A

hafREces©

K

eahAsro

tWofi(

0d

Available online at www.sciencedirect.com

Virus Research 131 (2008) 100–105

Short communication

Genetic characterization of equine influenza viruses isolatedin Italy between 1999 and 2005

A.M. Damiani a, M.T. Scicluna a, I. Ciabatti a, G. Cardeti a, M. Sala a, G. Vulcano a,P. Cordioli b, V. Martella c, D. Amaddeo a, G.L. Autorino a,∗

a National Reference Centre for Equine Diseases-Istituto Zooprofilattico Sperimentale delle Regioni Lazio e Toscana,Via Appia Nuova 1411, 00178 Rome, Italy

b Istituto Zooprofilattico Sperimentale della Lombardia e dell’Emilia Romagna, Via Bianchi 7, 25124 Brescia, Italyc Department of Animal Health and Well-being, University of Bari, S.p. per Casamassima km 3, 70010 Valenzano, Bari, Italy

Received 28 May 2007; received in revised form 31 July 2007; accepted 1 August 2007Available online 24 September 2007

bstract

During local respiratory disease outbreaks, occurring in 2003 and 2004 in horse training stables within race-tracks in Rome, and on a studorse farm in Bari in 2005, four strains of equine influenza (EI) virus were isolated. All outbreaks occurred in flu-vaccinated horses. Here, were reporting the results of the genetic characterization of these isolates, together with that of another EI virus strain isolated in 1999 from a deadoal presenting pulmonary lesions. Alignment and phylogenetic analyses were carried out using the haemagglutinin amino acid sequences. Theome and Bari isolates were identified as members of the American lineage, closely related to other recent strains isolated in America as well as inurope, including the latest recommended American lineage vaccine prototype A/eq/SouthAfrica/4/2003. In contrast, the Italian 1999 isolate was

lustered within the European lineage. In Italy, the most recent outbreaks of EI have been caused by the currently circulating American-like strains,ven in vaccinated populations, confirming that vaccines should contain an updated representative strain of this lineage. Presently, companies aretill in the process of registering updated vaccines but no product is yet available on the market.

2007 Published by Elsevier B.V.

; Euro

tdle1ccDV(t

eywords: Equine influenza virus; Genetic characterization; American lineage

Equine influenza (EI) is a highly contagious respiratory dis-ase, causing marked economic losses to the equine industry asresult of a long convalescence and high cost of therapy of sickorses. During infection caused by the EI virus (EIV), a type

Orthomyxovirus, the horse may show the following typicaligns: high pyrexia, dyspnoea, coughing and serous to mucopu-ulent nasal discharge as well as myalgia, anorexia and swellingf regional lymph nodes (Daly and Mumford, 2001).

Two antigenically different subtypes of EIV have been iden-ified: equine influenza A H7N7 and equine influenza A H3N8.

hile the latter is the subtype responsible for widespread

utbreaks in vaccinated and unvaccinated horses, the last con-rmed outbreak caused by an H7N7 subtype virus was in 1979Daly et al., 2004a). Phylogenetic studies have demonstrated

∗ Corresponding author. Tel.: +39 06 79099449; fax: +39 06 79099450.E-mail address: gianluca.autorino@izslt.it (G.L. Autorino).

Alwg2

g

168-1702/$ – see front matter © 2007 Published by Elsevier B.V.oi:10.1016/j.virusres.2007.08.001

pean lineage; Vaccine selection

hat since 1986, the equine influenza A H3N8 subtype hasiverged in two distinct genetically and antigenically differentineages, ‘European’ and ‘American’ (Daly et al., 1996; Lait al., 2001; Oxburgh et al., 1994; Oxburgh and Klingeborn,999). Viruses of the two lineages co-circulate, therefore vac-ines should include virus strains representative of both lineagesurrently circulating in the field (Daly et al., 1996, 2004b).uring the Expert Surveillance Panel on Equine Influenzaaccines, headed by the Office International des Epizooties

OIE) in 2006, it was recommended that vaccines should con-ain an A/eq/SouthAfrica/4/2003 (H3N8)-like virus for themerican lineage and an A/eq/Newmarket/2/1993 (H3N8)-

ike virus for the European lineage. Another recommendationas the constant surveillance necessary to spot any emer-

ence of antigenic and genetic variants of EI viruses (OIE,006).

In response to these recommendations, this study reports theenetic characterization of EI viruses isolated in Italy between

Rese

1p

tmottfshPasAoiRod(anrabiv

uitaRi2sa3TfcapsoctisHCTACCt

AntwAAEwu

oAauPpwcr(ttccHodtoAO

sts((2eAAa2ttfA(faaAA

A.M. Damiani et al. / Virus

999 and 2005 and their relatedness to recommended vaccinerototypes.

In April–May 2003, in the Capannelle race-track of Rome,horoughbred racehorses, although flu-vaccinated, presented

ild signs of respiratory disease. Nasal swabs submitted tour laboratory were inoculated into the allantoic cavity of 8o 10day old embryonated hen’s eggs. After 48–72 h incuba-ion at 37 ◦C, the allantoic fluid was harvested and checkedor the presence of a haemagglutinating virus employing thetandard methods (OIE, 2004). The viruses isolated from twoorses were analysed in nucleoprotein and H3 genes RT-nestedCRs (Oxburgh and Hagstrom, 1999) in order to confirm thems influenza virus type A and equine H3 subtype. The twotrains were, respectively, named as A/eq/Roma/1/2003 and/eq/Roma/2/2003. In the next 2 years, another two outbreaksf respiratory disease were detected. The first occurred dur-ng February–March 2004 in the Tor di Valle race-track ofome, affecting flu-vaccinated standard bred horses. The sec-nd was identified on a stud farm located in Southern Italy,uring December 2005, again in flu-vaccinated Saddle horsesMartella et al., 2007). Samples were sent to the laboratorynd processed as described above. Two strains of EIV, desig-ated as A/eq/Roma/2004 and A/eq/Bari/2005, were isolated,espectively, from the 2004 and 2005 outbreaks. Genetic char-cterization of the four isolates was then performed as describedelow. An EIV strain, isolated in 1999 from a dead foal present-ng pulmonary lesions, was also genetically characterised: thisirus was named A/eq/Brescia/1999.

Viral RNA was extracted from 140 �l of allantoic fluidsing the QIAamp viral RNA mini kit (Qiagen) accord-ng to the manufacturer’s instructions. For sequencing ofhe HA1 portion of the HA gene, reverse transcriptionnd PCR were performed using the SuperScript one-stepT-PCR with Platinum Taq (Invitrogen) in the follow-

ng conditions: 1× final concentration reaction mix with00 �M each of dNTPs and 1.2 mM of MgCl2, 0.3 �M ofense (H3HA1/1: 5′-AGCAAAAGCAGGGGATATTTCTG-3′)nd anti-sense (H3HA1/2: 5′-GCTATTGCTCCAAAGATTC-′) primers, 1 �l of RnaseOut (Invitrogen), 1 �l of RT/Platinumaq mix and 5 �l of viral RNA in a total 50 �l reaction. Theollowing thermal profile was run: 50 ◦C × 30′, 94 ◦C × 2′, 30ycles of 1′ at 94 ◦C, 1′ at 58 ◦C and 2′ at 72 ◦C followed by 7′t 72 ◦C. To compensate for possible errors of nucleotide incor-oration from the Taq polymerase, the reactions and successiveequencing were done in triplicate. The PCR amplified productsf about 1080 bp were purified using the QIAquick PCR purifi-ation kit (Qiagen) and extended by the ABI PRISM BigDyeerminator v1.1 cycle sequencing kit (Applied Biosystem) using,n addition to the H3HA1/1 and H3HA1/2 primers, otherix primers (H3HA1/3: 5′-TGAGGTGACAAATGCTACTG-3′;3HA1/4: 5′-TTGGGGATGCTTCCATT-3′; H3HA1/F525: 5′-TGAATTGGCTAACAAAATC-3′; H3HA1/R747: 5′-CTTA-CCTGCCTGATTGACC-3′; H3HA1/7: 5′-TCCTAATGAT-

AACAGTAATG-3′; H3HA1/8: 5′-GGAGCGATGCATAGT-AGG-3′). The sequencing products were then purified withentrisep columns (Applied Biosystem) and the reaction mix-

ures were analysed using capillary electrophoresis on an

(2

t

arch 131 (2008) 100–105 101

BI PRISM 310 Genetic Analyzer (Applied Biosystem). Theucleotide sequences were assembled and translated usinghe SeqMan and EditSeq programs of the Lasergene soft-are package version 5 (DNASTAR). The HA1 sequences of/eq/Roma/1/2003, A/eq/Roma/2/2003, A/eq/Roma/2004 and/eq/Brescia/1999 were deposited in GenBank (accession nos.F536318, EF536319, EF536320 and EF536321, respectively),hereas the one corresponding to the A/eq/Bari/2005 strain isnder accession no. EF117330 (Martella et al., 2007).

The 329 amino acid (AA) sequences of the HA1 portionf the HA (where the AA number 1 corresponds to the firstA after the signal peptide and the last AA is the arginine

t the cleavage site between HA1 and HA2) were alignedsing the Clustal W method of the above-mentioned software.hylogenetic analysis was performed using the Phylip softwareackage, version 3.6 (Felsenstein, 2001). The genetic distancesere calculated using ProtDist and a phylogenetic tree was

onstructed using the Neighbour-Joining method. The statisticalobustness of the tree was assessed by bootstrap resampling100 data sets) of the multiple alignment; bootstrap consensusree was built by using Consense. The tree was visualised usinghe Treview software, version 1.6.6 (Page, 1996). The geneticharacterization, consisting in sequence alignment and theonstruction of a phylogenetic tree, was performed using theA1 AA sequences of the five isolates together with those ofther 45 EIV strains. Forty-one out of the 45 sequences areeposited in GenBank and belong to strains isolated from 1991o 2005. To the former list we also added the HA1 sequencesf A/eq/SouthAfrica/4/2003, A/eq/Newmarket/5/2003,/eq/Holland/1/1995 and A/eq/Holland/2/1995 (provided byIE Reference Laboratory, Newmarket, UK).Sequence alignment revealed that putative N-glycosylation

ites located at positions 8, 22, 38, 53, 63, 165 and 285 ofhe HA1 are conserved. Moreover, by analyzing the AAubstitutions which allow the distinction between Americanpositions 261Arg → Lys and 276Thr → Ile) and Europeanpositions 78Val → Asp, 163Thr → Ile, 189Gln/Asn → Lys,07Lys → Glu, 213Ile → Val and 273Pro → Leu) lin-age isolates, the A/eq/Roma/1/2003, A/eq/Roma/2/2003,/eq/Roma/2004 and A/eq/Bari/2005 strains were classified asmerican lineage strains while the A/eq/Brescia/1999 strain

s an European lineage (Daly et al., 1996; Manuguerra et al.,000). High AA identities (99.4–99.7%) were observed amonghe Italian 2003, 2004 and 2005 isolates; when comparedo other post-2003 isolates of the same lineage they showedrom 97 to 100% AA identities. The two Rome 2003 strains,/eq/Roma/1/2003 and A/eq/Roma/2/2003, differed by 1 AA

174Phe → Tyr). When compared to the 2004 isolate, theormer strains presented 2 (263Arg → Gly and 174Phe → Tyr)nd 1 (263Arg → Gly) AA substitutions, respectively. Inddition to the substitutions at position 174 and 263 for/eq/Roma/1/2003 and A/eq/Roma/2004, respectively, the/eq/Bari/2005 presented other 2 additional AA changes

7Gly → Asp and 213Ile → Met) when compared to the Italian003 and 2004 isolates (Fig. 1A).

Aligned to the new recommended vaccine strain proto-ype for the American lineage A/eq/SouthAfrica/4/2003, the

102 A.M. Damiani et al. / Virus Research 131 (2008) 100–105

Fig. 1. Alignment of haemagglutinin (HA1) amino acid sequences. (A) A/eq/Roma/1/2003, A/eq/Roma/2/2003, A/eq/Roma/2004 and A/eq/Bari/2005 aligned withthe new (A/eq/SouthAfrica/4/2003) and the former (A/eq/Newmarket/1/1993) recommended vaccine prototypes for the American lineage. (B) A/eq/Brescia/1999aligned with the recommended vaccine prototype (A/eq/Newmarket/2/1993) for the European lineage and other recent isolates of the same lineage. N-linkedglycosylation sites are underlined. Putative antigenic sites in which the Italian isolates show amino acid substitutions when compared to vaccines prototypes areindicated in squares.

Rese

AasB31itAgr

s1tAAa

FAb

A.M. Damiani et al. / Virus

/eq/Roma/1/2003, A/eq/Roma/2/2003, A/eq/Roma/2004nd A/eq/Bari/2005 strains have only 3, 2, 3 and 4 AAubstitutions, respectively (Fig. 1A). Moreover, the Rome andari isolates share 8 of the 10 AA substitutions (5Thr → Ile,0Thr → Ser, 48Ile → Met, 58Val → Ile, 190Gln → Glu,93Glu → Lys, 272Ala → Val and 289Pro → Ser) presentn the A/eq/SouthAfrica/4/2003 strain, when compared to

he former vaccine prototype for the American lineage,/eq/Newmarket/1/1993 (Fig. 1A), including those at the anti-enic site B (positions 190 and 193) and C (position 48). Of theemaining two distinctive AA substitutions in the South African

pcAr

ig. 2. Phylogenetic tree for the haemagglutinin (HA1) amino acid sequences of equi/eq/Newmarket/1/1993 and A/eq/Newmarket/2/1993 are highlighted in gray. The trranch points. The bar represents 0.01 amino acid changes per site.

arch 131 (2008) 100–105 103

train, a further one is located at antigenic sites B (position59) and the other at E (position 78). In addition, when alignedo A/eq/Newmarket/1/1993 and A/eq/SouthAfrica/4/2003, the/eq/Roma/1/2003 and A/eq/Bari/2005 strains each present anA substitution, both located at antigenic site D (174Phe → Tyr

nd 213Ile → Met, respectively).For the A/eq/Brescia/1999, which belongs to the Euro-

ean lineage, AA identities of 97–99.4% are observed whenompared to other post-1998 isolates of the same lineage.lignment to the reference strain A/eq/Newmarket/2/1993

eveals the presence of six AA substitutions (5Thr → Ile,

ne influenza viruses. The Italian 1999–2005 isolates, A/eq/SouthAfrica/4/2003,ee has A/eq/Miami/1963 as outgroup. Bootstrap values are shown at the main

1 Rese

42cd8

IlittcKUtsaAg1aAa2ssd((ttt

eEstpAtm

tosAodgitiIsAiI

2u

saeAaarAgeSca

tisItftAst

A

P

R

D

D

D

D

D

F

L

M

04 A.M. Damiani et al. / Virus

8Ile → Met, 83Asn → Ser, 135Gly → Arg, 242Ile → Val and80Glu → Lys) (Fig. 1B). Four of these substitutions, whichan also be found in other post-1998 European-like isolates, areirectly associated with antigenic sites at positions 48 (site C),3 (site E), 135 (site A) and 242 (site D) (Fig. 1B).

The phylogenetic tree (Fig. 2), as expected, locates thetalian 2003, 2004 and 2005 isolates within the Americanineage. These strains form a cluster, together with recentsolates from United Kingdom and United States as well ashe A/eq/SouthAfrica/4/2003, into the Florida sub-lineage. Onhe other hand, the former vaccine prototype for the Ameri-an lineage, the A/eq/Newmarket/1/1993, is placed within theentucky sub-lineage, together with strains from Europe andnited States isolated before 2001. Interestingly, in two dis-

inct occasions Italian and British strains isolated during theame years are sub-clustered together. The A/eq/Roma/1/2003nd A/eq/Roma/2/2003 are place closed to the British/eq/Newmarket/5/2003 and, likewise, the A/eq/Bari/2005 isrouped with other 2005 British isolates. By alignment, a00% AA identity is observed between A/eq/Roma/2/2003nd A/eq/Newmarket/5/2003 and A/eq/Bari/2005 with both/eq/Essex/2/2005 and A/eq/Essex/3/2005. In the latter event,

lignment also revealed that A/eq/Bari/2005 and all the British005 isolates shared a distinctive motif (a Phe-Ile repeat) in theignal sequence (data not shown). The A/eq/Newmarket/5/2003train was isolated from a large outbreak in the United King-om that had began around a month earlier to the Italian oneNewton et al., 2006). Even in the other case the Essex outbreakDEFRA/AHT/BEVA, 2006) occurred around the same time ashe Italian one suggesting that in both occasions the spread ofhe infection was the result of movement of horses between thewo countries.

The A/eq/Brescia/1999, belonging to the European lin-age of EI viruses, is clustered with post-1998 isolates fromurope, differently from the reference European-like vaccinetrain A/eq/Newmarket/2/1993 (OIE, 2006) which is placedogether with viruses isolated before 1998. While the Euro-ean lineage is described to cause mild clinical infections, the/eq/Brescia/1999 strain was isolated from a dead foal, in which

he influenza virus might not have been the primary cause ofortality.In Italy, most of equine influenza vaccines contain representa-

ive strains (as whole inactivated viruses or HA and NA subunits)f an H7N7, an American-like H3N8 and European-like H3N8ubtypes. However, vaccines containing the new recommendedmerican lineage vaccine prototype A/eq/SouthAfrica/4/2003r like strains are not yet available. The results of this workemonstrate that recent Italian American-like isolates are moreenetically related to viruses isolated around the same periodn United Kingdom, United States and South Africa thano the former recommended vaccine strain for the Amer-can lineage A/eq/Newmarket/1/1993. The analyses of thetalian 2003–2005 isolates for their HA1 compositions has

hown that they share most of the substitutions present in the/eq/SouthAfrica/4/2003, some located at antigenic sites. This

s in line with what the Expert Surveillance Panel on Equinenfluenza (OIE, 2006) has reported for other recent 2005 and

M

arch 131 (2008) 100–105

006 EI isolates, who have also established antigenic correlationsing ferret sera.

The A/eq/Brescia/1999 isolate, an European-like H3N8ubtype, contains AA substitutions at antigenic sites thatre also found in other post-1998 isolates of the same lin-age when compared to the recommended vaccine prototype/eq/Newmarket/2/1993 strain. Despite these AA substitutions,

ntigenic characterization of a 2005 European-like isolate showshigh reaction in the haemagglutination inhibition test with fer-

et antisera raised against the European lineage reference strain/eq/Newmarket/2/1993 (OIE, 2006). In view of this high anti-enic reactivity, together with the absence of serious clinicalpisodes attributed to these viruses in the last years, the Experturveillance Panel on Equine Influenza recommends that EI vac-ines should still contain an A/eq/Newmarket/2/1993-like viruss a representative of the European lineage.

The majority of EIV isolates described here have been iden-ified as American lineage strains. The EI outbreaks occurredn flu-vaccinated populations as a consequence of a pre-umable vaccine breakdown. Genetic characterization of thetalian isolates has identified AA substitutions at antigenic siteshat likely account for the vaccine failure. Our study rein-orces the recommendation to use flu-vaccines updated forhe A/eq/SouthAfrica/4/2003-like virus as representative of themerican lineage of EI viruses, as well as reviewing immuni-

ation protocols to improve and ensure a baseline protection inhe horse population.

cknowledgment

This work was supported by the Italian Ministry of Healthroject IZSLT 01/03RC.

eferences

aly, J.M., Lai, A.C.K., Binns, M.M., Chambers, T.M., Barrandeguy, M., Mum-ford, J.A., 1996. Antigenic and genetic evolution of equine H3N8 influenzaA viruses. J. Gen. Virol. 77, 661–671.

aly, J.M., Mumford, J.A., 2001. Influenza infections. In: Lekeux, P. (Ed.),Equine Respiratory Diseases. International Veterinary Information Service,Ithaca, NY, USA, http://www.ivis.org/.

aly, J.M., Newton, J.R., Mumford, J.A., 2004a. Current perspectives on controlof equine influenza. Vet. Res. 35, 411–423.

aly, J.M., Yates, P.J., Newton, J.R., Park, A., Henley, W., Wood, J.L.N., Davis-Poynter, N., Mumford, J.A., 2004b. Evidence supporting the inclusion ofstrains from each of the two co-circulating lineages of H3N8 equine influenzavirus in vaccines. Vaccine 22, 4101–4109.

EFRA/AHT/BEVA, 2006. Equine Quarterly Disease Surveillance Report, vol.1, no. 4, http://www.aht.org.uk/pdf/equine1 vol4.pdf.

elsenstein, J., 2001. PHYLIP: Phylogenetic Inference Package 3.6. Distributedby the author. Department of Genetics, University of Washington, Seattle.

ai, A.C.K., Chambers, T.M., Holland Jr., R.E., Morley, P.S., Haines, D.M.,Townsend, H.G.G., Barrandeguy, M., 2001. Diverged evolution of recentequine-2 influenza (H3N8) viruses in the Western Hemisphere. Arch. Virol.146, 1063–1074.

anuguerra, J.-C., Zientara, S., Sailleau, C., Rousseaux, C., Gicquel, B., Rijks,

I., van der Werf, S., 2000. Evidence for evolutionary stasis and genetic driftby genetic analysis of two influenza H3 viruses isolated in France. Vet.Microbiol. 74, 59–70.

artella, V., Elia, G., Decaro, N., Di Trani, L., Lorusso, E., Campolo, M.,Desario, C., Parisi, A., Cavaliere, N., Buonavoglia, C., 2007. An outbreak

Rese

N

O

O

O

O

A.M. Damiani et al. / Virus

of equine influenza virus in vaccinated horses in Italy is due to an H3N8strain closely related to recent North American representatives of the Floridasub-lineage. Vet. Microbiol. 121, 56–63.

ewton, J.R., Daly, J.M., Spencer, L., Mumford, J.A., 2006. Description of theoutbreak of equine influenza (H3N8) in the United Kingdom in 2003, dur-ing which recently vaccinated horses in Newmarket developed respiratorydisease. Vet. Rec. 158, 185–192.

IE, 2004. Equine influenza. In: Manual of Diagnostic Tests and Vaccinesfor Terrestrial Animals, part 2, section 2.5, chapter 2.5.5, http://www.oie.int/eng/normes/mmanual/A 00083.htm.

IE, 2006. Conclusions and recommendations from the Expert SurveillancePanel on Equine Influenza Vaccines. Office International des Epizooties

O

P

arch 131 (2008) 100–105 105

Bulletin no 2006-2, pp. 35–36, http://www.oie/eng/publicat/en bulletins.htm.

xburgh, L., Berg, M., Klingeborn, B., Emmoth, E., Linne, T., 1994. Evolu-tion of H3N8 equine influenza virus from 1963 to 1991. Virus Res. 34,153–165.

xburgh, L., Hagstrom, A.A., 1999. A PCR based method for the identifica-tion of equine influenza virus from clinical samples. Vet. Microbiol. 67,

161–174.

xburgh, L., Klingeborn, B., 1999. Cocirculation of two distinct lineages ofequine influenza virus subtype H3N8. J. Clin. Microbiol. 37, 3005–3009.

age, R.D.M., 1996. Treeview: an application to display phylogenetic trees onpersonal computers. Comput. Appl. Biosci. 12, 357–358.