Vaccine 23 (2005) 5186–5195

The application of new techniques to the improved detection ofpersistently infected cattle after vaccination and contact

exposure to foot-and-mouth disease

S. Paridaa, S.J. Coxa, S.M. Reida, P. Hamblina, P.V. Barnetta, T. Inoueb,J. Andersona, D.J. Patona,∗

a Institute for Animal Health, Pirbright Laboratory, Ash Road, Pirbright, Surrey GU24 0NF, UKb National Institute of Animal Health, Kodaira, Tokyo, Japan

Received 30 November 2004; accepted 1 June 2005Available online 5 July 2005

Abstract

ventionals l contactc vaccinatedc osureT ersistence,t etected moref s isolation.A declinedo had beenc al tests tod test couldi st evidenceo nt FMDVi©

K

1

Ahmaa1

themA,ake

ateandare

ame,

froms buttinu-e with

0d

Detection of antibodies to the non-structural proteins (NSP) of foot-and-mouth disease virus (FMDV) was compared with conerological and virological methods and with RT-PCR for the identification of FMDV carrier animals obtained after experimentahallenge of vaccinated cattle. Transmission from carriers to sentinels was also monitored. Twenty FMDV vaccinated and five unattle were challenged by direct contact with five donor cattle excreting FMDV and monitored until 28 days post challenge-exp[1].welve vaccinated and three unvaccinated animals were retained up to 24 weeks post exposure to FMDV in order to monitor viral pransmission and antibody responses. In nine vaccinated animals, infection persisted beyond 28 days post exposure, virus being drequently and for longer in oesophagopharyngeal samples from these animals when examined by RT-PCR rather than by virulthough recovery of FMDV RNA became increasingly sporadic over time, the number of RNA copies detected in positive samplesnly slowly. Two naıve sentinel cattle housed with the persistently infected animals between 93 and 168 days after the latterhallenge-exposed to FMDV did not become infected. There were differences in the ability of commercially available serologicetect antibodies to FMDV non-structural proteins (NSP) in vaccinated and subsequently challenged cattle. Although no single

dentify all of the vaccinated cattle that became persistently infected, the most poorly recognised animals were those with the leaf virus replication based on other tests. The potential of the detection of antibodies to the 2B NSP of FMDV for diagnosing persiste

nfection was demonstrated.2005 Elsevier Ltd. All rights reserved.

eywords: Foot-and-mouth disease virus; Virus persistence; Vaccine differentiation; DIVA

. Introduction

Foot-and-mouth disease virus (FMDV) is in the genusphthovirus, family Picornaviridae and is the cause of aighly contagious vesicular disease of cloven hoofed ani-als. There are seven serotypes of the virus, which is smallnd unenveloped and has a positive sense RNA genome withlarge open reading frame encoding 12 viral proteins, L,

A, 1B, 1C, 1D, 2A, 2B, 2C, 3A, 3B, 3C, and 3D. These

∗ Corresponding author. Tel.: +44 1483 231012; fax: +44 1483 232621.E-mail address: david.paton@bbsrc.ac.uk (D.J. Paton).

are translated as a single polyprotein such that many ofalso exist in combinations, like 3ABC. Viral proteins 11B, 1C, and 1D are the structural proteins (SPs) that mup the shell of the virion. The other viral proteins participin replicatory and other functions within the host cellare mostly either not part of the virion structure or elsepresent in only trace amounts; hence their collective nnon-structural proteins, or NSPs.

Foot-and-mouth disease (FMD) has been eradicatedEurope, Australasia, Japan and parts of the Americaremains endemic in many other countries causing coning economic losses. The disease runs an acute cours

264-410X/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2005.06.012

S. Parida et al. / Vaccine 23 (2005) 5186–5195 5187

generally rapid onset of lesions and the early developmentof virus-neutralising antibodies whose appearance coincideswith viral clearance from the circulation. The virus has apredilection to replicate in epithelial cells including thoselining the distal oropharynx and the dorsal surface of the softpalate[2]. In these sites, infectious virus has been found forseveral months or even years in a proportion of recoveredruminants but not pigs[3]. Persistent infection with FMDis defined as the carriage of live virus beyond 28 days postinfection. Killed vaccines are available to provide clinicalprotection against FMD and to reduce virus transmission[4].However, properly vaccinated animals may still be infectedand may become persistently infected. Carrier animals per-sistently infected with FMDV can be identified by isolatingvirus in cell cultures from oesophago-pharyngeal (O-P) flu-ids collected with a probang sampling cup. RT-PCR providesan alternative approach for virus detection in O-P fluids, butthere are few publications that systematically compare thesensitivity of this method to that of virus isolation in highlysensitive bovine thyroid cell cultures[2].

In Europe and some other parts of the world, there isa desire to be less reliant on stamping out for the controlof future FMD incursions and instead to be able to vac-cinate animals and then to identify and remove those thathave been infected and especially any that are virus carri-ers. FMD serology has been used as part of FMD controls cteda iousv Ps ofF ac-c ti-SPa ed tos vac-c tog NSPt revi-o veb hicha ialr ea ani-m thes stenti andf s ofa ed ino imilart mi-h ntactb ,t s off withc withR ndt alsom

2. Materials and methods

2.1. Experimental animals

Full details of the vaccination and challenge of the exper-imental cattle, including virological and serological findingsup to 28 days post challenge (28 dpc) are to be found in Coxet al.[1]. All of the cattle were castrated males and weighedapproximately 130 kg at the start of the experiment. In brief,20 cattle were vaccinated with oil adjuvanted O Manisa vac-cine, previously shown to have a potency of 18PD50, fromthe UK FMDV antigen reserve. This vaccine antigen was ofrecent manufacture and complied with the latest requirementsfor freedom from NSPs. Five cattle acted as unvaccinatedcontrols and another five as FMDV donors. The donor cat-tle were infected with O UKG 34/2001 by tongue inoculationand placed in contact with the other 25 recipient animals for 5days, starting 1 day after the time of inoculation. Monitoringup to 28 dpc revealed that all of the unvaccinated recipient ani-mals developed FMD, whereas none of the vaccinees showedany signs of the disease. Nevertheless, virus replication wasdetected in some of the vaccinated cattle and nine animalswere scored as persistently infected based on the presence ofinfectious FMDV and/or FMDV genome in the oropharynxat or beyond 28 dpc.

At 28 dpc, twelve vaccinated and three unvaccinated cattlew pc;t ask oser f thet c. At3 on ins es inf cteda aa to acta l onew tinelt reo andO eklybr andv

2

Rw ce ofv al.[

2

asec OIE

chemes to help to identify and remove previously infenimals some of which may still harbour persisting infectirus. In the absence of vaccination, antibodies to the SMDV are indicative of past infection. However, use of vination complicates matters, since this can also elicit anntibodies. Therefore NSP antibody tests have been uspecifically detect antibodies due to infection in alreadyinated animals[5]. There is a shortage of suitable seraenerate reliable information on the effectiveness of

ests at detection of persistently infected animals in pusly vaccinated populations[6]. Experimental sera haeen mostly derived from vaccine potency tests in wnimals were challenged with live FMDV by an artificoute involving injection into the tongue[7,8]. Field sera arvailable in larger numbers and from naturally infectedals[9–14], but it is usually impossible to be certain of

tatus of the sampled population with respect to persinfection. The opportunity was therefore taken to retainollow up on the virological status and antibody responsegroup of cattle that had been vaccinated and challengrder to assess vaccine potency in circumstances more s

o the situation that could occur in the field, using a seeterologous virus challenge administered by direct coetween vaccinated and infected cattle[1]. In these animals

he detection of antibodies to the non-structural proteinoot-and-mouth disease virus (FMDV) was comparedonventional serological and virological methods andT-PCR for the identification of FMDV carrier animals a

he transmission of virus from carriers to sentinels wasonitored.

ere retained for long-term follow-up and kept until 168 dhe remaining animals were killed. One animal UV5 willed at 126 dpc due to a problem of recurrent bloat. Of thetained, nine of the twelve vaccinated cattle and one ohree non-vaccinees were scored as infected at 28 dp5 dpc, the animals had to be moved to accommodatimaller units and were split and housed in loose boxour groups of three or four animals. The persistently infenimals were in two groups and at 93 dpc a new FMD nıvege-matched steer was added to each of these penss a sentinel for continuing virus transmission. Sentineas housed with animals UV5, 9, 10, 11 and 13 and sen

wo with animals UV14, 17 and 19. All of the cattle webserved daily and a range of samples including blood-P fluid (probang fluid) was collected on at least a weasis. Sera and probang fluids were stored at−20 and−70◦C,espectively, prior to analysis by a range of serologicalirological tests.

.2. Virological tests

Virus isolation with confirmatory ELISA and RT-PCere used to examine probang samples for the presenirus and quantity of viral RNA, as described in Cox et1].

.3. Serological tests

The virus neutralisation test (VNT) and the solid phompetition ELISA (SPCE) were as described in the

5188 S. Parida et al. / Vaccine 23 (2005) 5186–5195

Diagnostic Manual[14]. The cut-off for the SPCE was at60% inhibition. The VNT was carried out in duplicate tocompare antibody titres against the O Manisa and O UKGviruses.

Serum antibodies to FMDV NSPs were measured by fourdifferent ELISAs, three of which are commercially avail-able and one is still under development. The commercialtests used were the Ceditest FMDV-NS (Cedi-Diagnostics),the FMDV NSP ELISA (United Biochemical Incorporated,UBI) and the CHEKIT-FMD-3ABC (Bommeli). Each wasused according to the manufacturer’s recommendations. TheCeditest measures competition between serum antibodies anda NSP 3B-specific monoclonal antibody for binding to the3ABC NSP of FMDV expressed by insect cells infectedwith a recombinant baculovirus[15,16]. The other two testsare indirect ELISAs that measure the binding of antibod-ies to a synthetic 3B peptide (UBI)[17–19] or a 3ABCNSP expressed in a recombinantE. coli (Bommeli) [20].The manufacturer’s recommended cut-off points were used,except that sera scoring inconclusive positive in the Bommelitest were considered positive for the purpose of test com-parisons. The other two tests do not have an inconclusiverange.

The fourth NSP test was an indirect ELISA measur-ing antibodies to a keyhole limpet haemocyanin conjugatedsynthetic 2B peptide. F96 Maxisorp Nunc Immuno platesw ft daya with5o rvel,1 erel r.A ox-i utiona s,t ddi-t ined /W)h ac-t pti-c d a4 ab-l lust ctedf intsb xpo-s

3

3

didn

3.2. Virological tests

The results of virus isolation and RT-PCR up to andbeyond 28 dpc are summarised inFig. 1. Additional probangsamples were collected and examined at 2, 12 and 16 dpcand the results of these tests are given in Cox et al.[1].The results obtained after 28 dpc mostly confirmed theoriginal categorisation of the cattle into three groups. Group1 comprised those that were vaccinated and did not appear tobecome persistently infected after challenge exposure, threeof which were kept beyond 28 dpc. Group 2 comprised thenine vaccinated cattle that developed persistent infectionsand Group 3 is the unvaccinated cattle, three of which wereretained until 168 dpc. A probang sample from a Group1 animal, UV8, was found RT-PCR positive on a singleoccasion at 42 dpc even though none of the samples collectedprior to 28 dpc were RT-PCR positive. Furthermore, noinfectious virus was ever recovered from this animal. Anunvaccinated but challenge exposed animal (UV22) that hadnot been scored as persistently infected at 28 dpc was alsoscored RT-PCR positive at 42 dpc.

Two animals, UV2 and UV5, that were categorised as per-sistently infected, were only virus positive at or after 28 dpcby RT-PCR. In a third animal, UV14, virus was only detectedon a single occasion after 28 dpc, in this case by virus iso-lation but not RT-PCR. Nineteen of 180 probang samplesc vac-c tion,w thefi sameg RT-Ps eacht rusw cteda adica t oft ve ats sitiveb ctedb col-l4 em thel nlys

3

xi-m aterr ultsn in),t cci-n ns).T by

ere directly coated overnight at +4◦C with 250 ng/ml ohe 2B peptide in carbonate/bicarbonate buffer. Nextfter three washes with PBS, plates were pre blocked% marvel and 1% Tween 20 in PBS for 1 h at 37◦Cn an orbital shaker. Sera prediluted 1:10 in 1% ma% tween 20 and 1% normal horse serum in PBS w

oaded and incubated for 1 h at 37◦C on an orbital shakefter washing three times with PBS, anti-bovine IgG per

dase conjugate (Sigma,UK) was added at a 1:5000 dilnd again incubated for 1 h at 37◦C. After three washe

he colour reaction was developed for 10 min by aion of the chromogen, 5.05 mM ortho-phenylene-diamihydrochloride (Sigma, UK) and substrate, 30% (Wydrogen peroxide (Sigma, UK) diluted 1:2000. The re

ion was stopped by using 1 M sulphuric acid. The oal density was recorded using an ELISA reader an92 mm filter. The cut-off point for the test was est

ished by calculating the mean optical density (OD) pwo standard deviations for test results from sera collerom the 25 cattle of this experiment at seven time poefore and after vaccination but prior to challenge eure.

. Results

.1. Sentinel animals

The two sentinel cattle showed no signs of FMD andot seroconvert to FMDV by any test.

ollected from the nine, Group 2, persistently infectedinees after 28 dpc were scored positive by virus isolahereas 56 samples were positive by RT-PCR. Duringrst 28 dpc the numbers of probang samples from theroup of animals found positive by virus isolation andCR were more similar at 54 and 67, respectively.Fig. 2hows the proportion of carrier animals detected atime point by virus isolation compared to RT-PCR. Vias isolated from seven of the nine persistently infenimals at or after 35 dpc. Isolation was rather spornd was not possible after 98 dpc. By RT-PCR, eigh

he persistently infected vaccinees were scored positiome time after 35 dpc and five animals were scored poetween 147 and 168 dpc. The RNA copy number detey quantitative RT-PCR in positive probang samples

ected between 35 and 168 dpc varied between log10 2.6 and.9 copies/ml (Table 1). Although detection of RNA becamore spasmodic with increasing time after infection,

evels of RNA in positive probang samples declined olowly.

.3. Serological tests

By O Manisa VNT (using the vaccine strain), approately eight vaccinated cattle showed a four-fold or gre

ise in antibody titre between 21 dpv and 16 dpc (resot shown). By O UKG VNT (using the challenge stra

his proportion had increased to 18 out of the 20 vaated cattle (UV15 and UV3 being the only exceptiohe overall results of testing the fifteen retained cattle

S. Parida et al. / Vaccine 23 (2005) 5186–5195 5189

Fig. 1. Summary of RT-PCR and virus isolation results on oesophagopharyngeal samples collected at different time points by probang cup.

VNT using virus homologous to the challenge strain areshown for each of the three categories of animals inFig. 3.The responses of individual animals to vaccination and chal-lenge exposure were very variable. However, there was atendency for the persistently infected animals to have higherlevels of neutralising antibodies that declined more slowly.Between 28 and 168 dpc the Group 1 animals showed a titrerange of between 1 in 11 and 1 in 90, whereas for Group2, this range was 1 in 16 to 1 in 355, and excluding UV2,

Fig. 2. Comparison of percentage of Group 2 cattle scored virus positive byRT-PCR and virus isolation over time.

from which live virus was not recovered after 10 days postinfection, the latter range would have been 1 in 45 to 1 in355.

The SPCE detected seroconversion in all 20 vaccinatedanimals at between 5 and 14 dpv (mean = 7 dpv). All of theseanimals remained seropositive (>89% inhibition) until theend of the experiment. The five unvaccinated cattle sero-converted at 7–12 dpc and also remained seropositive at >92percent inhibition.

Results obtained with the 2B peptide ELISA for sera fromthe three groups of cattle are shown inFig. 4. The cut-offpoint was calculated as >0.79 for vaccinated animals (meanOD plus two standard deviations for 140 sera collected from20 vaccinated cattle prior to challenge) and >0.67 for unvac-cinated animals (mean OD plus two standard deviations for35 sera collected from five unvaccinated cattle prior to chal-lenge). Two vaccinated animals (UV6 and UV19) had highbackground levels of reactivity in the 2B test that were presentbefore vaccination. The five unvaccinated cattle (Group 3)all showed a significant rise in test reactivity following con-tact exposure to FMDV infected cattle and this surpassed thecut-off threshold by 10–14 dpc. The test reactivity remainedabove the cut-off threshold until 105 dpc in one animal anduntil the end of the experiment at 168 dpc in the other two.

5190 S. Parida et al. / Vaccine 23 (2005) 5186–5195

Table 1Quantitative RT-PCR results from probangs samples of Group 2 cattle

Days postchallenge

Log10 RNA copies/ml from probang samples of Group 2 (persistently infected) cattle

UV2 UV14 UV5 UV10 UV13 UV11 UV19 UV9 UV17

35 dpc 3.3 3.3 3.1 4.0 2.6 3.3 4.642 dpc 3.6 4.0 4.0 4.649 dpc 3.5 3.3 3.5 3.7 4.3 3.8 4.256 dpc 3.8 3.6 4.2 4.3 4.2 4.363 dpc 3.9 4.070 dpc 3.0 3.7 4.3 3.777 dpc 3.1 4.0 3.6 4.484 dpc 3.7 3.1 3.991 dpc 4.0 4.298 dpc 3.9 4.9

105 dpc 2.9112 dpc 3.9119 dpc 4.2126 dpc 3.6 3.8 3.4133 dpc 4.1140 dpc147 dpc 3.8 2.9 4.1154 dpc 3.3161 dpc 3.3 3.4168 dpc 3.3

Blank spaces: no viral RNA detected.

Fig. 3. Reciprocal titres of virus neutralising antibody over time: (a) Group 1, vaccinees with no or transient post-exposure infection; (b) and (c) Group 2,vaccinees with persistent infection; (d) Group 3, non-vaccinees; (e) means of each Group.

S. Parida et al. / Vaccine 23 (2005) 5186–5195 5191

Fig. 4. Seroconversion in 2B peptide ELISA, measured as optical density reaction over time for individual animals: (a) Group 1, vaccinees with no or transientpost-exposure infection (data not shown for UV4, 7, 12, 15, 16 and 21, in which OD at all time points was <0.6); (b) and (c) Group 2, vaccinees with persistentinfection; (d) Group 3, non-vaccinees.

Apart from a few sporadic positive results and those fromanimal UV6, the cattle in Group 1 remained seronegative inthe test throughout the period of monitoring. In contrast, fiveof the Group 2 cattle showed a clear seroconversion in the test

between 14 and 42 dpc with antibody levels maintained abovethe cut-off threshold until 105–168 dpc. One animal, UV19showed a high background reaction from the outset and twoof the three others (UV10 and UV14) showed increased test

Table 2Days post vaccination or challenge at which serum samples scored positive by different NSP ELISAs

Cedi Bommeli UBI 2B

Group 1 UV3 – – – –UV4 – – – –UV6 – – – −4–4 dpv, 10–16 dpcUV7 – – – –UV8 42 dpc – – 42 dpcUV15 – – – –UV12a – – – –UV16a – – – –UV18a – – – 14 dpvUV20a 16 dpc – – –UV21a – – – –

Group 2 UV2 – – – –UV5b 28–126 dpc – – –UV9 28–168 dpc 63–168 dpc 63–91 dpc 14–168 dpcUV10 28–168 dpc – – –UV11 14–168 dpc 49–56, 126–133, 161–168 dpc 49–105 dpc 28–168 dpcUV13 14–168 dpc 91 dpc 35–84 dpc 42–63, 91–105 dpcUV14 – – – –UV17 12, 16–168 dpc 28–56, 161–168 dpc 49–70 dpc 10–147, 161–168 dpcUV19 10–168 dpc 56 dpc 56–84 dpc 3, 14, 21 dpv, 2, 12–63 dpc

G dpc

4, 168dpc8 dpc

–

roup 3 UV22 10–168 dpc 12–70UV23a 28 dpcUV24 12–21, 35–168 dpc 12–15UV25a 14–28 dpc 16–28UV26 12–168 dpc 10–16

, No antibody detected.a Animal not retained after 28 dpc.b Killed at 126 dpc.

12–49 dpc 12–98 dpc12–28 dpc

dpc 12–126 dpc 10–168 dpc16–28 dpc 12–28 dpc12–133 dpc 14–168 dpc

5192 S. Parida et al. / Vaccine 23 (2005) 5186–5195

Fig. 5. Detection of seroreactors by NSP ELISAs over time: (a) Group 2,vaccinees with persistent infection; (b) Group 3, non-vaccinees.

reactivities post-challenge that did not rise above the cut-offthreshold.

Summaries of the qualitative results of the different NSPserological tests are shown inTable 2andFig. 5. Sera fromthe original group[1] of eleven vaccinated cattle that did notbecome persistently infected (Group 1) were largely unreac-tive in NSP tests up to 28 dpc. The same was true for the threeanimals retained thereafter. Two animals were scored posi-tive by the Cedi test on isolated occasions, UV8 at 42 dpcand UV20 at 14 dpc. Two of the persistently infected vac-cinated cattle (Group 2, UV2 and UV14) remained consis-tently seronegative by all NSP tests up to 168 dpc. The otherseven persistently infected cattle were scored positive by theCedi test from 14 to 28 days onwards, until the end of theexperiment. The UBI and Bommeli tests scored sera fromthese animals positive on a more sporadic basis, detectingseroconversion later and for less time. Two carrier animals,UV5 and UV10 were not scored positive on any occasionby both the tests. In contrast to these results, the unvacci-nated group were scored positive by most of the tests from14 dpc onwards, although UV23 was only scored positiveat 28 dpc by Cedi test and three animals, UV22, UV24 andUV26 became seronegative by Bommeli and/or UBI tests atlater time points.

4

MDv MDo rs ofi hp -free

status for the purposes of international trade, post-vaccinalserosurveillance will be required to help demonstrate theabsence of FMDV infection. Countries or zones applyingemergency vaccination and wishing to re-establish the statusof FMD-free without vaccination would have to demonstrateabsence of persistently infected animals[22,23]. An impedi-ment to the development and application of these “vaccinate-to-live” policies has been the lack of validation data in supportof the serological tests that can detect post-vaccinal infection.

In the present study, we have challenged a relativelylarge group of cattle with FMDV by means of contactexposure with previously infected animals, the route ofinfection having been chosen to mimic a situation that mayoccur in the field more closely than is achieved by directinoculation of animals with large doses of FMDV. In aprevious report, we have described the clinical protectionafforded by vaccination (all vaccinees fully protected and allnon-vaccinees fully susceptible) and the categorisation ofthe animals into three groups after monitoring up to 28 dpc[1]. Group 1 initially comprised 11 vaccinated animals fromwhich virus could be recovered only transiently or not atall following challenge exposure. Group 2 comprised 9vaccinated animals in which virus infection had persisted upto 28 dpc. Group 3 were the five unvaccinated controls, twoof which showed evidence of virus persistence. A group of15 animals were retained for long-term follow-up, selectionh dpc.T and3

ntlyi CRr d byp eenr l thee ctedt omg beingd viralg fromG irusi ngs foundtt ve byR ost-i eens e andi after9 wasi y ofv sis-t bangs

anti-b tedf me

. Discussion

There is a heightened interest in the development of Faccination strategies that would enable rapid control of Futbreaks without the need to slaughter large numbe

nfected and at-risk animals[21]. In order to show that sucolicies have been successful and to quickly regain FMD

aving been based on the results obtained up until 28hese comprised three animals from each of groups 1, and all 9 animals from Group 2.

Categorisation of the retained cattle as persistenfected was based on a combination of VI and RT-Pesults from oesophagopharyngeal samples collecterobang sampling on a weekly basis from all of the fiftetained animals for a further 20 weeks, from 28 dpc untind of the experiment at 168 dpc. Virus was only dete

wice after 28 dpc, from all of the samples collected frroups 1 and 3, both occasions being at 42 dpc andetected only by RT-PCR. The persistence of virus orenome was confirmed in the nine vaccinated cattleroup 2 up to 42–168 dpc. A previous study comparing v

solation and RT-PCR for detection of FMDV in probaamples from vaccinated and needle-challenged cattlehe two methods comparable at all time points[24]. In con-rast, this study found more samples were scored positiT-PCR than by virus isolation, particularly at the later p

nfection time points. One animal, UV2, would not have bcored as persistently infected based on VI results alonnfectious virus was not recovered from any of the cattle8 dpc. After 56 dpc, recovery of virus or viral genome

ncreasingly spasmodic. One reason for the irregularitiral detection could be the difficulty in obtaining a conent sample of mucus and epithelium by means of a proampling cup.

Virus neutralisation tests and SPCELISA detectedodies to the FMDV structural proteins in serum collec

rom all of the animals. Although VN antibody titres in so

S. Parida et al. / Vaccine 23 (2005) 5186–5195 5193

cases declined to quite low levels towards the end of the exper-iment, the SPCELISA responses remained strong at the fixeddilution used in this test. Nearly all of the 25 cattle showed arise in antibody titre following challenge exposure, suggest-ing that they had mostly been infected with FMDV at thistime, although there is no control group that was vaccinatedbut not challenged to provide conclusive verification that theresponse was elicited by the challenge rather than a delayedeffect of vaccination. The peak level of antibody attained wasvery variable as was the decline thereafter, but vaccinated cat-tle in which virus persisted tended to maintain higher levelsof neutralising antibody than were found in vaccinated ani-mals that were only infected transiently. This supports thenotion that unusually high neutralising antibody levels foundin vaccinated cattle could be indicative of infection and viralpersistence. Under the conditions of this particular exper-iment, a homologous neutralising antibody titre of greaterthan or equal to 1 in 100 at any time between 4 and 24 weekspost infection would be highly suspicious. During large-scaleserosurveillance it may be possible to get a feel for what con-stitutes an unusually high neutralising titre of antibody inrelation to the time after vaccination and the vaccine used.

The different NSP ELISAs gave fairly consistent resultswith sera collected from the cattle in groups 1 and 3, i.e.mostly negative or mostly positive, respectively. However,antibodies were not detected by the UBI test in the unvac-c Anu oup1 wasp e).T bodyt thee eeks( aredt 1 in1 d att r ofe p int

ati entlyi teda fromC itivet BIt oup2 l tot eset ringw therb up2 SPt iso-l after1 coredR , the

sera from this animal were essentially unreactive in all ofthe NSP serological tests and not merely reactive at a levelthat did not reach the threshold for positivity. Seroconversionby VNT was also weak in UV2, reaching a peak value of1 in 128 at 28 dpc and stabilising at between 1 in 32 and 1in 45 for most of the monitoring period of this study. In thecase of UV14, there was a stronger seroconversion by VNTand sub-threshold increases in NSP serological responseswere observed. Although probang samples from this animalwere repeatedly positive up to 21 dpc, there was only a singlerecovery thereafter; virus isolation at 42 dpc.

This is the first published evaluation of 2B antibodiesas indicators of persistent infection. The prototype test thatwe used involved a 2B peptide coated directly to an ELISAplate and an indirect assay format. The method can be fur-ther optimised, and recent results have shown a significantimprovement from the use of a competitive assay formatwith an anti-2B monoclonal antibody (Inoue, unpublishedresults). Nevertheless, these preliminary results are encour-aging. Further work is needed to explore the nature of thenon-specific background reactivity of two of the cattle in thisstudy. These results were unexpected, since this effect hadnot been observed during the testing of large numbers of cat-tle in Thailand. The non-specific reactions greatly elevatedthe cut-off threshold for positive reactivity and the test wouldotherwise have had a sensitivity close to that of the Cedid

trastt en eta amet cu-l ersep liablyd i-t teda red r orlt pula-t asf m-p theC thea virusp sitiv-i Thep cci-n MDVb irusr mala ibodyr , thep d byU vely,r alsa cted

inated, infected cattle after 20 weeks post infection.nusual pattern of results was obtained for UV8, in Gr, in that only the 42 dpc serum sample from this animalositive in two of the NSP ELISAs (Cedi and 2B peptidhis sample also showed an elevated neutralising anti

itre to both O Manisa and O UKG virus, compared toquivalent samples collected in previous or subsequent wit was positive at 1 in 90 for both tests at 42 dpc, compo titres of 1 in 32 and 1 in 22 at 35 dpc and less than1 and 1 in 16 at 28 dpc). A probang sample collecte

his time was also scored positive by RT-PCR. A numbexplanations are possible, but the most likely is a mix-uhe identification of the animal at the time of sampling.

The different NSP ELISAs were not equally effectivedentifying vaccinated animals that had become persistnfected with FMDV (Group 2). None of the tests detecll of the persistently infected cattle, the best being thatedi-diagnostics, which scored seven of the nine pos

hroughout most of the period of monitoring. With the Uest, the maximum number of persistently infected Gr

cattle identified was five animals and this figure felhree for the Bommeli test. Furthermore, neither of thests detected antibodies throughout the full period duhich virus could be detected in individual animals, eiy virus isolation or by RT-PCR. The two cattle from Grothat were not found to be seropositive in any of the N

ests were UV2 and UV14. Infectious virus was notated from oesophagopharyngeal samples from UV20 dpc, but four subsequently collected samples were sT-PCR positive between 28 and 126 dpc. Interestingly

iagnostics test.The inconsistent NSP results for this study are in con

o those reported by Haas and Sorensen and by Moonl. [7,8] involving cattle that had been challenged at the s

ime interval after vaccination but by direct tongue inoation. The tongue inoculated cattle showed a less divattern of responses and the animals were more reetected by NSP ELISAs. Moonen et al.[8] reported a sens

ivity of 84% for the Bommeli test, using samples collect 21–180 days after inoculation with FMDV. This figuropped substantially if sera were examined from earlie

ater post-infection periods. Bronsvoort et al.[13] evaluatedhe performance of NSP tests in an unvaccinated poion of cattle in Cameroon where FMD is endemic. It wound that the relative sensitivity of the Bommeli test coared to VNT was only 23%, whereas the forerunner toedi test had a relative sensitivity of 71%. However, ifnalysis was restricted to samples taken from probangositive but apparently healthy animals, then the sen

ty of the two tests rose to 47 and 92%, respectively.resent study differs from earlier experiments, in that vaated and unvaccinated cattle have been exposed to Fy contact challenge rather than by needle inoculation. Veplication in recipient vaccinated cattle ranged from minind/or transient to extensive and/or persistent and antesponses largely mirrored these differences. At 28 dpcroportion of vaccinated cattle that had seroconverteBI, Bommeli and Ceditests was 0, 5 and 35%, respecti

ising to 0, 11 and 78% if only persistently infected animre considered. The proportions of Group 2 cattle dete

5194 S. Parida et al. / Vaccine 23 (2005) 5186–5195

by the UBI, Bommeli and Ceditests at 56 dpc were 44, 33and 78% respectively falling to 0, 22 and 78% at 126 dpc.However, by 56 and 126 dpc the number of cattle still shownto be harbouring virus or viral genome was only 7 and 5,respectively, and the proportions of these detected by the UBI,Bommeli and Ceditests were 71, 43 and 100% for 56 dpc and20, 40 and 100% at 126 dpc. Based on these findings, someof these tests could have difficulty to detect a low preva-lence of persistently infected cattle. For example, using a testwith an 80% sensitivity, then even if all animals are sampledand tested, 95% confidence of detecting one or more carrierswould require a prevalence of 1.5 and 15% in herd sizes of100 and of 10 cattle, respectively[25].

The likely prevalence of persistently infected cattle isclearly a key parameter in determining the predictive valueof testing for these animals in post-vaccinal serosurveillance.In this study, nine out of 20 vaccinated cattle (45%) becamecarriers, comparable to reports based on field studies[26].However, it could be envisaged that less severe challengesmight follow from indirect exposure to FMDV and that thismight result in a lower proportion. On the other hand, theinterval between vaccination and challenge was 21 days inthis study and when vaccinating in the field in the face of anoutbreak this period might be shorter, and vaccine might beless thoroughly applied. These factors would tend to lessenthe degree of protection and might contribute to a higherp

hatw rsis-t therr DVi -t usv tiono cedn 14),b s upt atu-r sitiveN posea irus,b ithern nly as

A

vi-r .

R

J, etency

FMD vaccination of cattle and the effect on virus excretion from theoropharynx. Vaccine 2005;23:1106–323.

[2] Alexandersen S, Zhang Z, Donaldson AI. Aspects of the persistenceof foot-and-mouth disease virus in animals—the carrier problem.Microbes Infect 2002;4:1099–110.

[3] Salt JS. The carrier state in foot-and-mouth disease—an immuno-logical review. Br Vet J 1993;149:207–23.

[4] Donaldson AI, Kitching RP. Transmission of foot-and-mouth dis-ease by vaccinated cattle following natural challenge. Res Vet Sci1989;46:9–14.

[5] Clavijo A, Wright P, Kitching P. Developments in diagnostic tech-niques for differentiating infection from vaccination in foot-and-mouth disease. Vet J 2004;167:9–22. Review.

[6] Mackay DK, Forsyth MA, Davies PR, Salt JS. Antibody to thenonstructural proteins of foot-and-mouth disease virus in vaccinatedanimals exposed to infection. Vet Q 1998;20(Suppl. 2):9–S11.

[7] Haas B, Sorensen KJ. Comparison of ELISAs for the differentiationof infection from vaccination by detection of antibodies to NSP3ABC of FMDV. In: Report of the European Commission for theControl of foot-and Mouth Disease, Session of the Research Groupof the Standing Technical Committee of the European Commissionfor the Control of Foot-and-Mouth Disease. 2002. p. 248–56.

[8] Moonen P, Van Der Linde E, Chenard G, Dekker A. Comparablesensitivity and specificity in three commercially available ELISAsto differentiate between cattle infected with or vaccinated againstfoot-and-mouth disease virus. Vet Microbiol 2004;99:93–101.

[9] Bulut AN, Alpay B, Ozcan C, Sareyyuboglu B. A serosurveillanceto determine the prevalence of antibodies against structural andnon-structural proteins of FMDV following the spring-2003 FMDvaccination campaign in Turkey. In: Report of the European Com-mission for the Control of foot-and Mouth Disease, Session of the

Euro-003.

[ TW.in

Rec

[ 01sionearch

om-2–5.

[ hezouth

ion: ae-

45:

[ nyaed

-and-Clin

[ ines.ony,

[ ckayth

oteinsac-

[ nti-pigs

roteinClin

revalence.Another uncertainty is the significance of the risk t

ould be posed to other susceptible animals by the peently infected cattle generated in this experiment. As in oeports, we failed to demonstrate transmission of FMnfection from persistently infected animals to naıve seninels [3]. The detection of viral genome but not infectioirus from some of the carrier animals could be an indicaf a reduced risk. Two of the nine Group 2 cattle produo detectable NSP antibody responses (UV2 and UVut infectious virus was only detected in these animalo 21 and 42 dpc, respectively. Overall, it appears that nally infected vaccinees that are not detected by a senSP test such as that from Cedi-diagnostics probablyn extremely low threat for the onward transmission of veing the animals with the least virus replication that eever became persistently infected or else harboured ohort-lived persistent infection.

cknowledgements

This work was funded by the UK Department for Enonment, Food and Rural Affairs through grant SE2918

eferences

[1] Cox SJ, Voyce C, Parida S, Reid SM, Hamblin PA, Paton Dal. Protection against direct contact challenge following emerg

Research Group of the Standing Technical Committee of thepean Commission for the Control of Foot-and-Mouth Disease. 2p. 36–9.

10] Chung WB, Liao PC, Yang PC, Chen SP, Jong MH, SheuSurveillance of FMD virus non-structural protein antibodiespig populations involved in an eradication programme. Vet2003;152:595–7.

11] Gill A. Experience in post-vaccinal surveillance following 20FMD epidemic in Uruguay. In: Report of the European Commisfor the Control of foot-and Mouth Disease, Session of the ResGroup of the Standing Technical Committee of the European Cmission for the Control of Foot-and-Mouth Disease. 2003. p. 3

12] Bergmann IE, Malirat V, Neitzert E, Beck E, Panizzutti N, SancC, et al. Improvement of a serodiagnostic strategy for foot-and-mdisease virus surveillance in cattle under systematic vaccinatcombined system of an indirect ELISA-3ABC with an enzymlinked immunoelectrotransfer blot assay. Arch Virol 2000;1473–89.

13] Bronsvoort BMD, Sorensen KJ, Anderson J, Corteyn A, TaVN, Kitching RP, et al. Comparison of two 3ABC enzyme-linkimmunosorbent assays for diagnosis of multiple-serotype footmouth disease in a cattle population in an area of endemicity. JMicrobiol 2004;42:2108–14.

14] OIE (2004). Manual of Standards for Diagnostic Tests and Vacc4th Edition. Office International des Epizooties, 12 rue de Pr75017 Paris, France.www.oie.int.

15] Sorensen KJ, Madsen KG, Madsen ES, Salt JS, Nqindi J, MaDK. Differentiation of infection from vaccination in foot-and-moudisease by the detection of antibodies to the non-structural pr3D, 3AB and 3ABC in ELISA using antigens expressed in bulovirus. Arch Virol 1998;143(8):1461–76.

16] Chung WB, Sorensen KJ, Liao PC, Yang PC, Jong MH. Differeation of foot-and-mouth disease virus-infected from vaccinatedby enzyme-linked immunosorbent assay using nonstructural p3AB as the antigen and application to an eradication program. JMicrobiol 2002;40:2843–8.

S. Parida et al. / Vaccine 23 (2005) 5186–5195 5195

[17] Shen F, Chen PD, Walfield AM, Ye J, House J, Brown F,et al. Differentiation of convalescent animals from those vacci-nated against foot-and-mouth disease by a peptide ELISA. Vaccine1999;17(23–24):3039–49.

[18] Wang CY, Chang TY, Walfield AM, Ye J, Shen M, Zhang ML, et al.Synthetic peptide-based vaccine and diagnostic system for effectivecontrol of FMD. Biologicals 2001;29:221–8.

[19] Liu S, Chang TY, Walfield AM, Zhang ML, Sokoll KK, Chen SP, etal. Synthetic peptide-based serosurveillance and vaccine system forFMD. In: Proceedings of the European Commission for the Controlof foot-and Mouth Disease. 2002. p. 303–21.

[20] Bruderer U, Swam H, Haas B, Visser N, Brocchi E, Grazioli S, et al.Differentiating infection from vaccination in foot-and-mouth-disease:evaluation of an ELISA based on recombinant 3ABC. Vet Microbiol2004;101:187–97.

[21] Royal Society Report on infectious diseases in livestock. Scien-tific questions relating to the transmission, prevention and controlof epidemic outbreaks of infectious disease in livestock in Great

Britain. The Royal Society 2002; Latimer Trend Ltd., Cornwall,London.

[22] Anon. Council Directive 2003/85/EC on Community measuresfor the control of foot-and-mouth disease repealing Directive85/511/EEC and Decisions 89/531/EEC and 96/665/EEC and amend-ing Directive 92/46/EEC. Official Journal of the European UnionL306 2003; volume 46, 22 November 2003.

[23] OIE (2004). Terrestrial Animal Health Code, 13th Edition, OfficeInternational des E*pizooties, 12 rue de Prony, 75017 Paris, France.www.oie.int.

[24] Moss A, Haas B. Comparison of the plaque test and reverse tran-scription nested PCR for the detection of FMDV in nasal swabs andprobang samples. J Virol Methods 1999;80:59–67.

[25] Cameron AR, Baldock FC. A new probability formula for surveysto substantiate freedom from disease. Prev Vet Med 1998;34:1–17.Review.

[26] Sutmoller P, Gaggero A. Foot and mouth disease virus carriers. VetRec 1965;77:968–9.