vaccines and viral antigenic diversity

Rev. sci. tech. Off. int. Epiz., 2007, 26 (1), 69-90

Vaccines and viral antigenic diversityJ.A. Mumford

Cambridge Infectious Diseases Consortium, Department of Veterinary Medicine, University of Cambridge,Madingley Road, Cambridge CB3 0ES, United Kingdom

SummaryAntigenic diversity among ribonucleic acid (RNA) viruses occurs as a result ofrapid mutation during replication and recombination/reassortment betweengenetic material of related strains during co-infections. Variants which have aselective advantage in terms of ability to spread or to avoid host immunitybecome established within populations. Examples of antigenically diverseviruses include influenza, foot and mouth disease (FMD) and bluetongue (BT).Effective vaccination against such viruses requires surveillance programmes tomonitor circulating serotypes and their evolution to ensure that vaccine strainsmatch field viruses. A formal vaccine strain selection scheme for equineinfluenza has been established under the auspices of the World Organisation forAnimal Health (OIE) based on an international surveillance programme. Aregulatory framework has been put in place to allow rapid updating of vaccinestrains without the need to provide full registration data for licensing the updatedvaccine. While there is extensive surveillance of FMD worldwide and antigenicand genetic characterisation of isolates, there is no formal vaccine strainselection system. A coordinated international effort has been initiated to agreeharmonised approaches to virus characterisation which is aimed at providingthe basis for an internationally agreed vaccine matching system for FMDsupported by the OIE. The emergence and spread of BT in Europe have resultedin an intensification of vaccine evaluation in terms of safety and efficacy,particularly cross-protection within and between serotypes. The most importantrequirement for producing vaccines against viruses displaying antigenicdiversity is a method of measuring antigenic distances between strains anddeveloping an understanding of how these distances relate to cross-protection.Antigenic cartography, a new computational method of quantifying antigenicdistances between strains has been applied to human and equine influenza toexamine the significance of viral evolution in relation to vaccine strains. Thismethod is highly applicable to other important pathogens displaying antigenicdiversity, such as FMD.

KeywordsAntigenic cartography – Antigenic diversity – Bluetongue – Cross-protection – Foot and mouth disease – Influenza – Serotype – Surveillance – Topotype – Vaccine strainselection.

IntroductionUnderstanding the genetic diversity of viral pathogens andhow it is modulated by host immunity, transmission bottle-necks, epidemic dynamics and population structures isessential for the development of effective control measures(26). Ribonucleic acid (RNA) viruses with their short

replication times, particular propensity to mutate duringreplication, and other strategies for diversification, are aparticular challenge (27). The best-known example ofantigenic diversity of a virus and its importance forvaccines is that of human influenza for which there is in-depth knowledge of virus serotypes, their evolution andtheir significance for vaccine efficacy. The global

surveillance and monitoring of human influenza andemergence of new viruses from animal reservoirs areembedded in the Global Influenza Programme of theWorld Health Organization (WHO) (67) and the basicrequirements for effective surveillance, outbreak responseand updating of vaccine strains are well established. Thedevelopment of the programme has required thecoordination of a network of reference laboratories, anannual strain review mechanism, acceptance ofrecommendations on vaccine strains by nationalauthorities, internationally accepted standards for vaccinesand an updating mechanism that can respond rapidly tochanging epidemiological conditions. This article focuseson diseases of veterinary species which have similarrequirements and reviews progress in understandingpathogen diversity and in establishing systems to identifyappropriate vaccine strains in response to changingepidemiological situations.

There are a number of viral diseases affecting animalswhich are antigenically diverse and require similarapproaches to control by vaccination. Probably the moststudied in relation to vaccine strain selection are (i) influenza, and in particular equine influenza (43), and(ii) foot and mouth disease (FMD) (69). Both diseases arecaused by viruses demonstrating a high degree of antigenicdiversity and evolution. Additionally, there are otherveterinary viruses which, although they do not show thesame degree of antigenic evolution, do display multipleserotypes. Such viruses include, for example, theorbiviruses, bluetongue (BT) and African horse sickness(AHS), where serotype identification is important forappropriate vaccine strategies.

Mechanisms producing viral diversityDuring replication of viruses, mistakes occur in the processof producing copies of viral nucleic acid which are knownas mutations. Viruses containing ribonucleic acid (RNA)generate a higher rate of mutation than viruses containingdeoxyribonucleic acid (DNA) because there is no effectiveproof-reading mechanism in the replication strategiesemployed by RNA viruses (20). As a result, ‘clouds’ ofmutants or quasi species are generated during infection,however, many fail to transmit, a phenomenon known astransmission bottle-necks.

If random mutations have some selective advantage interms of viral fitness (ability to replicate within the hostand transmit and spread in a population) or avoidance ofthe immune response (ability to avoid neutralisation byantibody generated by earlier related strains) then thesemutations may become fixed in the population of progenyviruses (7). These processes are well recognised in anumber of RNA viruses such as influenza, FMD and BT.

Genetic and antigenic driftThe progressive accumulation of random geneticmutations is known as genetic drift which may or may notresult in changes in amino acid sequence of viral proteins.If the genetic code for amino acid changes then this resultsin altered antigenic characteristics and is known asantigenic drift. There are a number of factors which drivethe selection of antigenic variants and in some populationsantigenic variants co-exist while in others emergingvariants replace earlier viruses. These processes are knownas viral evolution and understanding its basis andpredicting likely trends are an important aspect ofcontrolling virus diseases (7).

Antigenic diversity arising from recombination and reassortmentGenetic and associated antigenic changes can also occur asa result of deletions and genetic rearrangements caused bynucleic acid splicing and recombination events, as hasbeen reported for foot and mouth disease virus (FMDV). Ininfluenza virus infections for example, a key event arisingfrom the segmented genome is the reassortment of genesduring mixed infections of the same cells with twodifferent viruses. This is an important mechanism for theemergence of new influenza viruses and has been reportedfor human, avian and swine influenzas (11, 37). Theprocess of generating a new influenza virus with a uniquecombination of surface glycoproteins by reassortment iscalled antigenic shift. This process has also been reportedin BT virus (BTV). Although virus virulence is not beingconsidered in detail here, it is noteworthy thatreassortment of surface glycoproteins from one virus on anovel background of internal genes from another virus cansignificantly alter the pathogenicity of influenza viruses.

Selection and survival of variantsKey factors affecting the selection of variants relate to thevirus, the host immune response and the population sizeand structure. For example, viruses with a high infectivityhave a selective advantage as they are more successful intransmission. Viruses with altered antigenic sites,particularly those involved in virus–cell attachment, maybe capable of avoiding neutralising antibody present in apopulation as a result of previous infection. Thisphenomenon of immune selection of variants isparticularly important for infections where immunity isnot lifelong, where there is a high rate of mixing withinhost populations and where animals are exposed torepeated infections by closely related viruses.

Selection and survival of variant viruses is also affected byhost population structure and size, which is well illustratedby considering influenza of different species. Viral

Rev. sci. tech. Off. int. Epiz., 26 (1)70

evolution in human influenza has been extensively studiedand within the subtypes, one strain largely replacesanother on a global scale, as viruses replicate in a partiallyimmune population creating the need to escape theimmune response.

A similar pattern is seen in equine influenza, althoughthere is not such a strong effect due to lower populationdensities and lower rates of infection. Additionally, theremay be little mixing among different populations, whichencourages evolution of discreet co-existing lineages underpotentially different selection pressures.

By comparison, influenza in pigs and domestic poultryheld in isolated environments is, in general, reliant on therapid introduction and constant availability of youngimmunologically naïve hosts within the breeding andfarming structures. There are few opportunities for re-infection as stock is slaughtered at a young age.Maintenance of infection in such naïve populations andabsence of partially immune older animals does notprovide the same driving forces for immune selection andlack of mixing between farms encourages development ofmultiple lineages. For example, swine influenza, whileshowing antigenic diversity with multiple strains co-existing, shows less immune-driven evolution.

Thus, antigenic variability is driven not only by the abilityof the viruses to mutate and their ability to transmitbetween hosts, but also by the opportunities for survivalthat present themselves as a result of the immuneenvironment and population size and structure.

Equine influenza: addressing issues of antigenicdiversity in relation to vaccinesBackground to vaccination with equine influenzaIn 1956 the H7N7 subtype of equine influenza was firstisolated in Prague and the prototype was designated asA/equine1/Prague/56. Within a 7 year period a secondequine influenza virus of the H3N8 subtype was isolatedfrom horses in Florida and was designatedA/equine2/Miami/63. Both subtypes caused majorepidemics and in the mid 1960s vaccination against equineinfluenza was introduced. The early vaccines contained theprototype strains of the H7N7 and H3N8 subtypes grownin eggs, inactivated and combined with an oil adjuvant.The early products were not widely accepted as they werehighly reactogenic, but as acceptable adjuvants were found

vaccination became the accepted means of control,particularly in performance animals such as race horses.

In well-vaccinated populations vaccine breakdownattributable to H7N7 viruses was rare or non-existent,however, repeated infections with the H3N8 subtype havebeen reported over a long period. Much research has beenundertaken to establish the contribution of vaccinepotency and antigenic variation to this observed vaccinefailure (43, 44).

Virus structure and variabilityInfluenza viruses are single-stranded RNA viruses withsegmented genomes comprised of 8 segments (genes)coding for structural components of the virus particle andnon-structural components important for replicationwithin host cells. The two most important structuralproteins demonstrating genetic and antigenic variationwhich are relevant to protection and vaccination are theenvelope glycoproteins, the haemagglutinin (HA) and theneuraminidase (NA). Of these, the HA is particularlyimportant as it mediates virus attachment to the host celland antibody induced against the HA neutralises virusinfectivity. The ability of the virus to evolve in terms of theantigenic character of the HA (antigenic drift) is crucial foravoidance of population immunity and immunity derivedfrom inactivated vaccines, which is largely reliant onantibody to HA. The NA is involved in elution of virusfrom cells and the spread of infection between cells, butalthough the NA is known to vary, there is little information on the impact of its antigenic drift onvaccine efficacy.

Antigenic and genetic variation of equine influenza virusesAs with other influenza A viruses, both subtypes of equineinfluenza exhibit genetic and antigenic variation. Theevolution of the HA gene has been well studied because ofits importance in relation to virus neutralisation andprotection. Attention has been focused on the A/equine/2(H3N8) virus as this has been the predominant straincirculating since the 1960s and more importantly becausethere have been repeated reports of vaccine breakdown inthe field. The majority of studies on the antigenic characterof the HA and its relationship to viral neutralisation havebeen conducted using haemagglutination inhibition (HI)tests, exploiting the fact that influenza viruses naturallyagglutinate erythrocytes and that antibody inhibitingagglutination is a measure of virus neutralisation (VN).Much antigenic analysis of influenza viruses has relied onthe use of ferret sera as this species is susceptible toinfection with influenza and provides strain specificantisera which can discriminate between strains in HI tests.

Rev. sci. tech. Off. int. Epiz., 26 (1) 71

In 1983, Hinshaw reported that there had been majorantigenic drift in viruses isolated between 1979 and 1981as compared with the prototype virus Miami/63 (31).However, they also recognised, based on antigenic analysiswith ferret sera in HI tests, that some viruses similar to theprototype Miami/63 virus were co-circulating with themore recent variants. On the basis of this data theyrecommended that additional strains (Fontainebleau/79 orKentucky/81) should be included in vaccines, which at thetime contained only the prototype H3N8 virus, Miami/63.At that time surveillance and virus collection was sporadicand there was no certainty that the strains selected asvaccine strains were representative of the predominantstrains circulating.

Subsequent genetic analysis (32) based on sequencing theHA gene of a larger panel of viruses from around the world,revealed that the A/equine/2 HA gene was evolvingessentially as a single lineage, however, antigenic analysisrevealed that the resultant changes to the amino acidsequence gave rise to viruses which were both similar toand distinct from the prototype H3N8 virus, Miami/63. Itwas noted that the pattern of evolution was similar to thatseen in human influenza and it was proposed that it wasdriven by immunological pressure, i.e. the existingimmunity to historical viruses present in the olderpopulation. This study demonstrated that not only was thedegree of antigenic drift important, i.e. the number ofmutations which had arisen and become fixed in the HAmolecule, but their identity and location was alsoimportant because genetically distant viruses couldnevertheless react in a similar way in HI tests.

In both these studies antigenic differences betweenprototype and recent strains were measured using HI testsand post infection ferret or rabbit sera or monoclonalantibodies. Where fourfold differences in reactivity of serawith different virus strains could be detected it wasconcluded that the viruses were significantly different interms of antigenicity, which may have implications forvaccines. At that time no attempt was made to assess thesignificance of such antigenic differences for vaccineefficacy in the target species. The significance of fourfolddifferences in HI tests was assumed to have immunologicalrelevance based on experience with human influenzaviruses.

The conclusion that the antigenic drift might compromisevaccine efficacy was not accepted by others. Burrows et al(13) concluded that the antigenic differences detectedbetween the prototype strain Miami/63 and the newvariants Fontainebleau/79 and Kentucky/81(demonstrated using ferret sera and monoclonalantibodies) were unlikely to be important because postvaccination sera from horses vaccinated with Miami/63was highly cross-reactive with the recent 1979 isolates(13). This lead to a debate about the relevance of antigenic

differences detected using post infection and postvaccination sera from laboratory animals as compared tosera from target species (43) and hindered progress in theunderstanding of significance of antigenic variation inequine influenza viruses in relation to vaccine efficacy.

Genetic and antigenic drift has been periodically reportedfrom a number of different countries (34, 52). However, aparticularly important observation was made in a jointstudy by OIE Reference Laboratories in the UnitedKingdom (UK) and in the United States of America (USA).These laboratories examined viruses from 1963 to 1994and revealed that genetic and antigenic variants were co-circulating as a result of a divergence in the single lineageof the H3N8 viruses (originally described by Kawaoka et al.[32]) into two sublineages representing isolates originatingfrom the Americas on the one hand and viruses fromEurope and Asia on the other (15). However, these lineagesdid not remain geographically separate and in the early1990s American-like viruses were identified in Europeprobably reflecting the significant traffic of horses from theUSA to Europe for racing (Fig. 1).

At that time vaccines manufactured in America containedAmerican isolates and most vaccines manufactured inEurope contained European viruses. Thus, horsesvaccinated with European viruses were reliant on cross-protection when exposed to viruses from the Americanlineage and vice versa.

The two sublineages of the H3N8 viruses have continuedto evolve and sequencing has revealed the appearance of anumber of clades (subgroups) within the lineages, some ofwhich have geographic origins, e.g. the South Americanbranch of the American sublineage (34).

Evidence of antigenic drift affecting vaccine efficacy in the fieldVaccine breakdown has been reported during a number ofoutbreaks of influenza A/equine/2 over many years, butthis had been largely attributed to poor vaccine efficacy, orvaccination schedules which did not accommodate theshort duration of immunity provided by the earlyinactivated vaccines. In 1976 (68) and 1979 (12)vaccinated horses became infected, but those horses whichsuccumbed to infection had low or undetectable antibodyat the time of exposure. Thus, at this stage there was nofirm evidence for antigenic drift being the explicit cause ofvaccine failure.

In contrast, in 1989 a major epidemic of equine influenzaA/equine/2 occurred in the UK and elsewhere and firstcases were identified in regularly vaccinated army horseswith high levels of antibody prior to infection (36).Although the infection was generally mild in well


vaccinated horses it spread rapidly through populationsindicating that levels of virus shedding were significanteven in the absence of severe clinical signs. At the time ofthe outbreak, available vaccines contained the prototypestrain Miami/63 and a strain from the 1979-1981 epidemicsuch as Fontainebleau/79, Kentucky/81, Brentwood/79 orBorlange/79.

In the intervening ten years between 1979 and 1989 therehad been a major improvement in vaccine potency as aresult of the introduction of challenge models in the targetspecies to assess vaccine efficacy and establishment ofacceptability thresholds for vaccines in terms of antigencontent (measured as µg HA) and levels of antibody(measured by Single Radial Haemolysis [SRH]) that areconsistent with protection. As a result many of theEuropean vaccines available at that time had demonstrableefficacy against homologous strains as judged by HAcontent, serological responses generated and protectionagainst challenge infection (46, 47). These observationsfurther supported the conclusion that significant antigenicdrift had occurred in 1989.

Significance of antigenic variation measured byhaemagglutination inhibition tests in relation tovaccine efficacy in the target speciesAs knowledge of the genetic and antigenic diversity andevolution of equine influenza grew and field observations

suggested antigenic drift may have played a role in vaccinebreakdown, it became essential to establish the significanceof antigenic variability as measured by HI tests with ferretsera for vaccine efficacy in the target species.

A series of four viruses spanning a period of 26 years(Miami/63, Fontainebleau/79, Kentucky/81 andSuffolk/89) were examined in a cross-protection study inponies in which groups of ten ponies were vaccinated withtwo doses of inactivated vaccines prepared from each straincontaining equivalent HA content and challenged with arecent isolate Sussex/89 (43). Protection was measured interms of serological responses, virus excretion and clinicalsigns following challenge. The key findings from this studywere that vaccines derived from both recent and historicviruses provided equally effective clinical protection interms of reduction in pyrexia and coughing in vaccinates ascompared to unvaccinated controls. In contrast, the abilityof vaccines to protect against infection and suppress virusexcretion following challenge was directly related to theantigenic relatedness of the challenge and vaccine viruses,with the Miami/63 vaccine allowing significantly morevirus excretion than the Suffolk/89 virus most closelyrelated to the challenge virus Sussex/89 (Table I). Thisdifference in protection could not be attributed todifferences in potency because similar levels of HI antibodyto the challenge virus (Sussex/89) were stimulated by theMiami/63 and the Suffolk/89 vaccines (Table II).Furthermore, SRH antibody levels to Sussex/89 werehigher in the Miami/63 vaccine group than in theSuffolk/89 vaccine group (Table III).


Fig. 1Phylogenetic tree of H3N8 equine influenza viruses showing divergence from a single lineage into the American and Eurasiansublineages

‘American-like’ lineage ‘European-like’ lineage

Two viruses fromdifferent lineages

isolated on the same dayand in the same location

in 1993

Newmarket/2/93UK/00

Miami/63

Kentucky/94

1989 epidemic

1979 epidemic

1963 epidemic

Kentucky/98

UK/98

Newmarket/1/93

This study was the first to demonstrate that antigenicdifferences between equine influenza strains detected byHI tests with ferret antisera were significant for vaccineefficacy in the target species, particularly with respect toprotection against infection and virus excretion. However,it also demonstrated that vaccines containing viruses ill-matched to epidemiologically relevant strains provided adegree of clinical protection which could mask infectionwhile allowing copious amounts of virus to be excreted.These data supported the conclusion that for the control ofinfluenza at the herd level it is important that vaccinescontain virus strains which match currently circulatingstrains in order to minimise virus shedding.

These observations also raised the question of geographicvariation and its importance for vaccine strain selection.While the majority of vaccines available in the USA,Europe and centres of thoroughbred racing around theworld are produced by large multinational companies,other vaccines are made locally for specific populations, forexample in South America, Japan, Eastern Europe andIndia. It became important to explore whether antigenicdifferences between viruses of different locations are likelyto affect vaccine efficacy.

Competition animals travel extensively and internationallyand it is likely that such horses are exposed to viruses fromdifferent locations. While originally it was held that equineinfluenza evolved as a single lineage, the observationsmade in the early 1990s revealed that the A/equine/2lineage diverged into American and Eurasian sublineages.Subsequent to that observation a further sublineage of theAmerican-like viruses has been recognised as originatingfrom South America (34) (Fig. 2). It is central tointernational control of equine influenza to understand thesignificance of the antigenic differences between thesesubpopulations (or clades) for vaccine efficacy.

With this objective in mind a series of vaccination andchallenge studies in the target species have been performedto examine cross-protection between strains arising fromthe American and Eurasian lineages. The prototype virusesNewmarket/2/93 (Eurasian) and Newmarket/1/93(American) were selected and used in cross-protectionstudies in horses (16, 82). As with the study to examine thesignificance of temporal antigenic drift, it was found thatthe vaccines containing viruses from the two lineagesprovided a significant degree of cross-protection againsteach other in terms of suppression of clinical signs such ascoughing and pyrexia. Interestingly it was also found thatthe American lineage virus protected equally well againstthe European virus as against the homologous Americanvirus in terms of infection and reduction in virus excretion(82). In contrast, the European lineage virus vaccine wasnot as effective in protecting against infection and virusexcretion when challenged with the American lineage virusas compared with the protection afforded against ahomologous challenge (16).

While the differences between the protection observedusing the different vaccines were subtle underexperimental conditions in limited groups of ponies, it hasbeen demonstrated, using mathematical models, that thelikely impact of such variations in suppression of virusexcretion on immunity in a population is significant (53)(Fig. 3). Furthermore, field observations have supportedthis conclusion. In a limited outbreak of the Europeanlineage virus, it was found that horses vaccinated with aproduct containing a European virus and with SRHantibody levels above the protective threshold wereprotected against infection (50). In contrast, in a similar


Table IVirus excretion following aerosol challenge with A/equine/2(H3N8) virus Sussex/89 from ponies vaccinated withmonovalent vaccines

Vaccine groupNumber of ponies Mean duration

excreting virus (days)

Miami/63 9/10 3.6 *

Fontainebleau/79 9/10 3.3 *

Kentucky/81 8/10 2.5 **

Suffolk/89 5/9 1.6 ***

Controls 10/10 5.1

* p<0.05, **p<0.01, ***p>0.001 (compared to controls)

Table IICross-reactivity of haemagglutination inhibition (HI) antibodystimulated by two doses of monovalent vaccine

Vaccine Mean HI titres to virus strains

M/63 F/79 K/81 S/89

Miami/63 1.58* 0.95 1.32 0.7

Fontainebleau/79 0.9 1.15 1.40 0.95

Kentucky/81 1.08 1.11 1.54 0.9

Suffolk/89 0.7 0.85 1.0 1.0

* HI titre log 10

Table IIICross-reactivity of single radial haemolysis (SRH) antibodystimulated by two doses of monovalent vaccine

Vaccine Mean SRH antibody to virus strain

M/63 F/79 K/81 S/89

Miami/63 125.7* 148.2 103.7 143.7

Fontainebleau/79 41.6 54.0 46.2 55.6

Kentucky/81 74.0 75.2 82.7 69.6

Suffolk/89 18.0 46.1 48.9 75.2

* mean area of zone of haemolysis to specified strain (mm2)

Annual review of vaccine strains and criteriafor changing strainsWhile cross-protection studies in the target species are theultimate test of the significance of antigenic drift, it is notpractical to base vaccine strain selection on such studiesbecause of the difficulty of accessing influenza-free ponies,and the cost and time required to undertake large animalexperimentation. This holds true for many virus vaccines.Therefore, in order to identify a reliable predictor ofsignificant antigenic drift, there has been considerableeffort to examine the relationship between protection inthe target species, protection in hamsters as a small animalmodel and antigenic differences discriminated by HI testsusing ferret, horse and hamster sera (16).

As already mentioned, ferrets produce highly strain-specific sera following infection with influenza strains,whereas horse sera are more cross-reactive. However,analysing the reactivity of post infection ferret sera in HItests remains a useful way to compare the antigenicdifferences between strains and it provides an indication ofcross-protection (Fig. 5).

Surveillance and equine influenza expertsurveillance panelAs there is considerable international traffic of Equidae, it isimportant to conduct surveillance on a global scale andthere are continuing efforts to collect viruses from aroundthe world for sequencing and antigenic analysis. While thenumbers of viruses screened are low by comparison withhuman influenza, surveillance has provided a picture ofthe evolution of equine H3N8 strains and the importanceof inadequately vaccinated animals in the transmission ofviruses globally. Based on the WHO model for surveillance,


Fig. 2Phylogenetic tree of the American lineage of H3N8 equineinfluenza viruses showing South American clade

American sub-lineage

South American clade

Newmarket/1/93

Newmarket/03

outbreak caused by an American lineage virus, horsesvaccinated with a European virus vaccine were notprotected even when antibody levels were above theprotective threshold (49) (Fig. 4). Thus, predicting thelikely efficacy of a vaccine is based not only on potency butalso suitability of the vaccine strains in the field.

Fig. 3A model of the probability of outbreaks occurring throughout ayear in which horses are vaccinated on a 6 monthly basis witheither a strain that matches the outbreak strain or aheterologous strain

Vaccine strain heterology significantly increases risk of outbreaks

Week (starting Jan.1)

homologous vaccineheterologous vaccine

Prob

abili

ty o

f epi

dem

ic �

5%

5 10 15 20 25 30 35 40 45 50

1

0.8

0.6

0.4

0.2

0

analysis of viruses and vaccine strain selection, an EquineInfluenza Expert Surveillance Panel has been set up underthe auspices of the OIE to review on an annual basisoutbreaks of equine influenza, vaccine performance,antigenic and genetic character of new virus isolates and totake decisions on the need to update vaccine strains. Thepanel includes experts from the WHO collaboratinglaboratories at the National Institute of Medical Researchand the National Institute for Biological Standardisationand Control in London, the three OIE Equine InfluenzaReference Laboratories in Germany, the UK and the USAand other experts involved in equine influenzasurveillance. Their conclusions are reported annually bythe OIE.

Developing criteriaOriginally, the criteria that were applied to decisions aboutthe need to change vaccine strains were based on thoseused for human influenza and included vaccinebreakdown in the field, fourfold differences detected in HItests with ferret sera between vaccine strains andpredominant field isolates, discrimination between vaccineand field viruses by post-vaccinal equine sera and geneticsequence of the HA1 molecule. Additionally, these criteriahave been judged against cross-protection studies in horsesand hamsters in order to validate their relevance to thecriteria applied to equine influenza viruses. It has becomeclear that post-vaccinal horse sera are generally unable todiscriminate between viruses unless there are majorantigenic differences, therefore this test has become lessimportant in the decision-making processes. Decisions tochange vaccine strains are normally conservative and areonly recommended when there are measurable antigenicdifferences as a result of significant genetic mutationsbetween vaccine and predominant field strains andevidence of vaccine breakdown. For example, Europeanlineage viruses which can be discriminated from vaccinestrains based on fourfold differences with ferret sera butwhich have not established and spread in the equinepopulation have not warranted a recommendation tochange vaccine strains.

To date, strain differences identified with ferret sera appearto correlate well with limited cross-protection studiesconducted in horses, however, patterns of cross reactivitybetween panels of ferret sera and viruses are complex anddifficult to interpret by eye. Recent advances incomputational methods are revolutionising the way suchdata can be analysed and a method known as antigeniccartography has been applied to historical data fromhuman influenza and equine influenza (65). This


Field studies on vaccine performance

1995 outbreakVaccine strain = EuropeanOutbreak strain = Europeani.e. HOMOLOGOUS challenge

1998 outbreakVaccine strain = EuropeanOutbreak strain = Americani.e. HETEROLOGOUS challenge

Protected Infected Protected Infected

SRH

antib

ody

(mm

2 )

SRH

antib

ody

(mm

2 )

Fig. 4Prechallenge single radial haemolysis (SRH) antibody in protected and susceptible horses in an outbreak where the field and vaccinestrains were homologous or heterologous

300

150

0

300

150

0

Fig. 5Antigenic distances between equine H3N8 viruses measuredwith haemagglutination inhibition tests using post infectionferret sera

American lineage

Ken/81

N/1/93

Ken/90

Ell/89

Aru/91

Ken/91

Sx/89Yve/89

Lam/92

H/1/95

Suf/89

N/2/93

788/91

HK/92

Ber/94

European lineage

technique provides a visual image of the antigenicdistances between viruses and how they cluster and isbeginning to provide a method to assess whether variantsanalysed are evolving along a main lineage or whether theyare unusual variants distant from the predominantantigenic types. Such information is very useful in decidingwhich strains are most suitable for selection as vaccineviruses, i.e. which have the widest cross-reactive repertoirewith viruses in the field. The method has been adopted forthe annual selection of human influenza vaccine strainsand is being developed for equine influenza.

Regulatory framework for updating vaccine strains

The OIE Manual of Diagnostic Tests and Vaccines forTerrestrial Animals (Terrestrial Manual) (81) providesdetailed recommendations for vaccine strains and vaccinepotency testing. The standards it contains are generally inline with the European Pharmacopoeia Monograph oninactivated equine influenza vaccines and the standardunder development by the United States Department ofAgriculture (USDA).

The majority of equine influenza vaccines are inactivatedwhole virus (46) or viral subunits (47) combined with anadjuvant. The immune response of the horse to vaccinationis relatively short-lived and multiple doses are required tomaintain complete protection against infection, although adegree of clinical protection is provided with fewer doses.

The basis of vaccine potency for inactivated vaccines is wellunderstood and relates to the amount of immunologicallyactive HA contained in the vaccine and the efficacy of theadjuvant in enhancing circulating antibody to HA (80).Many studies in immunologically naïve horses havedemonstrated a direct relationship between µg HA invaccines (79) and antibody responses in horses measuredusing an SRH test (44). Furthermore, the level of SRHantibody stimulated is indicative of the level of protectionacquired against challenge infections in vaccinated horses,with 150 mm2 being identified as the threshold forprotection, provided that the vaccine contains a virusantigenically similar to that being used to test the vaccineby challenge infection (44). Furthermore, this thresholdfor protection against experimental infection is valid for afield situation (50). Therefore, the efficacy of a vaccine in afield situation can be predicted based on accuratemeasurement of immunologically active HA in the vaccine,SRH antibody stimulated by the HA in combination withadjuvant and protection against challenge infection;however, the predictions will only be accurate if the virusused as a standard for the single radial diffusion (SRD), oras antigen in the single radial haemolysis (SRH), or aschallenge virus for experimental infection, is antigenicallyindistinguishable from the vaccine strain.

The requirements of the European Pharmacopoeia forlicensing equine influenza vaccines are described inMonograph No. 249 and utilise these relationships. Testingrequires measurement of vaccine antigen, antibodyresponses in horses, and challenge infection studies with atleast one virus included in the vaccine (23).

Vaccine strains are recommended by the Equine InfluenzaExpert Surveillance Panel and are published by the OIE.Currently, it is recommended that vaccines should containrepresentatives of the Eurasian and American sublineagesof the H3N8 virus. Inclusion of H7N7 virus is no longerrecommended on the basis that such a virus has not beenisolated for more than 20 years. Viruses originatingbetween 1989 and 1993 are still accepted for the Europeanlineage, however, recent antigenic drift and field outbreakscaused by American lineage viruses have lead to arecommendation that vaccine viruses should be updated torepresentatives from 2003 such as South Africa/2003. Theselection of virus strain is not prescriptive but selectedstrains must be shown to be antigenically similar to thoserecommended.

Fast track licensing systemOnce new recommendations are made it is highly desirablethat vaccines are updated as quickly as possible and to thisend a fast track licensing system has been developed forupdating vaccine viruses in Europe. These Guidelines,which have been developed by the ImmunologicalWorking Group of the European Medicines EvaluationAgency (22), recognise the well-established relationshipbetween µg immunologically active HA in the vaccine,levels of SRH antibody generated in the target species andprotection against challenge infection. They operate on theprinciple that if a vaccine has been licensed according tothe European Pharmacopoeia standards, which also usethese relationships in their requirements for potency andefficacy testing, and that in the process of updating avaccine strain no other parameter of the vaccine ischanged, then manufacturers are only required todemonstrate safety and the ability of the final product togenerate protective levels of antibody in the target speciesagainst the new strain. This obviates the need for challengestudies and generation of duration data, significantlyreducing the testing required to license the updatedvaccine.

International considerations for vaccine strain selection and standardisation of licensing proceduresThe majority of vaccines are made in the USA or Europe,and efforts are ongoing to harmonise licensing proceduresbetween the European Pharmacopoeia and the USDA.


A series of WHO/OIE consultations have been held towork towards international harmonisation of vaccinestandards (45). In recent years, challenge tests have beenaccepted by the USDA as useful for efficacy and adocument is now under review to provide a fast tracklicensing system for updating strains for vaccines producedin the USA (70).

Influenza of other speciesThe same basic principles apply for influenza of otherspecies, but control processes other than vaccination maybe more suitable. The relevance of antigenic diversity hasbeen examined for swine influenza vaccines (72) and goodcross-protection has been demonstrated between divergentstrains. This has been attributed to the use of very potentadjuvants in swine vaccines which may compensate forantigenic differences. Thus, to date, vaccine strain selectionhas not become an important issue for swine influenza.

With the recent outbreaks of H5 and H7 avian influenzathere is an increasing interest in vaccination as a method ofcontrol to avoid massive slaughter of infected flocks. Themain aspect of genetic variation studied in avian influenzahas been the switch to highly pathogenic virus from viruseswith low pathogenicity of the same serotype. However,there have been some recent reports of antigenic driftoccurring under the immune pressure of vaccination (35).Therefore, it is likely that if vaccination becomes widelyused to protect poultry against avian influenza moreattention to strain evolution and vaccine strain selectionwill be required.

Foot and mouth disease virusIntroductionThe potential impact of antigenic diversity on the controlof FMD is well recognised, however, the task of setting upadequate response systems is enormous and needs to takeaccount of the number of serotypes and subtypes, hostrange, political and socio-economic constraints. Since the2001 outbreak of FMD in Europe caused by serotype O,there have been renewed efforts to improve the proceduresin place for surveillance of FMD on an international scale.These include collection and submission of viruses toreference laboratories, and development of the scientificand technical approaches to examining antigenic diversityamong FMDV strains and assessment of its relevance forvaccine strain selection. These issues have been examinedin a number of reviews on FMDV vaccines (17, 18) andvaccine strain selection (54) and are addressed in the mostrecent foot and mouth disease chapter in the OIETerrestrial Manual (Chapter 2.1.1.). The importance of

having an early warning system for emergence of variantstrains is well recognised. The ability to rapidly analysenew viruses and measure their antigenic relatedness toexisting vaccine strains is crucial to providing effectivevaccines used in rapid response control programmes andfor laying down new viruses in vaccine banks.

Genetic and antigenic variabilityThe molecular basis of antigenic variation in FMDV hasbeen extensively studied and it is well known that FMDVexhibits a high degree of genetic and antigenic variation(21). As with other RNA viruses such as influenza, thishigh level of variation is attributable to the error-pronereplication of viral RNA and the lack of a proof-readingmechanism associated with the viral replicase (20, 66).Thus, mutations are constantly being produced in progenyviruses and subsequently selected for or against as the virusis transmitted within a population, depending on whetherthe mutations are beneficial for virus survival (29, 38).There are 7 serotypes of FMDV known as O, A and C(historically regarded as European types), Asia-1, and SAT1, 2 and 3 (from the South African Territories) (6, 75).Within each serotype there are varying degrees of diversitywith subtypes recognised in some serotypes. There is aparticularly high diversity among SAT 1 and 2 viruseswhich has been ascribed to generation of variants inpersistently infected buffalo (75). Antibody generated byinfection or vaccination against one serotype fails to cross-protect against all other types. Furthermore, antigenicdifferences within a serotype may be so great that there islittle or no cross-protection between strains of the sameserotype (3).

During infection some mutations are selected under theinfluence of immune pressure (10), while others becomefixed even in the absence of immune pressure (19, 64).This viral evolution can occur in distinct populations ofsusceptible animals in separate geographic locations (Fig. 6) (77), resulting in the maintenance and evolution ofdistinct lineages within an FMDV virus serotype (40, 75).These so-called topotypes are an important feature ofFMDV as they may have significantly different antigeniccharacteristics which could impact on vaccine efficacy (Fig. 7) (60).

Virus structure and antigenic sitesFoot and mouth disease virus is a small non-envelopedpositive-stranded RNA virus belonging to thePicornaviridae family. The single-stranded RNA iscomprised of a large open reading frame (ORF) encoding asingle polypeptide which undergoes proteolytic cleavage toform non-structural proteins involved in virus replicationand four structural proteins (VP1, VP2, VP3, VP4) which


are incorporated into the virus capsid. VP1, VP2 and VP3are exposed on the viral surface and carry major antigenicsites. An important cell attachment site with a conservedstructure is located between variable regions on the highlyimmunogenic loop of VP1 which protrudes from thecapsid surface. This region is capable of elicitingneutralising antibody and its variable nature leads to bothintra- and inter-typic antigenic variation (73). Some otherepitopes (or antigenic sites) are dependent on the tertiarystructure of the virus particle (41) and are only present inthe intact virus known as the 146S particle (named on thebasis of its sedimentation coefficient). Additionally,different FMDV types are able to attach to different celltypes using a range of cellular receptors (25) and different host species may preferentially recognisedifferent antigenic sites (1). Thus, the virus epitopesinvolved in attachment to cells and virus neutralisation are complex.

Foot and mouth disease virus vaccines and vaccine banks

Foot and mouth disease virus vaccines are generallypurified inactivated whole-virus particles combined withadjuvants (their production and use is reviewed by Doel[17, 18] and Ahl et al [2]). During the manufacturingprocess the antigenic content of the vaccine is measured asthe amount of 146S particles. Following inactivation andcombination with adjuvant, potency is measured in termsof ability to generate virus neutralising antibody, with theultimate test of efficacy being challenge infection ofvaccinated cattle with a challenge virus homologous withthe vaccine virus. While there is some data on therelationship of antigenic content, antibody responses andprotection against infection, it has not been possible todescribe these relationships for all the serotypes and


Fig. 6Location of different topotypes of SAT 1 strains of foot and mouth disease in AfricaSource: W. Vosloo

SAT 1:

Topotype I

Topotype II

Topotype III

Topotype IV

Topotype V

Topotype VI

Topotype VII

Topotype VIII

subtypes within them. There are international standardsfor potency recommended by the OIE and these relate tonormal routinely used vaccines. Vaccines are manufacturedand supplied by local laboratories around the world as wellas by multinational companies, and depending on thesource of vaccine there is more or less adherence torecommended standards. Efficacy under field conditions ishighly variable depending on quality and potency ofvaccines, strain matching tests and species infected.However, modern vaccines properly standardised arereported to be efficacious (5).

As well as vaccines designed for routine use there is arequirement for stockpiles of emergency vaccines invaccine banks maintained in disease-free countries such asthose in Western Europe, North America and Australasia(4, 24). The vaccines are stored as a safeguard againstincursions of disease against which the population willhave no immunity. Since it is not possible to predict whichserotypes may cause an outbreak, it is desirable for vaccinebanks to store a full spectrum of serotypes and subtypes torespond to any potential eventuality. These vaccines arestored as virus concentrate over liquid nitrogen and in an


100

TAN/2/99TAN/1/99TAN/60/99

TAN/19/96TAN/51/99

TAN/5/96

100

10098

94

86

70

9875

73

99

100

100

100

100

100

100

90

88

5%

VI

V I

H

G

F

E

D

C

B

A

J

K

98

100

KEN/4/98

KEN/9/91TAN/2/96TAN/1/96

MAL/1/85TAN/37/99

TAN/3/80TAN/2/77

ZAM/2/93ZIM/14/98

ZIM/7/99ZIM/5/99

KNP/196/91MOZ/3/77

SAR/9/81ZIM/3/88

ZAM/1/99BOT/37/98

BOT/2/98BOT/8/98

BOT/24/77BOT/1/68

UGA/3/99

UGA/1/97NIG/6/76NIG/25/75NIG/14/76NIG/14/75NGR/4/76NIG/20/76

NGR/2/76NIG/1/76

NIG/8/76UGA/13/74

NIG/3/80NIG/5/81

NIG/10/81NIG/2/79

SUD/3/76SUD/4/76

SUD/13/74SUD/8/74SUD/9/74

I

II

III

IV UGA/7/99

KEN/11/91

Fig. 7Phylogenetic tree of SAT 1 viruses isolated in East Africa between 1971 and 2000Source: W Vosloo

emergency are diluted to concentrations higher thannormal vaccines as the aim is to arrest spread of infectionwith a single dose. Understanding the impact of straindiversity between vaccine strains and field strains is veryimportant for predicting the likely contribution ofemergency vaccination strategies to the eradication of theinfection. There is also an important interplay betweenvaccine potency and strain diversity, as highly potentvaccines containing a heterologous strain may be aseffective in control as a well matched vaccine strain in a low potency vaccine and at present there is little data toinform governments of the best vaccines to select in a crisis.

Vaccine strainsVaccine strains are selected on the basis of a number ofcharacteristics, but good growth characteristics and theability to elicit an antibody response which is broadlycross-reactive within a subtype are the most important(17). This is a major challenge for vaccine manufacturersglobally, but particularly for the providers of vaccine bankswhich hold a range of vaccines or vaccine concentrates toenable disease-free countries to respond to incursions ofFMDV with vaccination programmes.

Epidemiology of foot and mouth disease and the use of vaccinesInactivated vaccines are in routine use in some regionswhere FMDV is endemic and the virus types included inthe vaccines reflect those which are prevalent in the region.In general the SAT1, SAT2 and SAT3 types have beenrestricted to sub-Saharan Africa, with only occasionalincursions into North Africa and the Middle East.Serotypes O, A and C have also been reported from theAfrican continent (Fig. 8). In South America there havebeen intensive efforts to eradicate FMD through avaccination policy, but type O and A viruses continue to beisolated (3), and there has been a need in recent years tomodify vaccine strains in response to a variant virus of theA serotype. In Asia there are large unmonitored reservoirsand types O, A and Asia 1 are endemic in some regions(62). In South Eastern Europe types O and A, andoccasionally Asia 1, have also been reported in recent years(61). Of particular note is the dramatic spread of type O (pan-Asian lineage), which was first reported fromNorthern India, but spread east to Taipei China and westto the Middle East and the Balkans. Eventually it wasshipped to South Africa in 2000 and reached Europe in 2001.

The ability for this virus to spread rapidly throughpopulations and to be transported in the form ofcontaminated products is a clear indication of the

importance of horizon scanning as part of a preparednesspolicy (59). It is essential to maintain an awareness ofcurrent virus types and the strains within those typeswhich are circulating and this is a major challenge giventhe diversity of strains even within one continent (Fig. 8)(76). A cornerstone of effective vaccination programmes tocontrol and eradicate the disease in endemic areas and toprevent incursions into normally disease-free areas is theuse of vaccines containing strains that are well matched tothe outbreak strains. The huge logistical problems toachieving this on a global scale are reviewed by Paton et al.(54). In some regions such as South America there are wellcoordinated surveillance programmes and vaccine strainselection systems, whereas in other regions there is littleattempt to monitor circulating strains or submit viruses tonational or reference laboratories for characterisation.

Initial characterisation and selection for vaccine matching tests

Isolates collected from around the world are submitted toOIE reference laboratories and the Food and AgricultureOrganization (FAO) World Reference Laboratory for footand mouth disease (Institute of Animal Health [IAH],Pirbright, UK) for identification, genetic analysis andserological typing. Identification is normally achieved byenzyme-linked immunosorbent assay (ELISA) with a panelof type specific antisera. Sequencing of part of the VP1gene allows comparison with other viruses already typedand submitted to the database (33). This comparison veryoften allows the origin of the outbreak strain to be locatedas virus topotypes can be identified in this way. As anexample there are at least 8 topotypes of serotype O. Thisdata can give an indication of whether the virus strainsubmitted has been isolated before or whether it is unusualand warrants vaccine matching tests, given the highmutation rate and consequent variable nature of FMDV,where possible several isolates from the same outbreak arecharacterised and submitted for vaccine matching.

Foot and mouth disease virus international surveillance and virus typing

In parallel with genetic studies cross-neutralisation testswith reference sera that have been prepared to previouslycharacterised viruses are conducted to examine the cross-reactivity between outbreak strains and the availablevaccine strains. ELISA tests are also used to examineantigenic relationships. The purpose of this exercise is toidentify the virus type and ascertain whether the isolatesare closely related to currently held vaccine strains of therelevant type or are antigenically distinct. As alreadymentioned, it is important, particularly for procurers ofvaccines, to appreciate that within a single type there may


be a wide spectrum of strains, some of which barely cross-react and therefore would not cross-protect.

Laboratory tests used internationally to characterise viruses andmatch them to vaccine strainsAs a result of the independent regional efforts to addressthe problem of vaccine strain selection and disparateapproaches used by local and multinational vaccinemanufacturers, a number of tests have been developed for

comparing field isolates with vaccine strains. These includethe calculation of R values (relationship values) fromserological cross-reaction studies using VN, complementfixation and ELISA tests to compare the reactivity ofoutbreak and vaccine strains with antisera to vaccine virus(54). Additionally, in South America this approach ofcomparing vaccine and field viruses serologically has beenrefined by using sera from vaccinated cattle which weresubsequently challenged, thus allowing a prediction ofprotection to be made based on the serological cross-reactivity (54). The major drawback with all these tests isthat there has been little standardisation or harmonisation


Fig. 8Location of foot and mouth disease outbreaks and their serotypes recorded in the African Continent

of techniques and reagents or provision of internationalreagents for standardisation. Furthermore, there is onlyvery limited data from cross-protection studies usingemerging viruses as challenge viruses against heterologousvaccine strains. Thus, interpretation of such data in termsof vaccine efficacy in a field situation remains uncertain.

Future initiativesIt is clear that with increasing international trade andtravel, FMDV has the means by which to spread rapidlyaround the world. It is essential, particularly for thedisease-free regions, to maintain effective horizon scanningso that they are prepared for the emergence of new strains.

To date, there has been no internationally coordinatedprogramme of collection and review of FMDV isolates as isconducted for equine influenza. However, the laboratory atthe IAH, Pirbright, which is the FAO World ReferenceLaboratory for foot and mouth disease, has characterisedmany viruses from around the world. Other laboratorieshave played similar roles at a regional level. It has beenrecognised that to respond to the challenges of the straindiversity of FMDV these resources need to be pooled (54).

Following the 2001 outbreak in Europe a ‘coordinatedaction’ has been funded by the European Union to enableOIE reference laboratories round the world to create anetwork of information and reagents in order to harmoniseapproaches to virus characterisation and comparison withvaccine strains. This will bring together expertise from theUK, South America, Russia and sub-Saharan Africa andprovide an opportunity for international harmonisation.The aims are to develop standardised methods of bestpractice; collect, characterise and archive viruses whichrepresent FMDVs global diversity; exchange reagents andinformation to facilitate efficient vaccine matching and toreport annually to the OIE and FAO.

Orbiviruses: bluetongue andAfrican horse sicknessStructure and variabilityBluetongue and AHS viruses are members of the Orbivirusgenus in the family Reoviridae. They are arthropod-borne(Culicoides sp.) viral diseases of ruminants and equidaerespectively (14, 74). Orbiviruses have double-strandedRNA segmented genomes and as such have the potentialfor displaying broad antigenic diversity, as evidenced bythe 24 serotypes of BT and 9 serotypes of AHS. Asexpected, the replication of the RNA genome of orbivirusesis also prone to errors due to lack of a proof-reading

polymerase. Diversity is also generated by gene segmentswapping during mixed infections (27). However, the rateof evolution in arthropod-borne viruses is lower than insingle-host pathogens such as equine influenza and it ishypothesised that it is limited by the alternating hostreplication cycles (Culicoides sp. and ruminants) whichdemand a compromise in fitness levels to enable the virusto replicate in both vertebrate and invertebrate cells (78).

The 10 genome segments code for seven structuralproteins (VP1-7) and three non-structural proteins (NS1-3). VP2 is the major component of the outer capsid and themain antigen responsible for cell attachment and virusneutralisation, although VP5, another component of thecapsid, also plays a minor role. There is some cross-protection between serotypes within each virus and this isattributed both to a degree of cross-neutralisation betweenserotypes with similar VP2 antigenic structures and also tocell-mediated immunity driven by the less variable internalantigens.

The gene segments evolve independently of one another bygenetic drift in a host-specific fashion generatingquasispecies populations in both ruminants and insects. Ithas also been shown that random mutations occurring invertebrate cells may become fixed when ingested byCulicoides sp. (9). Thus, there are many complexopportunities for genetic and antigenic diversity.

The genetic diversity of BT has been exploited forepidemiological studies. Analysis of genes coding for theconserved VP3 or the NS3 proteins can be used forgeographic typing and tracing (9, 27) whereas the VP2gene segregates strains according to serotype (8).Nevertheless, in a recent investigation of BT in theMediterranean Basin complete sequence analysis of theVP2 gene has proved very useful in identifying topotypeswithin a serotype and in tracing sources of infection (56).

Vaccines and antigenic diversityCurrently, most available vaccines for BT and AHS areclassical attenuated vaccines developed by passagingviruses in embryonated eggs (BT) or mice (AHS) and areproduced in tissue culture (74). These attenuated vaccinestrains are not without risk and their main use has been tocontrol the diseases in sub-Saharan Africa, therefore,knowledge of the impact of viral diversity on vaccineefficacy is limited. The low levels of cross-reactivitybetween serotypes have been exploited for vaccinationagainst both BT and AHS. Thus, it is not necessary toinclude all serotypes in live vaccines in order to providerelatively broad protection against a range of serotypes (14, 74). In general, the success of this strategy has beenassessed from field rather than experimental studies. Thecurrent inactivated vaccine contains serotypes 2 and 4 and


cross-protection against other serotypes has not beenreported.

The recent outbreaks of BT, serotypes 1, 2, 4, 8, 9 and 16,in the Mediterranean Basin (28) have focused attention ongenetic and antigenic diversity of BT (56) and how it mayrelate to vaccine efficacy in the field. The use of the liveattenuated BT vaccines in Europe and subsequent spreadof the vaccine virus has also revealed the potential safetyissues relating to live vaccines. The recent spread of BTserotype 8 in northern Europe (42) further focusesattention on appropriate vaccine strategies to respond tochanging the epidemiological situation in Europe.Historically there has been much research to developsubunit vaccines as alternative vaccine candidates to bothBT and AHS (58, 57) and to explore common antigensbetween serotypes. However, if inactivated vaccinestrategies are pursued, antigenic diversity within andbetween serotypes will have much greater importance.

While it is recognised that VP2 is highly variable acrossand within serotypes, it is also recognised that the VP2genes retain common regions across serotypes which mayexplain the degree of cross-reactivity observed betweensome serotypes. Similar observations have been made forAHS (55). The challenge is to assess how important theobserved diversity is in terms of neutralisation andprotection in the target species. To date, there have beenfew studies to examine this question. However, it wasobserved that there was a high homology at the molecularlevel between Italian isolates and the vaccine strain forBTV-2 (51), which was consistent with observed protectionin the field (28, 63). In contrast, there was low genetichomology between the BTV-9 isolated in Italy and thevaccine strain, although cross-protection wasdemonstrated in a challenge study (G. Savini, unpublishedfindings). Interestingly, when amino acid sequences, asopposed to nucleotide sequences, were compared therewas a higher degree of homology between the two BTV-9strains. Thus, it appears that important epitopes relating tocell attachment may have been preserved in spite of thepropensity for the virus to diversify genetically (56). Also,the observed protection may be in part due to the fact thatlive attenuated vaccines generate neutralising antibody to anumber of surface epitopes on other viral proteins as wellas elicit cell-mediated immunity.

Clearly, with the increasing importance of BT (andpotentially AHS) in the changing global climaticconditions, there is a need to increase our understanding ofvaccine efficacy against intra- and inter-typic variants ofthese viruses. This will require more cross-protectionstudies in the target species and analysis of protection inrelation to antigenic characteristics.

Summary and conclusionsThis article refers to the antigenic diversity of threedifferent types of RNA viruses and briefly reviews itspotential significance for different vaccination strategies.Although the genetic basis of virulence has not beenaddressed in this chapter it is crucial to the understandingof vaccine efficacy given that the immunity provided byvaccines can be overcome if infections are rapid withinhost or create high virus doses and spread rapidly through populations.

There are obviously many more viruses displaying similarcharacteristics which are generating intensive researchefforts to examine antigenic diversity in relation to control.The appearance of bat lyssaviruses in Europe has initiatedefforts to understand the antigenic significance of differentlineages with respect to vaccination (48). Similarly, theexplosion of infectious bursal disease infections in poultryhas created huge interest in this avian birnavirus, where itis essential to understand the relative contribution ofchanges in virulence and antigenicity to the epidemiologyof the disease (30, 71).

Ribonucleic acid viruses will remain an enormouschallenge in disease control as new variant viruses emerge.However, prospects of responding more effectively areincreasing. Collaborations between virologists,computational experts and mathematicians are opening upexciting new opportunities for monitoring viral diversityand predicting likely changes. As genome sequencingbecomes a routine and rapid technique it becomes easier totrack large numbers of viruses and assess genetic distancesbetween isolates, and, consequently, compiling largedatabases becomes possible. As genetic data accumulatesin parallel with antigenic data it is becoming possible toidentify amino acid changes which are silent and thosewhich have significant antigenic impact. Such studies arealready ongoing for influenza and where profound changesin antigenicity of the HA have been associated with singleamino acid substitutions, the causal nature of theobservations are being examined using reverse genetics.

The development of microarray-based identification ofantigenic variants of FMD virus provides prospects forspeeding up the analysis of antigenic variation among largenumbers of strains and, eventually, of vaccine strainselection (39).

To date, antigenic analysis of FMD viruses has relied onexamination of R values based on VN tests or ELISA, andanalysis of influenza has been based on the examination ofcross HI data. The development of a sophisticatedcomputational method called antigenic cartography (65)for measuring antigenic distances between strains hasprovided a step change in the way epidemiological data for


human influenza is reviewed annually and vaccine strainsselected. This approach can provide a multidimensionalimage of the antigenic distances between viruses, how theycluster and the direction of their evolution. It has greatpotential for other viruses requiring this process of reviewand selection. It can be applied to historical data ofserological reactions between viruses and sera used tocompare strains. When linked with challenge datademonstrating protection by vaccines, as is possible forequine influenza, antigenic cartography is providing realinsight into the important antigenic changes affectingcross-protection.


Les vaccins et la variabilité antigénique des virus

J.A. Mumford

RésuméLa variabilité antigénique des virus à acide ribonucléique (ARN) est le résultat dela mutation rapide qui intervient lors de la réplication et de larecombinaison/réassortiment de matériel génétique de souches apparentées,pendant une co-infection. Les souches variantes bénéficiant d’un avantagesélectif en termes de capacité de se propager ou de contourner l’immunité del’hôte s’établissent au sein des populations. Le virus de l’influenza, le virus de lafièvre aphteuse et le virus de la fièvre catarrhale du mouton sont des exemplesde virus présentant une variation antigénique. Pour être efficaces contre cesvirus, les stratégies de vaccination doivent s’accompagner de programmes desurveillance visant à détecter les sérotypes en circulation et à retracer leurévolution afin d’assurer un parfait appariement entre les souches vaccinales etles souches sauvages. Sous les auspices de l’Organisation mondiale de la santéanimale (OIE), un dispositif de sélection de souches vaccinales du virus de lagrippe équine a été mis en place, fondé sur un programme international desurveillance. Un cadre réglementaire autorise désormais la réactualisationrapide des souches vaccinales sans qu’il soit nécessaire de fournir toutes lesdonnées d’enregistrement de ces vaccins réactualisés. La fièvre aphteuse faitl’objet d’une surveillance rigoureuse partout dans le monde, recourant à lacaractérisation antigénique et génétique des isolats, mais il n’existe aucunsystème formel de sélection des souches vaccinales. Une initiative a étéentreprise à l’échelle internationale pour harmoniser les méthodes decaractérisation des virus, dans le but d’établir la base d’un futur systèmed’appariement des vaccins vis-à-vis de la fièvre aphteuse, accepté sur le planinternational et soutenu par l’OIE. En raison de l’émergence et de la propagationde la fièvre catarrhale du mouton en Europe, l’évaluation de l’innocuité et del’efficacité de vaccins contre cette maladie a été intensifiée, notamment en cequi concerne la protection croisée vis-à-vis de chaque sérotype et entresérotypes. Le principal critère pour produire des vaccins dirigés contre des virusprésentant une variabilité antigénique est de disposer d’une méthode permettantde mesurer la distance antigénique entre les souches et de mieux appréhenderles relations entre ces distances et les mécanismes de protection croisée. Unenouvelle méthode de modélisation informatique permettant de chiffrer ladistance entre souches, appelée cartographie antigénique, a été appliquée auxvirus de la grippe humaine et équine dans le but d’élucider l’évolution de ces

Clearly, success in this field will depend onmultidisciplinary teams including clinical virologists,epidemiologists, molecular biologists and mathematiciansto exploit the new opportunities available.


Vacunas y variabilidad antigénica de los virus

J.A. Mumford

ResumenLas rápidas mutaciones originadas por la replicación yrecombinación/reordenamiento de material genético de cepas afines eninfecciones simultáneas provocan la variabilidad antigénica de los virus ARN.Aquellas variantes cuya ventaja selectiva les permite propagarse, o evitar lainmunidad del huésped, se establecen en las poblaciones. Entre los virus quepresentan variabilidad antigénica pueden mencionarse los responsables de lainfluenza, la fiebre aftosa y la lengua azul. Para que la vacunación contra esosvirus sea eficaz es preciso recurrir también a programas de vigilancia de losserotipos circulantes y su evolución a fin de asegurarse de que las cepasvacunales neutralizan a los virus de campo. Se ha establecido un sistema oficialde selección de cepas vacunales contra la influenza equina, bajo los auspiciosde la Organización Mundial de Sanidad Animal (OIE), basado en un programa devigilancia internacional. Ese marco reglamentario permite actualizarrápidamente las cepas vacunales sin necesidad de presentar todos los datospara obtener la autorización de comercialización de la vacuna actualizada. Sibien la fiebre aftosa es objeto de una estrecha vigilancia en todo el mundo,caracterizándose los antígenos y genes de las muestras, aún no se dispone deun sistema oficial de selección de cepas vacunales. Con el apoyo de la OIE, seha dado inicio a una iniciativa internacional conjunta para armonizar losmétodos de caracterización de virus y echar los cimientos de un sistema decomparación de cepas vacunales contra la fiebre aftosa aceptadointernacionalmente. La aparición y propagación de la lengua azul en Europacondujeron a intensificar la evaluación de la inocuidad y eficacia de lasvacunas, en particular, la protección cruzada contra cada serotipo, y entre ellos.La condición más importante para producir vacunas contra virus que muestranvariabilidad antigénica consiste en recurrir a un método de medida de lasdistancias antigénicas entre cepas y comprender la relación entre esasdistancias y la protección cruzada. La cartografía antigénica, un nuevo métodoinformático para medir las distancias antigénicas entre cepas, se ha aplicado alos virus de la influenza humana y equina con objeto de estudiar la importanciade su evolución en relación con las cepas vacunales. Este método puedeaplicarse muy fácilmente a otros importantes agentes patógenos que presentanvariabilidad antigénica, como el virus de la fiebre aftosa.

Palabras claveCartografía antigénica – Fiebre aftosa – Influenza – Lengua azul – Protección cruzada –Selección de cepas vacunales – Serotipo – Topotipo – Variabilidad antigénica –Vigilancia.

virus par rapport aux souches vaccinales. Cette méthode est parfaitementapplicable à d’autres agents pathogènes présentant une variabilité antigénique,tels que le virus de la fièvre aphteuse.

Mots-clésCartographie antigénique – Fièvre aphteuse – Fièvre catarrhale du mouton – Grippe –Protection croisée – Sélection de souche vaccinale – Sérotype – Surveillance – Topotype– Variabilité antigénique.

References1. Aggarwal N. & Barnett P.V. (2002). – Antigenic sites of foot-

and-mouth disease virus (FMDV): an analysis of thespecificities of anti-FMDV antibodies after vaccination ofnaturally susceptible host species. J. gen. Virol., 83, 775-782.

2. Ahl R., Haas B., Lorenz R.J. & Whittmann G. (1990). –Alternative potency test of FMD vaccines and results ofcomparative antibody assays in different cell systems andELISA. In Report of the Session of the Research Group of theStanding Technical Committee of the European Commissionfor the Control of Foot-and-Mouth Disease, Lindholm,Denmark, 25-29 June. Food and Agriculture Organization,Rome, 51-60.

3. Araujo Jr J.P., Montassier H.J. & Pinto A.A. (2002). –Extensive antigenic and genetic variation among foot-and-mouth disease type A viruses isolated from the 1994 and1995 foci in Sao Paulo, Brazil. Vet. Microbiol., 84, 15-27.

4. Barnett P.V., Samuel A.R. & Statham R.J. (2001). – Thesuitability of the ‘emergency’ foot-and-mouth disease antigensheld by the international Vaccine Bank within a globalcontext. Vaccine, 19 (15-16), 2107-2117.

5. Barteling S.J. (2002). – Development and performance ofinactivated vaccines against foot and mouth disease. In Footand mouth disease: facing the new dilemmas (G.R. Thomson,ed.). Rev. sci. tech. Off. int. Epiz., 21 (3), 577-588.

6. Bastos A.D.S., Haydon D.T., Sangare O., Boshoff C.I., Edrich J.L. & Thomson G.R. (2003). – The implications ofvirus diversity within the SAT 2 serotype for control of foot-and-mouth disease in sub-Saharan Africa. J. gen. Virol., 84,1595-1606.

7. Boni M.F., Gog J.R., Andreasen V. & Feldman M.W. (2006). –Epidemic dynamics and antigenic evolution in a single seasonof influenza A. Proc. biol. Sci., 273, 1307-1316.

8. Bonneau K.R., Zhang N., Zhu J., Zhang F., Li Z., Zhang K.,Xiao L., Xiang W. & MacLachlan N.J. (1999). – Sequencecomparison of the L2 and S10 genes of bluetongue virusesfrom the United States and the People’s Republic of China.Virus Res., 61 (2), 153-160.

9. Bonneau K.R., Mullens B.A. & MacLachlan N.J. (2001). –Occurrence of genetic drift and founder effect duringquasispecies evolution of the VP2 and NS3/NS3A genes ofbluetongue virus upon passage between sheep, cattle andCulicoides sonorensis. J. Virol., 75, 8298-8305.

10. Borrego B., Novella I.S., Giralt E., Andreu D. & Domingo E.(1993). – Distinct repertoire of antigenic variants of foot-and-mouth disease virus in the presence or absence of immuneselection. J. Virol., 67 (10), 6071-6079.

11. Brown I.H. (2000). – The epidemiology and evolution ofinfluenza viruses in pigs. Vet. Microbiol., 74 (1-2), 29-46.

12. Burrows R., Denyer M., Goodridge D. & Hamilton F. (1981).– Field and laboratory studies of equine influenza virusesisolated in 1979. Vet. Rec., 109, 353-356.

13. Burrows R. & Denyer M. (1982). – Antigenic properties ofsome equine influenza viruses. Arch. Virol., 73, 15-24.

14. Coetzer J.A.W. & Guthrie A.J. (2004). – African horsesickness. In Infectious diseases of livestock, 2nd Ed. (J.A.W.Coetzer & R.C. Tustin, eds). Oxford University Press, 1231-1246.

15. Daly J.M., Lai A.C., Binns M.M., Chambers T.M.,Barrandeguy M. & Mumford J.A. (1996). – Antigenic andgenetic evolution of equine H3N8 influenza A viruses. J. gen.Virol., 77, 661-671.

16. Daly J.M., Yates P.J., Newton J.R., Park A., Henley W., Wood J.L.N., Davis-Poynter N. & Mumford J.A. (2004). –Evidence supporting the inclusion of strains from each of thetwo co-circulating lineages of H3N8 equine influenza virus invaccines. Vaccine, 22 (29-30), 4101-4109.

17. Doel T.R. (2003). – FMD vaccines. Virus Res., 91, 81-99.

18. Doel T.R. (2005). – Natural and vaccine induced immunity toFMD. Curr. Top. Microbiol. Immunol, 288, 103-131.

19. Domingo E., Diez J., Martinez M.A., Hernandez J., Holguin A., Borrego B. & Mateu M.G. (1993). – Newobservations on antigenic diversification of RNA viruses.Antigenic variation is not dependent on immune selection. J. gen. Virol., 74 (10), 2039-2045.

20. Domingo E. & Holland J.J. (1997). – RNA virus mutationsand fitness for survival. Annu. Rev. Microbiol., 51, 151-178.

21. Domingo E., Baranowski E., Escarmis C. & Sobrino F.(2002). – Foot-and-mouth disease virus. Comp. Immunol.Microbiol. infect. Dis., 25 (5-6), 297-308.

22. European Agency for the Evaluation of Medicinal Products(EMEA) (1998). – EMEA/CVMP/112/98 Note for Guidance –Harmonisation of requirements for equine influenza vaccines:specific requirements for substitution or addition of a strainor strains. Available at: http://www.emea.eu.int/pdfs/vet/regaffair/011298en.pdf (accessed on 12 February 2007).

23. European Pharmacopoeia (1998). – Monograph on EquineInfluenza Vaccines. 3rd Ed, Council of europe, Strasbourg.

24. Forman A.J. & Garland A.J.M. (2002). – Foot and mouthdisease: the future of vaccine banks. In Foot and mouthdisease: facing the new dilemmas (G.R. Thomson, ed.). Rev.sci. tech. Off. int. Epiz., 21 (3), 601-612.

25. Fry E.E., Newman J.W.I., Curry S., Najjam S., Jackson T.,Blakemore W., Lea S.M., Miller L., Burman A., King A.M.Q.& Stuart D.I. (2005). – Structure of foot-and mouth diseasevirus serotype A1061 alone and complexed witholigosaccharide receptor: receptor conservation in the face ofantigenic variation. J. gen. Virol., 86, 1909-1920.


26. Gould A.R. (2004). – Virus evolution: disease emergence andspread. Aust. J. experim. Agric., 44 (11), 1085-1094.

27. Grenfell B.T., Pybus O.G., Gog J.R., Wood J.L.N., Daly J.D.,Mumford J.A. & Holmes E.C. (2004). – Review: unifying theepidemiological and evolutionary dynamics of pathogens.Science, 303, 327-332.

28. Hammoumi S., Breard E., Sailleau C., Russo P., Grillet C.,Cetre-Sossah C., Albina E., Sanchis R., Pepin M., Guibert J.M.& Zientara S. (2003). – Studies on the safety andimmunogenicity of the South African bluetongue virusserotype 2 monovalent vaccine: specific detection of thevaccine strain genome by RT-PCR. J. vet. Med., 50, 316.

29. Haydon D.T., Samuel A.R. & Knowles N.J. (2001). – Thegeneration and persistence of genetic variation in foot-and-mouth disease virus. Prev. vet. Med., 51 (1-2), 111-124.

30. Heine H.G., Haritou M., Pailla P., Fahey K. & Azad A. (1991).– Sequence analysis and expression of the host-protectiveimmunogen VP2 of a variant strain of infectious bursaldisease virus which can circumvent vaccination with standardtype 1 strains. J. gen. Virol., 72, 1835-1843.

31. Hinshaw V.S., Naeve C.W., Webster R.G., Douglas A., Skehel J.J. & Bryans J. (1983). – Analysis of antigenicvariation in equine 2 influenza A viruses. Bull. WHO, 61 (1),153-158.

32. Kawaoka Y., Bean W.J. & Webster R.G. (1989). – Evolution ofthe hemagglutinin of equine H3 influenza viruses. Virology,169, 283-292.

33. Knowles N.J. & Samuel A.R. (2003). – Molecularepidemiology of foot-and-mouth disease virus. Virus Res., 91 (1), 65-80.

34. Lai A.C., Rogers K.M., Glaser A., Tudor L. & Chambers T.(2004). – Alternate circulation of recent equine-2 influenzaviruses (H3N8) from two distinct lineages in the UnitedStates. Virus Res., 100 (2), 159-164.

35. Lee C.W., Senne D.A. & Suarex D.L. (2004). – Effect ofvaccine use in the evolution of Mexican lineage H5N2 avianinfluenza virus. J. Virol., 78, 8372-8381.

36. Livesay G.J., O’Neill T., Hannant D., Yadav M.P. & Mumford J.A. (1993). – The outbreak of equine influenza(H3N8) in the United Kingdom in 1989: diagnostic use of anantigen capture ELISA. Vet. Rec., 133 (21), 515-519.

37. Marozin S., Gregory V., Cameron K., Bennett M., Valette M.,Aymard M., Foni F., Barigazzi G., Lin Y. & Hay A. (2002). –Antigenic and genetic diversity among swine influenza A H1N1 and H1N2 viruses in Europe. J. gen. Virol., 83 (4), 735-745.

38. Martin M.J., Nunez J.I., Sobrino F. & Dopazo J. (1998). – A procedure for detecting selection in highly variable viralgenomes: evidence of positive selection in antigenic regionsof capsid protein VP1 of foot-and-mouth disease virus. J. virol.Meth., 74 (2), 215-221.

39. Martin V., Perales C., Abia D., Ortiz A.R., Domingo E. &Briones C. (2006). – Microarray-based identification ofantigenic variants of foot-and-mouth disease virus: abioinformatics quality assessment. BMC Genomics, 7, 117.

40. Martinez M.A., Dopazo J., Hernandez J., Mateu M.G.,Sobrino F., Domingo E. & Knowles N.J. (1992). – Evolutionof the capsid protein genes of foot-and-mouth disease virus:antigenic variation without accumulation of amino acidsubstitutions over six decades. J. Virol., 66 (6), 3557-3565.

41. Mateu M.G., Hernandez J., Martinez M.A., Feigelstock D.,Lea S., Perez J.J., Giralt E., Stuart D., Palma E.L. & Domingo E. (1994). – Antigenic heterogeneity of a foot-and-mouth disease virus serotype in the field is mediated by verylimited sequence variation at several antigenic sites. J. Virol.,68 (3), 1407-1417.

42. Mellor P. (2006). – Bluetongue in sheep in the Netherlands.Preliminary outbreak assessment, 18 August 2006.Information provided by the Bluetongue OIE ReferenceLaboratory (Institute of Animal Health, Pirbright, UnitedKingdom) to the International Animal Health Division,Department for Environment, Food and Rural Affairs,London. Available at: http://www.defra.gov.uk/animalh/diseases/monitoring/pdf/bt-nlsheep180806.pdf.

43. Mumford J.A. (1998). – Control of influenza from aninternational perspective. Equine infectious diseases, VIII. R & W Ltd, Newmarket, 11-24.

44. Mumford J.A. & Wood J. (1992). – Establishing anacceptability threshold for equine influenza vaccines. Dev. biol. Standard., 79, 137-146.

45. Mumford J. & Wood J. (1993). – WHO/OIE meeting:consultation on newly emerging strains of equine influenza.18-19 May 1992, Animal Health Trust, Newmarket, Suffolk,United Kingdom. Vaccine, 11 (11), 1172-1175.

46. Mumford J.A., Wilson H., Hannant D. & Jessett D.M. (1994).– Antigenicity and immunogenicity of equine influenzavaccines containing a Carbomer adjuvant. Epidemiol. Infect.,112, 421-437.

47. Mumford J.A., Jessett D.M., Rollinson E.A., Hannant D. &Draper M.E. (1994). – Duration of protective efficacy ofequine influenza immunostimulating complex/tetanusvaccines. Vet. Rec. 134, 158-162.

48. Nel L.H. (2005). – Vaccines for lyssaviruses other than rabies.Expert Rev. Vaccines, 4 (4), 533-540.

49. Newton J.R., Verheyen K., Wood J.L.N., Yates P.J., Mumford J.A. (1999). – Equine influenza in the UnitedKingdom in 1998. Vet. Rec., 145 (16), 449-452.

50. Newton J.R., Townsend H.G.G., Wood J.L.N., Sinclair R.,Hannant D., Mumford J.A. (2000). – Immunity to equineinfluenza: relationship of vaccine-induced antibody in youngthoroughbred racehorses to protection against field infectionwith influenza A/equine-2 viruses (H3N8). Equine vet. J., 32 (1), 65-74.


51. Orru G., Santis P.D., Solinas F., Savini G., Piras V. & Caporale V. (2004). – Differentiation of Italian field and SouthAfrican vaccine strains of bluetongue virus serotype 2 usingreal-time PCR. J. virol. Meth., 122 (1), 37-43.

52. Oxburgh L., Berg M., Klingeborn B., Emmoth E. & Linne T.(1994). – Evolution of H3N8 equine influenza virus from1963 to 1991. Virus Res., 34 (2), 153-165.

53. Park A.W., Wood J.L.N., Daly J.M., Newton J.R., Glass K.,Henley W., Mumford J.A. & Grenfell B.T. (2004). – Theeffects of strain heterology on the epidemiology of equineinfluenza in a vaccinated population. Proc. Biol. Sci., 271,1547-1555.

54. Paton D.J., Valarcher J.-F., Bergmann I., Matlho O.G.,Zakharov V.M., Palma E.L. & Thomson G.R. (2005). –Selection of foot and mouth disease vaccine strains – a review.Rev. sci. tech. Off. int. Epiz., 24 (3), 981-993.

55. Potgieter A.C., Cloete M., Pretorius P.J. & van Dijk A.A.(2003). – A first full outer capsid protein sequence data-set inthe Orbivirus genus (family Reoviridae): cloning, sequencing,expression and analysis of a complete set of full-length outercapsid VP2 genes of the nine African horsesickness virusserotypes. J. gen. Virol., 84, 1317-1326.

56. Potgieter A.C., Monaco F., Mangana O., Nomikou K., Yadin H. & Savini G. (2005). – VP2-segment sequenceanalysis of some isolates of bluetongue virus recovered in theMediterranean Basin during the 1998-2003 outbreak. J. vet.Med., 52, 372-379.

57. Roy P. (1996). – Multiple gene expression in baculovirussystem. Third generation vaccines for bluetongue disease and African horsesickness disease. Ann. N.Y. Acad. Sci., 791, 318-332.

58. Roy P., Bishop D.H., Howard S., Aitchison H. & Erasmus B.(1996). – Recombinant baculovirus-synthesized Africanhorsesickness virus (AHSV) outer-capsid protein VP2provides protection against virulent AHSV challenge. J. gen.Virol., 77 (9), 2053-2057.

59. Royal Society (2002). – Royal Society inquiry into infectiousdiseases in livestock. Summary and main recommendations.Policy Doc. 19/02. Available at: http:// www.royalsoc.ac.uk/inquiry/index/idl_sum.pdf (accessed on 12 February 2007).

60. Sahle M., Dwarka R.M., Venter E.H. & Vosloo W. (2006). –Comparison of SAT-1 foot-and-mouth disease virus isolatesobtained from East Africa between 1971 and 2000 withviruses from the rest of sub-Saharan Africa. Arch. Virol., 28 December 2006 (E-pub ahead of print).

61. Samuel A.R., Knowles N.J. & MacKay D.K.J. (1999). –Genetic analysis of type O viruses responsible for epidemicsof foot-and-mouth disease in North Africa. Epidemiol. Infect.,122, 529-538.

62. Sanyal A., Gurumurthy C.B., Venkataramanan R., HemadriD. & Tosh C. (2003). – Antigenic characterisation of foot-and-mouth disease virus serotype Asia 1 field isolates usingpolyclonal and monoclonal antibodies. Vet. Microbiol., 93 (1),1-11.

63. Savini G., Goffredo M., Monaco F., De Santis P. & Meiswinkel R. (2003). – Transmission of bluetongue virus inItaly. Vet. Rec., 152 (4), 119.

64. Sevilla N., Verdaguer N. & Domingo E. (1996). –Antigenically profound amino acid substitutions occurduring large population passages of foot-and-mouth diseasevirus. Virology, 225 (2), 400-405.

65. Smith D.J., Lapedes A.S., de Jong J.C., Bestebroer T.M.,Rimmelzwaan G.F., Osterhaus A.D. & Fouchier R.A. (2004).– Mapping the antigenic and genetic evolution of influenzavirus. Science, 305 (5682), 371-376.

66. Steinhauer D., Domingo E. & Holland J.J. (1992). – Lack ofevidence for proofreading mechanisms associated with anRNA virus polymerase. Gene, 122, 281-288.

67. Stohr K. (2003). – The global agenda on influenzasurveillance and control. Vaccine, 21 (16), 1744-1748.

68. Thomson G.R., Mumford J.A., Spooner P.R., Burrows R. &Powell D.G. (1977). – The outbreak of equine influenza inEngland: January 1976. Vet. Rec., 100, 465-468.

69. Thomson G.R. & Bastos A.D.S. (2004). – Foot-and-mouthdisease. In Infectious diseases of livestock, 2nd Ed. (J.A.W.Coetzer & R.C. Tustin, eds). Oxford University Press, 1323-1365.

70. United States Department of Agriculture (USDA) (2006). –USDA Veterinary Services Memorandum No. Draft-017. Viralstrain changes in equine influenza and swine influenzavaccines. Available at: www.aphis.usda.gov/vs/cvb/memos/VSMemoDraftD017.pdf (accessed on 12 February 2007).

71. Van den Berg T. (2000). – Acute infectious bursal disease inpoultry: a review. Avian Pathol., 29, 175-194.

72. Van Reeth K., Labarque G., De Clercq S. & Pensaert M.(2001). – Efficacy of vaccination of pigs with different H1N1swine influenza viruses using a recent challenge strain anddifferent parameters of protection. Vaccine, 19 (31), 4479-4486.

73. Verdaguer N., Sevilla N., Valero M.L., Stuart D., Brocchi E.,Andreu D., Giralt E., Domingo E., Mateu M.G. & Fita I.(1998). – A similar pattern of interaction for differentantibodies with a major antigenic site of foot-and-mouthdisease virus: implications for intratypic antigenic variation. J. Virol., 72 (1) 739-748.

74. Verwoerd D.W. & Erasmus B.J. (2004). – BluetongueChapter. In Infectious diseases of livestock, 2nd Ed. (J.A.W.Coetzer & R.C. Tustin eds.). Oxford University Press, 1201-1220.

75. Vosloo W., Bastos A.D., Kirkbridge E., Esterhuysen J.J., van Rensburg D.J., Bengis R.G., Keet D.W. & Thomson G.R.(1996). – Persistent infection of African buffalo (Synceruscaffer) with SAT-type foot-and-mouth disease viruses: rate offixation of mutations, antigenic change and interspeciestransmission. J. gen. Virol., 77 (7), 1457-1467.


76. Vosloo W., Bastos A.D.S., Sangare O., Hargreaves S.K. &Thomson G.R. (2002). – Review of the status and control offoot and mouth disease in sub-Saharan Africa. In Foot andmouth disease: facing the new dilemmas (G.R. Thomson,ed.). Rev. sci. tech. Off. int. Epiz., 21 (3), 437-449.

77. Vosloo W., Dwarka R.M., Bastos A.D.S., Esterhuysen J.J.,Sahle M. & Sangare O. (2004). – Molecular epidemiologicalstudies of foot-and-mouth disease virus in sub-Saharan Africaindicate the presence of large numbers of topotypes:implications for local and international control. Appendix 22 of the Report of the Session of the Research Group of theStanding Technical Committee of the European Commissionfor the Control of Foot-and-Mouth Disease, 11-15 October,Chania, Crete, Greece. Food and Agriculture Organization,Rome, 149-158.

78. Weaver S.C., Brault A.C., Kang W. & Holland J.J. (1999). –Genetic and fitness changes accompanying adaptation of anarbovirus to vertebrate and invertebrate cells. J. Virol., 73 (5),4316-4326.

79. Wood J.M., Mumford J.A., Folkers C., Scott A.M. & Schild G.C. (1983). – Studies with inactivated equineinfluenza vaccine 1. Serological responses of ponies to gradeddoses of vaccines. J. Hyg. (Camb.), 90, 371-394.

80. Wood J.M., Schild G.C., Folkers C., Mumford J.A. &Newman R.W. (1983). – The standardisation of inactivatedequine influenza vaccines by single-radial immunodiffusion.J. biol. Standard., 11, 133-136.

81. World Organisation for Animal Health (OIE) (2004). –Manual of Diagnostic Tests and Vaccines for TerrestrialAnimals, Chapter 2.5.5, Equine influenza, 5th Ed. OIE, Paris.

82. Yates P. & Mumford J.A. (2000). – Equine influenza vaccineefficacy: the significance of antigenic variation. Vet. Microbiol.,74 (1-2), 173-177.


vaccines and viral antigenic diversity

Documents