link between percentage sequence homology and antigenic relationships. It has been reported that quite distantly related isolates may have similar antigenic characteristics due to molecular mimicry at important antigenic sites [18,19]. Conversely, very close sequence homology may conceal large antigenic differences. [20] described a monoclonal antibody escape mutant which was isolated by serial selection with a panel of ve neutralizing monoclonal antibodies. This mutant virus is able to evade neutralization by polyclonal antisera elicited against the parental virus (O1/Kaufbeuren/FRG/66). However, if a comparison is made between the mutant virus and the parent strain the percentage nucleotide difference is only 1.17%. This clearly demonstrates the danger in making the assumption that if two isolates are closely related genetically that the same vaccine can be used. It is important therefore that, when a eld isolate is studied, appropriate techniques are used simultaneously to gain the required information and enable a well thought out disease control programme to be implemented. The serological tests used in this study, are able to show the likely efcacy of vaccines to control outbreaks. The sequence data, on the other hand, by reference to a large database of virus sequences provides more detailed information on possible sources of the strain causing an outbreak. This is particularly highlighted by the Asia1 Saudi 13/92 isolate whose sequence was identical to that of the Asia1 India vaccine strain and strongly suggests that incompletely inactivated vaccine may have been applied in the eld. Phylogenetic analysis can pinpoint where animal movements or importation may have been responsible. Decisions to implement effective animal control measures

2116

P.V. Barnett et al. / Vaccine 19 (2001) 21072117 [5] Mackay DKJ. Serological surveillance for FMD in North Africa. Report, Closed Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot-and-Mouth Disease and the Foot-andMouth Disease Subgroup of the Scientic Veterinary Committee of the Commission of the European Community, Paris, 1999. [6] Booth JC, Rweyemamu MM, Pay TWF. Dose response relationships in a microneutralization test for foot-and-mouth disease viruses. J Hyg 1978;80:31 42. [7] Knowles NJ, Samuel AR. Polymerase chain reaction amplication and cycle sequencing of the ID (VP1) gene of foot-andmouth disease viruses. Report, Closed Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot-and-Mouth Disease, Vienna, Austria, Rome, Appendix 8, 1994: 45 53. [8] Murray V. Improved double-stranded DNA sequencing using the linear polymerase chain reaction. Nucleic Acids Res 1989;17:88 9. [9] Felsenstein J. PHYLIP Phylogeny inference package (version 3.2). Cladistics 1989;5:164 6. [10] Rweyemamu MM, Hingley PJ. Foot-and-mouth disease virus strain differentiation: analysis of the serological data. J Biol Stand 1984;12:323 37. [11] Barnett PV, Statham RJ. Long term stability and potency of antigen concentrates held by the International Vaccine Bank. Report, Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot-and-Mouth Disease and the Foot-and-Mouth Disease Subgroup of the Scientic Veterinary Committee of the Commission of the European Community, UK. Appendix 38, 1998: 2725. [12] Kitching RP, Rendle R, Ferris NP. Rapid correlation between eld isolates and vaccine strains of foot-and-mouth disease virus. Vaccine 1988;6:403 8. [13] Doel TR, Williams L, Barnett PV. Emergency vaccination against foot-and-mouth disease: rate of development of immunity and its implications for the carrier state. Vaccine 1994;12(7):592 600. [14] Cox SJ, Barnett PV, Dani P, Salt JS. Emergency vaccination of sheep against foot-and-mouth disease: protection against disease and reduction in contact transmission. Vaccine 1999;17:185868. [15] Salt JS, Williams L, Statham R, Barnett PV. Further studies on the rate of development of protection in cattle given emergency vaccination against FMD. Report Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot-and-Mouth Disease and the Foot-and-Mouth Disease Subgroup of the Scientic Veterinary Committee of the Commission of the European Community, Moeldling, Vienna, Austria, Appendix 17, 1995: 90 7. [16] Salt JS, Barnett PV, Dani P, Williams L. Emergency vaccination of pigs against foot-and-mouth disease: protection against disease and reduction in contact transmission. Vaccine 1998;16(7):746 54. [17] Martinez MA, Dopazo J, Hernandez J, Mateu MG, Sobrino F, Domingo E, Knowles NJ. Evolution of the capsid protein genes of foot-and-mouth disease virus: antigenic variation without accumulation of amino acid substitutions over six decades. J Virol 1992;66:3557 65. [18] Samuel AR, Knowles NJ, Kitching RP. Serological and biochemical analysis of some recent type A foot-and-mouth disease virus isolates from the Middle East. Epidemiol Infect 1988;101:577 90. [19] He rnandez J, Martinez MA, Rocha E, Domingo E, Mateu MG. Generation of a subtype-specic neutralization epitope in footand-mouth disease virus of a different subtype. J Gen Virol 1992;73:213 6. [20] Crowther JR, Farias S, Carpenter WC, Samuel AR. Identication of a fth neutralizable site on type O foot-and-mouth

can then be taken on a more informed and rational basis. Another aspect of sequence studies are that if some isolates are found to be genetically divergent to existing isolates it may highlight the need to study these isolates more fully with serological tests to gain further insight into their antigenic properties. The recent appearance of genetically and antigenically distinct outbreak viruses in Turkey in 1996 and 1999, that prompted the development of new vaccine strains, appear to be excellent candidates. In the future it would be advantageous to be able to predict the likely antigenic consequences of changes in sequence. With the continuing accumulation of data and studies with monoclonal antibodies this could ultimately be a distinct possibility. All the antigens stored in the IVB are tested annually for stability and efcacy in vivo and in vitro incorporating guinea pig potency tests, sucrose density gradient and SDS PAGE analysis. The antigens are therefore unique in respect to the amount of quality control testing undergone. In many cases these antigens have also been examined experimentally with differing adjuvant formulations and for the rapidity with which they confer protection in the target host. This permits a high degree of condence in vaccine stability, potency and efcacy. The only remaining area of doubt is antigenic relevance to current and future FMD outbreaks. Retrospective studies have shown that the criteria used to select the majority of the banks antigens though intuitively based were accurate. Techniques now available allow the selection to be more scientically based. Nevertheless, the appropriation of further strains remains an important consideration particularly with regard to the Asia1 and C serotypes, which are each only represented by a single strain, and the newly developed and antigenically unique A Iran 96 vaccine strain. These observations underline the importance of continually surveying and characterising eld isolates. Acknowledgements This work was supported nancially by MAFF, UK (project number SE2805). References
[1] Kitching RP. A recent history of foot-and-mouth disease. J Comp Pathol 1998;118:89108. [2] Beck E, Strohmaier K. Subtyping of European foot-and-mouth disease virus strains by nucleotide sequence determination. J Virol 1987;61:16219. [3] Donaldson AI, Doel TR. Foot-and-mouth disease: the risk for Great Britain after 1992. Vet Rec 1992;131:11420. [4] Samuel AR, Knowles NJ, Mackay DKJ. Genetic analysis of type O viruses responsible for epidemics of foot-and-mouth disease in North Africa. Epidemiol Infect 1999;122(3):52938.

P.V. Barnett et al. / Vaccine 19 (2001) 21072117 disease virus following characterization of single and quintuple monoclonal antibody escape mutants. J Gen Virol 1993;74:1547 53. [21] Weddell GN, Yansura DG, Dowbenko DJ, Hoatlin ME, Grub-

2117

man MJ, Moore DM, Kleid DG. Sequence variation in the gene for the immunogenic capsid protein VP1 of foot-and-mouth disease virus type A. Proc Natl Acad Sci USA 1986;82:2618 22.