A

wctsFs©

K

1

drlpibtb

f

0d

Vaccine 25 (2007) 7120–7124

Antigenic differentiation of classical swine fever vaccinal strainPAV-250 from other strains, including field strains from Mexico

Susana Mendoza a,∗, Pablo Correa-Giron a,b, Edgar Aguilera a, German Colmenares b,Oscar Torres a, Tonatiuh Cruz a, Andres Romero a,Eliseo Hernandez-Baumgarten a, Abel Ciprian a

a Coordinacion de Estudios de Posgrado, Facultad de Estudios Superiores-Cuautitlan (Graduate Studies Coordination, Faculty of SuperiorStudies-Cuautitlan), Universidad Nacional Autonoma de Mexico, Av. Primero de Mayo S/N, Campo I, Cuautitlan Izcalli,

C.P. 54700, Apartado Postal 222, Cuautitlan Izcalli, Estado de Mexico, Mexicob CENID-Microbiologıa Animal, Instituto Nacional de Investigaciones Forestales, Agrıcolas y Pecuarias (INIFAP),

Carretera Mexico-Toluca, Km. 151/2 Cuajimalpa, Palo Alto, Mexico, D.F., C.P. 05110, Mexico

Received 27 February 2007; received in revised form 5 July 2007; accepted 26 July 2007Available online 14 August 2007

bstract

Twenty-nine classical swine fever virus (CSFv) strains were grown in the PK15 or SK6 cell lines. Antigenic differentiation studiesere performed using monoclonal antibodies (McAbs), produced at Lelystad (CDI-DLO), The Netherlands. The monoclonals which were

lassified numerically as monoclonals 2–13. Epitope map patterns that resulted from the reactivity with McAbs were found to be unrelatedo the pathogenicity of the viruses studied. Antigenic determinants were recognized by McAbs 5 and 8, were not detected in some Mexican

trains; however, sites for McAb 6 were absent in all strains. The PAV-250 vaccine strain was recognized by all MAbs, except by MAb 6.urthermore, the Chinese C-S vaccine strain was found to be very similar to the GPE− vaccine. None of the studied Mexican vaccines or fieldtrains was found to be similar to the PAV-250 vaccine strain.

2007 Elsevier Ltd. All rights reserved.

madpgdtM

eywords: CSF virus; Monoclonal antibodies; PAV-250 vaccinal strain

. Introduction

Classical swine fever virus (CSFv) is a highly contagiousisease of pigs affecting nervous, respiratory, digestive andeproductive systems, producing vascular lesions in endothe-ium and reticulo-endothelial cells. It is characterized by theresence of generalized haemorrhages and infarcts in thenternal organs [1,2]. Bovine viral diarrhoea virus (BVDv),order disease virus (BDv) and CSFv belong to the pes-

ivirus genus of the Flaviviridae family. Those viruses areoth structurally and antigenically related [3–5].

∗ Corresponding author. Tel.: +52 52 56 23 20 58;ax: +52 52 56 23 20 58.

E-mail address: seme@servidor.unam.mx (S. Mendoza).

caavatt

264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2007.07.045

Immunologically, CSFv was initially considered as aonotypic virus. It was not until the development and avail-

bility of monoclonal antibodies (McAbs) that antigenicifferences were detected, although differences have beenreviously reported regarding pathogenicity and immuno-enicity [6]. Monoclonal antibodies obviate the greatifficulties found in distinguishing between the various pes-iviruses and allow the clear differentiation of CSFv isolates.

cAbs have been recommended for CSFv identification andlassification [7], and as means to differentiate between CSFvnd BVDv [8]. Previously, complex cross-neutralization testsnd pig inoculations were required to distinguish between the

arious pestiviruses. The need to distinguish between BVDvnd CSFv arose during the development and production ofissue culture vaccines against CSFv. Culture contamina-ion with BVDv from fetal calf serum was also a problem

ccine 2

ftcvawBtta1i

orwTda1tm

tiraFb

edvop

tpdiaato

aicoapu

cCu

sa[tescMc

2

2

oLc

(

(

(

2

St

2

Lcover slips. Four tubes were used for each virus dilution.Each strain was diluted 10-fold and mixed with 200,000PK15 cells/ml. After incubation for 1 h with gentle stirring,

S. Mendoza et al. / Va

or vaccine production and for diagnostic laboratories sincehis virus, as well as CSFv, react with polyclonal sera. Bothytopathogenic and non-cytopathogenic strains of the threeiruses are included in this genus, although the CSFv gener-lly is non-cytopathogenic [9]. Contamination of cell culturesith BVDv is a serious problem to vaccine producers sinceVDv has produced outbreaks with signs that are similar

o CSFv in piglets born from vaccinated sows [8,9]. Pes-iviruses contain three proteins in their envelope: Erns, E1nd E2. These proteins either simply, or jointly give rise to3 epitopes (numbered 1–13) that have been found to clustern four regions: A–D [10].

Nishimori et al. [11] developed a different panel of mon-clonal antibodies (McAbs II and IV–VI) that recognizedegions A2, A3 or D, B and C. Another panel of McAbsas developed in Germany for use on tissue sections [12].hose monoclonals were highly specific, as demonstrated byirect immunofluorescence and biotin–avidin techniques. Innother study, eight McAbs were prepared from the Alfort87 “strain” [13]. Those McAbs recognized conserved epi-opes, and it was demonstrated that they were directed against

ajor E1 envelope glycoprotein.Studies with McAbs in several countries revealed that

he various isolates have conserved regions, probably thosenvolved in adherence to the target cells, as well as variableegions that permit strain differentiation. However, the vari-ble regions seem to have little or no bearing on pathogenicity.ield isolates are obtained from individual animals or out-reaks generally contain a mixture of several biotypes.

Mexico is currently engaged in a swine fever control andradication program. For this purpose, the country has beenivided into three regions: region 1: control with intensiveaccination; region 2: eradication with vaccination (insteadf eradication without vaccination); region 3: disease freehase.

Because phases 1 and 2 require intensive vaccination,his restricted use of vaccine prompted the research team toerform a cooperative study with the Lelystad Institute toetermine the antigenic composition of the vaccine currentlyn use and compare it with ones used in the recent past. Welso compared the vaccinal strains with the field strains toscertain that all are CSFv, which, therefore, would allow uso detect any field outbreaks caused by vaccine strains, or byther pestiviruses.

The PAV-250 strain [2], prepared in the PK15 cell line, islive modified viral vaccine against CSFv, that has been used

n Mexico from 1979 to date (Jimenez, personal communi-ation). During the first 4 months of 2002, 100,645 dosesf CSF were sold; 95.2% of them were the PAV-250 strainnd 4.8% corresponded to the PAV-1 vaccine. This last strain,repared in suspended cultures of pig red bone marrow, wassed between 1977 and 2002.

The Minnesota strain was produced in pig kidney cellultures, and was used in Mexico prior to 1971–1995. Thehinese strain was commercialized around 1971 and usedntil 1997. This strain was produced in rabbits. The Japanese (

5 (2007) 7120–7124 7121

train (GPE) appeared in the commercial literature frombout 1987 and was produced in guinea pig cell cultures14]. All those vaccines were available prior to, and duringhe beginning of the campaign, when outbreaks of the dis-ase were present. The possibility of low pathogenicity wildtrains should be considered. In the present study, identifi-ation of the viruses as CSFv, antigenic differences betweenexican CSFv vaccines, strains isolated from field clinical

ases and reference strains were analyzed.

. Materials and methods

.1. Viral stains

Twenty-nine strains of CSFv were supplied by the Virol-gy Laboratory, Central Veterinary Institute (CDI-DLO)1 ofelystad, The Netherlands. The strains were identified by theountry origin as follows:

(a) Germany: high virulence (H) Behring.b) U.S.A.: BAI (H); Cornell (H); Ames (H); 331, medium

virulence (M); New Lederle (H).(c) France: Alfort (H); Alfort 2.3.1 (M); Alfort 2.3.2 (M).d) Japan: ALD (H) and GPE, low virulence (L).

(e) The Netherlands: Brescia 1.1.1 (H); Brescia 2.1.1 (H);Baker A 1.2.1 (H); Henken (L); Cedipest (L). Fieldstrains included Jongen (H) and Wild Boar (H).

(f) Poland: field strains Spruit 2 (H); Jongerbreur (H).g) Mexico: vaccinal strains included Minnesota (L); PAV-1

(L); China C-S (L) and PAV-250 (L). Field strains wereVC-55 (H); VC126 (H); VC-039 (ND); VC-127 (ND).The pathogenic reference strain ALD (H) and field strainsthat were isolated and identified between 1996 and 1998were furnished by one of us (Correa). Field strains andthe ALD reference strains were adapted to the PK15 cellline and lyophilized for transport from Mexico to theLelystad Institute.

.2. Cell lines

The CSF strains from the Lelystad Institute were grown inK6 and/or PK15 cell lines [15]. The Mexican strains were

itrated in SK6 and/or PK15 cells.

.3. Viral titrations

To standardize tests, all CSF strains were titrated ineighton tube cultures of PK15 cells that contained glass

1 From the “Centro Nacional de Servicios de Diagnostico en Salud AnimalCENASA), SAGARPA”, Santa Ana Tecamac, Mexico.

7 ccine 2

tiabwwim

2a

LlMwb1odt

iw9m

iuser

3

3

c

TA

T(d

122 S. Mendoza et al. / Va

he screw cap tubes were each inoculated with 1 ml of the var-ous virus-cell dilutions. Screw caps were left slightly loose,nd the tubes were incubated for 4 days at 37◦ in a CO2 incu-ator. At the end of the incubation period, each cover slipas washed with PBS, fixed with cold acetone and stainedith a polyclonal antibody conjugated with fluorescein isoth-

ocyanate. Titers were determined by the Reed and Muenchethod [16].

.4. Immunoperoxidase assay using monoclonalntibodies

This study employed McAbs provided by the CDI-DLO,elystad. Each CSF strain was evaluated by infecting mono-

ayer cell cultures and then, after 3 days, applying the 12Abs conjugated to horseradish peroxidase. The McAbsere developed against the Brescia strain [10,17] and num-ered as 2–13. To determine the optimal virus dilution,

0-fold dilutions of each strain were made and four wellsf a 96-well microplate were filled with 20 �l of each virusilution. Twenty microliters of a PK15 cell suspension washen added and plates were maintained for 3 days in a 5% CO2

waew

able 1ntigenic analysis of 29 CSFv “strains” using 12 monoclonal antibodies directed a

he shaded areas represent strains whose epitopes are recognized by the monoclonal aW) Weak reaction; (H) highly pathogenic strains; (M) medium pathogenicity strainetermined.

5 (2007) 7120–7124

ncubator. The plates were then fixed, washed and incubatedith the McAbs for 1 h; they were developed with 3-amino--ethil-carbazole (AEC) and observed, using an invertedicroscope [10,17,18].Tests were considered positive when the cytoplasm of

nfected cell monolayers stained brown by the reaction prod-ct of the enzyme. When a monolayer was only partiallytained, it was considered as a weak reaction. Three differ-nt technicians tested each CSFv strain consecutively and theesults were then compared.

. Results

.1. Pestivirus identification

Monoclonals 2–4, 7 and 9–11, wich recognize epitopeslustered in domains A1 and A2, identified all strains tested,

hile recognition of epitopes on domains A3 and D was vari-

ble (MAbs 12, 13 (W)). Common clusters were in all of theight Mexican “strains” as CSFv. None of them reacted eitherith BVDv or with BDv specific MAbs, confirming both the

gainst epitopes of the Brescia strain

ntibodies, while the unshaded areas represent strains that are not recognized.s; (L) low pathogenicity strains; (ND) strains whose pathogenicity was not

ccine 2

iwm

3s

gMeIiasaoWJ

3sV

aArssBrww

eTtB

rso

4

updt[Eift

acT

tRd

[ntei

tia

4

dtwtsMats

hdwtwtta1t(ftvet

vttc

S. Mendoza et al. / Va

dentity of the Mexican strains included in this study. Theeak reaction of vaccine strains Minnesota and PAV-1 toonoclonal 13 is taken to identify a partly different epitope.

.2. Monoclonal antibody differentiation of the CSFtrains

Results obtained with MAbs 2–13, which cover the anti-enic spectrum of the Brescia strain, are shown in Table 1.onoclonal 5 failed to recognize most European strains,

xcept the Brescia 1.1.1, Brescia 2.1.1 2 and Baker A strains.t did not identify any of the Japanese or US viruses, butt react with three of the Mexican strains (PAV-250, VC-55nd VC-126). It weakly recognized the China C-S vaccinetrain. Monoclonal 6 reacted with one US strain (Cornell),nd weakly with another one (Ames). It identified the samenes from The Netherlands as monoclonal 5, and also theild Boar strain and the two Poland strains (Spruit 2 and

ongerbreur).Monoclonal 8 identified only two US strains (Cornell and

31), but all strains from The Netherlands. It also reactedtrongly with the Mexican strains China C-S, PAV-250 andC-127, but weakly with Minnesota and PAV-1.Monoclonal 12 recognized the German strain (Behring)

nd all of the US strains. Two French strains (Alfort andlfort 2.3.1) were not recognized; however, there was a weak

eaction with one French strain (Alfort 2.3.2). Mab 12 reactedtrongly with Japanese strain ALD, but it failed to react withtrain GPE. Four strains from The Netherlands (Brescia 1.1.1,rescia 2.1.1, Baker A 1.2.1 and CEDIPEST) gave strong

eactions, but the two polish strains (Spruit 2 and Jongerbreur)ere negative. On the other hand this monoclonal reactedith all the Mexican strains.Monoclonal 13 reacted with the majority of the strains,

xcept for two from the US (Cornell and 331) and one fromhe Netherlands (Jongen). However, there was a weak reac-

ion with three Dutch strains (Henken, CEDIPEST and Wildoar); and two Mexican strains (Minnesota and PAV-1).

The PAV-250 vicinal strain’s antigenic determinants wereecognized by all MAbs, except for Mab 6, which corre-ponds to subdomain B. This strain was distinctive from thether studied strains.

. Discussion

This study compared field and vaccinal strains of CSFvsed in Mexico using an immunoassay that employed aanel of 13 peroxidase-labeled McAbs. The failure of McAbsirected to conserved epitopes of BVDv and BDv indicatedhe lack of contamination by these, or other pestiviruses10,19]. The antigenic determinants of the A domain, the

2 viral glycoprotein (gp 51–54), are recognized as the most

mportant ones for viral neutralization. Those domains wereound in all of the studied strains, which confirm their iden-ity as CSFv. The A1 subdomain of the A region, which is

l

to

5 (2007) 7120–7124 7123

lso a conserved neutralization site located in the same gly-oprotein, was present in all the studied strains, as shown inable 1 [15,17].

The McAbs used in this study were not able to differen-iate between strains of high, medium or low pathogenicity.esults similar to those reported here were obtained in studiesone in The Netherlands [10,19].

Results of studies carried out in Japan, and elsewhere,11,12] are difficult to compare to ours since different termi-ology and different sets of monoclonals were used. In ordero compare the various strains using the McAbs from differ-nt laboratories, studies would be required to be carried outn the same laboratory, using the various sets of McAbs.

To assist the National Eradication Campaign in Mexico,he initial goal was to distinguish the pathogenic field strainssolated in Mexico from the vaccinal strains, using McAbs indiagnostic tests that would differentiate them.

.1. Strains

The low pathogenicity strains (PAV-1 and PAV-250) wereerived from the same source, but they had different reactivityo Mabs 5 and 6 (Table 1). This minor antigenic variationas independent of the number of passages required to adapt

hem to grow in cell cultures [10]. The PAV-250 vaccinaltrain was the only strain studied that was not recognized by

Ab 6 (domain B), but was recognized by the other MAbs,nd we take this to mean that it has not been transmittedo non-vaccinated pigs, as it is not represented in the fieldtrains.

The Mexican field strains included in this study revealedeterologous reactions in their secondary epitopes that wereistantly related to the conserved neutralization sites. Thisas not surprising since they had been isolated at different

imes, often with a separation of several years. This variationas restricted to secondary neutralization sites of CSFv. On

he other hand the pathogenic strain ALD, which served ashe source of the GPE− vaccine [11], kept most of its originalntigenic identity. The reactivity of Mexican strains to MAb2, directed to subdomain A3, was found in all strains, excepthe Chinese C-S strain. Epitope reactivity with domain BMAb6), was not observed in the Mexican strains; epitopesrom region C (MAbs 5 and 8) were also absent in some ofhese strains, e.g., PAV-1 and Minnesota. Only the PAV-250accine has been used in phases 1 and 2 of the Mexican CSFvradication campaign. This vaccine was shown to be superioro the other classical swine vaccines available in Mexico [2].

Pigs vaccinated with PAV-250, and challenged on post-accination day 14 with virulent ALD-CSFv, did not transmithe challenge virus to susceptible pigs. Notwithstanding,here was low level shedding of virulent virus from the vac-inated pigs following challenge; however, virus shed was at

evels below the infectious dose [20].

Since the McAbs used in this study were developed againsthe Brescia strain, the failure of some epitopes to be rec-gnized in the Mexican vaccines and field strains suggests

7 ccine 2

tepsfiTumsioowv

A

tatfi

R

[

[

[

[

[

[

[

[

[

[

Gen Virol 1990;71:531–40.

124 S. Mendoza et al. / Va

hat the Mexican viruses differ from the Brescia strain. How-ver, the Minnesota and PAV-1 vaccinal strains gave similaratterns. Patterns also were similar with other swine fevertrains, but none of the studied Mexican field strains wereound similar to the PAV-250 vaccinal strain. Similar find-ngs were made in regard to the other strains examined.herefore, PAV-250 may be differentiated from other strainssing the panel of Mabs employed in this study. With theethods used in this study antigenic variation between CSFv

trains/isolates could be detected; however the origin of thesolates could not be determined. In conclusion, the technol-gy described here allows the detection of antigenic variantsf CSFv if they occur in vaccinated pigs. The method alsoould detect the rare cases of BVD infection in pigs againstaccinated CSF.

cknowledgements

We thank DVM David Trujillo and Gabino Sanchez forheir technical support, DVM Renate M. Thummler Blumnd Dr.Leland Carmichel (Prof. Emeritus) for her help inhe preparation of this manuscript. This study was partiallynanced by CONACYT 1082-PB.

eferences

[1] Van Oirschot JT. Classical swine fever. In: Straw BE, Mengeling WL,D’Allaire S, Taylor DJ, editors. Diseases of swine. 7th ed. IA, USA:Iowa, State University Press; 1999. p. 159–72.

[2] Correa-Giron P. Characterısticas mas importances de la vacunaPAV-250 y de las vacunas contra la Fiebre Porcina Clasica (FPC)usadas en Mexico. In: Morilla GA editor. La Fiebre Porcina Clasicaen las Americas. Symposium International Puebla-1998, INIFAP,SAGARPA, IICA, FUPPUE, Mexico, 2000. p. 301–23.

[3] Terpstra C. Detection of border disease antigen in tissues of affectedsheep and in cell cultures by immunofluorescence. Res Vet Sci1978;25:350–5.

[4] Westaway EG, Brinton MA, Gaidamovich SYA, Horzinek MC, IgarashiA, Kaarianen L, et al. Togaviridae. Intervirology 1985;24:125–39.

[5] Francki R.I.B, Fauquet C.M, Knudson O.L, Brown F. Classificationand nomenclature of viruses. Sixth Rep Int Committee Taxon Viruses1995;4(24).

[

5 (2007) 7120–7124

[6] Laude H. Nonarbo-togaviridae: Comparative hydrodynamic propertiesof the pestivirus genus. Arch Virol 1979;62:347–52.

[7] Edwards S, Moening V, Wensvoort G. The development of aninternational reference panel of monoclonal antibodies for the dif-ferentiation of hog cholera virus from other pestivirus. Vet Microbiol1991;29:101–8.

[8] Tepstra C, Wensvoort G. Natural infections of pigs with bovine viraldiarrhea virus associated with signs resembling swine fever. Res VetSci 1988;45:139–42.

[9] Wenvoort G, Terpstra C. Bovine viral diarrhea infections in piglets fromsows vaccinated against swine fever with contaminated vaccine. ResVet Sci 1988;45:134–8.

10] Wensvoort G. Topographical and functional mapping of epitopesof hog cholera virus with monoclonal antibodies. J Gen Virol1989;70:2865–76.

11] Nishimori T, Yamada S, Shimizu M. Production of monoclonalantibodies against classical swine fever and their use for anti-genic characterization of Japanese isolates. J Vet Med Sci 1996;58:707–10.

12] Zhou Y, Moening V, Coulibaly D, Dahle J, Liess B. Differentia-tion of hog cholera and bovine virus diarrhea. Zentralbl Vet 1989;36:76–82.

13] Weiland G, Moening V, Coulibaly Z, Dahle J, Leder L, Liess B. Iden-tification of conserved epitopes on a Hog Cholera virus protein. ArchVirol 1990;111:213–25.

14] Sasahara J, Kumagai T, Shimizu Y, Furuchi S. Field experiments ofHog Cholerae live vaccine prepared in guinea pig kidney cell culture.Natl Inst Annu Health Quart 1969;9:83.

15] Wensvoort G, Terpstra C, De Kluijver EP, Kragten C, Warnaar JC. Anti-genic differentiation of pestivirus strains with monoclonal antibodiesagainst hog cholera virus. Vet Microbiol 1989;21:9–20.

16] FAO. Organizacion de la Naciones Unidas para la Agricultura y laAlimentacion. Manual de Tecnicas de Diagnostico Virologico. Red deCooperacion Tecnica. 1a Ed. Santiago de Chile. Organizacion de lasNaciones Unidas para la Agricultura y la Alimentacion; 1993.

17] Van Rijn PA, Bossers A, Wensvoort G, Moormann RJM. Classicalswine fever virus (CSFV) envelope glycoprotein E2 containing onestructural antigenic unit protects pigs from lethal CSFV challenge. JGen Virol 1996;77:2737–45.

18] Terpstra C, Bloemraad M, Gielkens ALJ. The neutralizing peroxidase-linked assay for detection of antibody against swine fever virus. VetMicrobiol 1984;9:113–20.

19] Wensvoort G, Boonstra J, Bodzinga BG. Immunoaffinity purificationand characterization of the envelope protein El of hog cholera virus. J

20] Gonzalez C, Pijoan C, Ciprian A, Correa P, Mendoza S. Airborne trans-mission of Hog Cholera virus in susceptible pigs and the effect ofvaccination with the PAV-250 HCV strain on the airborne transmisionof HVC. J Vet Med Sci 2001;63:991–6.