Search

cDNA copies of the M and N genes of equine arteritis virus (EAV) isolates were synthesized by reverse transcription followed by polymerase chain reaction amplification. The cDNA was subjected to a cycle sequencing strategy using Taq polymerase, and the nucleotide and derived amino acid sequences of 10 virus isolates were compared. The M and N genes of all isolates had the same initiation and termination sites as the prototype Bucyrus strain and the encoded proteins were conserved between viruses. Comparison of nucleotide sequence homologies and phylogenetic tree analysis implied the existence of three EAV variants originating from the U.S.A. (Bucyrus), Austria (Vienna) and Switzerland (Bibuna), and suggested that RNA recombination between EAV isolates may have occurred.

Equine influenza vaccines containing inactivated whole virus and Carbomer adjuvant stimulated higher levels and longer lasting antibody to haemagglutinin in ponies than vaccines of equivalent antigenic content containing aluminium phosphate adjuvants. Five months after the third dose of vaccine containing Carbomer adjuvant, ponies were protected against clinical disease induced by an aerosol of virulent influenza virus (A/equine/Newmarket/79, H3N8). In contrast ponies which received vaccine containing aluminium phosphate adjuvant were susceptible to infection and disease. There was an inverse correlation between prechallenge levels of antibody detected by single radial haemolysis (SRH) and duration of virus excretion, pyrexia and coughing. All ponies with antibody levels equivalent to SRH zones of > or = 154 mm2 were protected against infection and all those with levels > or = 85 mm2 were protected from disease.

Antigenic differences between three of six equine influenza virus (H3N8) MDCK cell- and egg-derived pairs have been demonstrated using monoclonal and polyclonal antibodies. Sequencing of the haemagglutinin (HA) genes revealed amino acid changes in four of the six virus pairs. These data contrast with those for human isolates of influenza virus in that it was predominantly tissue culture-isolated equine virus and not egg-derived virus which displayed heterogeneity. Some of the molecular changes involved are located within the vicinity of the cell receptor-binding site (positions 156, 158 and 222) whereas others are in the vicinity of the HA1-HA2 cleavage site (positions 18 and 32 of HA1 and position 12 of HA2). Our results indicate that the host cell can play a part in selecting antigenic variants of equine influenza virus and suggest that the egg, and not cell culture as is the case for human isolates, is the preferred host for vaccine and antigenic studies.

Seven previously untreated five-month-old New Forest ponies received two doses of equine influenza immunostimulating complex vaccines, one with and one without an immunopurified tetanus toxoid component, given by deep intramuscular injection six weeks apart, followed by a booster dose without tetanus toxoid five months later. Fifteen months after the third dose of vaccine, the ponies were challenged by exposure to an aerosol of influenza A/Equine 2/Sussex/89 (H3N8), a virus isolated from a recent outbreak of influenza A/equine 2 in Britain. The challenge produced severe clinical signs of influenza (pyrexia and coughing) in five unvaccinated control ponies. Four of the vaccinated ponies were completely protected against clinical disease, and two of these were also protected against infection as demonstrated by their lack of an antibody response after challenge. No coughing was recorded among the vaccinated ponies, and only three of the seven vaccinated ponies experienced a transient mild pyrexia. The mean duration and severity of the pyrexia among the vaccinated ponies was significantly less (P < 0.01) than among the controls, and the excretion of virus was almost eliminated, thus demonstrating the protective efficacy of the vaccines 15 months after vaccination. Monitoring of tetanus antitoxin antibodies showed that protective levels (> or = 0.01/iu/ml) were maintained for at least 20 months after vaccination.

Welsh Mountain ponies were inoculated with an isolate of Streptococcus pneumoniae, SPE 1618 (capsular type 3) recovered from the equine respiratory tract: 10 ml of a suspension of 10(8) or 10(9) cfu/ml were instilled intratracheally. Fever was observed after either dose but the greater concentration also produced coughing, ocular and nasal discharge, depression and enlargement of submandibular lymph nodes. Cytological evidence of infection was also observed in tracheal washings during the first week after inoculation and corresponded with isolation of S. pneumoniae from the washes. Morbid anatomical and histopathological examinations of selected animals revealed focal pneumonia affecting the ventral lung, especially the cardiac area and accessory lobe, with a propensity to affect the right lung. S. pneumoniae was isolated directly in pure culture from these lesions or was demonstrable by immunostaining of macrophages bearing specific capsular type 3 antigen. By 10 days after inoculation, the ponies were healthy and had developed antibodies to S. pneumoniae. S. pneumoniae was therefore a primary pathogen in the horse under the conditions of the challenge.

In July 1989 influenza A/equine-2 (H3N8) was isolated from a nasopharyngeal swab taken from a non-thoroughbred horse exhibiting acute clinical respiratory disease. This was the first isolation of equine influenza virus in the United Kingdom since 1981. Subsequent investigations of acute respiratory disease in horses indicated that the infection was dispersed throughout the UK. However, unlike the previous epidemic of 1979, the first horses from which the virus was isolated had been vaccinated. This outbreak of influenza provided an opportunity to evaluate an antigen capture ELISA, directed against the influenza virus nucleoprotein, as a rapid method for detecting virus in the nasopharyngeal secretions of naturally infected horses.

Seven hybrid cell lines of mouse myeloma cell line NSO and spleen cells of BALB/c mice producing monoclonal antibodies (MAbs) against equine influenza A/Equi-2/Ludhiana/87 (H3N8) virus were developed. These MAbs were purified, isotyped and characterised by enzyme linked immunosorbent assay (ELISA), fluorescent antibody test (FAT), haemagglutination inhibition (HI) and virus neutralization (VN) tests. The titres of ascitic fluids induced by hybridomas as estimated by ELISA ranged from 1:25,600 to 1:51,200. Monoclonality of these clones was confirmed using a panel of 5 viral antigens, each belonging to a single isotype. MAbs (5) belonged to IgM and one each to IgG1 and IgG2a. Two epitopes appeared to be closely resembling by HI and VN tests but other two epitopes appeared to be different.

An experimental (ISCOM) vaccine previously shown to protect hamsters from lethal challenge with equid herpesvirus-1 (EHV-1), was tested in horses. Vaccination with EHV-1 ISCOMs induced serum antibodies to the major virus glycoproteins gp10, 13, 14, 17, 18 and 21/22a, whereas antibody responses to gp2 were weak or absent. High levels of virus neutralising antibody of long duration were induced, but did not prevent challenge infection with virus of the homologous strain. However, in the vaccinated ponies there was a significant reduction in clinical signs, nasal virus excretion and cell associated viraemia compared with age-matched unvaccinated controls. There was a strong correlation between pre-challenge levels of serum virus neutralising antibody and the duration and total amount of virus excreted from the nasopharynx.

Equid herpesvirus-1 (EHV-1; Ab4 isolate) was inoculated unilaterally into the cavum vaginale of four pony colts under general anaesthesia. The animals were monitored daily for evidence of scrotal or testicular swelling and euthanased electively on days 3, 4, 6 and 12 after infection. Detailed pathological examination of the male genital tract was carried out. In animals examined at days 3 and 4 after infection, replication of EHV-1 was detected bilaterally in mesothelial and endothelial cells of the parietal and visceral vaginal tunics. The mesothelial infection had resolved by day 12 after infection, with no evidence of direct extension to deeper genital organs. None of the four colts showed significant scrotal or testicular swelling.

Two A/Equi-2 (H3N8) isolates were obtained during the 1987 Indian equine influenza epizootic. The sequence of the Ludhiana/87 HA1 gene revealed that this isolate was very similar to recent European and North American isolates of equine influenza. In contrast, the Bhiwani/87 HA1 gene was nearly identical to the Miami/63 prototype H3 sequence. These results support the antigenic analysis previously carried out on these isolates using monoclonal antibodies. However, the finding that Bhiwani/87 is so similar to Miami/63, coupled with the finding that equine H3N8 influenza viruses have previously been shown to evolve along a single lineage, suggests that Miami/63 virus from either a vaccine or laboratory source may be the origin of the Bhiwani/87 isolate.

The N-terminal fragment comprising residues +1 to +50 (gB1-50) of equine herpesvirus type 1 (EHV-1) glycoprotein B was expressed as a glutathione S-transferase fusion protein in Escherichia coli. Recombinant gB1-50 (rgB1-50) was recognized in immunoblots by sera from rabbits immunized with EHV-1 and by convalescent-phase sera from horses with natural EHV-1 infections. An enzyme-linked immunosorbent assay (ELISA) for monitoring antibody levels against EHV-1 was developed by using rgB1-50, and its specificity was assessed with a panel of reference antisera against other equine viruses. A specific cross-reaction was detected with EHV-4, which was confirmed by inhibition ELISA. Convalescent-phase sera from horses with natural EHV-1 or EHV-4 infections possessed antibody titers against rgB1-50 ranging from 1:2,000 to 1:64,000, indicating the presence of an immunodominant antigenic site. The study demonstrated the potential application of rgB1-50 as a diagnostic antigen and highlights the glutathione S-transferase fusion system as a simple and effective method of producing purified milligram quantities of antigen.

Six monoclonal antibodies directed against EHV-1 glycoprotein 13 were characterized. Five antibodies neutralized EHV-1 and were directed against a single antigenic site which comprised type-specific and type cross-reactive epitopes. Inhibition of monoclonal antibody binding to this site by post-infection equine sera suggests that it is a target of host antibody during natural infection with either EHV-1 or EHV-4.

The haemagglutinin (HA) gene from the equine influenza H3N8 isolate Suffolk/89 has been cloned by reverse transcription and polymerase chain reaction amplification. The nucleotide sequence of the HA gene was determined from two independently cloned copies of the gene and was found to be most closely related to recent American isolates supporting the idea that most isolates of equine H3N8 are evolving as a single lineage. When the predicted amino acid sequence of the Suffolk/89 HA was examined, changes had taken place in at least four of the major antigenic sites, A, B, C, and D when compared to the sequences of the isolates used in the current vaccines (Miami/63 and Fontainebleau/79). Surprisingly, when the Suffolk/89 isolate was tested in haemagglutination inhibition (HI) assays with a panel of six mouse monoclonal antibodies, no differences were observed between the Suffolk/89 and the Fontainebleau/79 isolates, suggesting that this panel of monoclonal antibodies may recognise a limited subset of the major antigenic sites. Three anti-HA horse heterohybridoma monoclonals were able to distinguish between the Suffolk/89 and Fontainebleau/79 viruses, demonstrating that the horse does recognise these isolates as being antigenically different. The results of the work suggest that the isolates used in current equine influenza vaccines may need updating.

Twelve Welsh Mountain pony mares in late gestation were infected intranasally with EHV-1 (AB4 isolate) at dose rates from 10(3) to 10(7.3) TCID50. This resulted in 3 cases of paresis, at Days 9, 10 and 12 after inoculation, and 5 abortions, at Days 6, 9, 18, 19 and 20. Euthanasia was performed between Days 6 and 21, with collection of uterine specimens for histopathology, virus isolation and immunoperoxidase staining from the pregnant horn, non-pregnant horn and body. EHV-1 replication in endometrial vessels was detected as early as Day 6 and was maximal at Days 9-11, when widespread thromboischaemic damage was present. By Days 15-19 in mares remaining pregnant, EHV-1 antigen expression in the endometrium was sparse, despite residual lesions but little associated thrombosis. Endometrial vascular pathology varied considerably in degree and extent, and no consistent predilection sites for replication within the uterus were apparent.

An antigen capture enzyme-linked immunosorbent assay (ELISA) was developed for the detection of equine herpesvirus type-1 (EHV-1) antigens in nasal swab specimens. The test was designed as a solid phase, amplified sandwich assay in which an EHV-1 specific monoclonal antibody was used to capture virus antigen and polyclonal antisera used to detect antigen bound to the test plates. Eight monoclonal antibodies were tested for their ability to capture virus antigen and one was selected for routine use. The sensitivity and specificity of the ELISA was compared with that of virus isolation using swabs from ponies which were experimentally infected with EHV-1. Of 72 nasal swabs collected, 32 were found to be positive by both virus isolation (VI) and ELISA, a further 15 samples were positive by VI alone, but none of the samples were positive by ELISA and negative by VI. This yielded an overall assay sensitivity of 68% and specificity of 100%. The assay proved useful for diagnosis since virus antigen was detected during the first four days post-infection which corresponded to the acute phase of disease when some clinical symptoms were apparent. In addition, the assay could be completed within one day when antibody coated plates were available.

From 1988 to 1991, 51 pregnant pony mares were challenged intranasally or by aerosol with an isolate of EHV-1 (AB4) originally recovered from a quadriplegic mare. This resulted in 32 abortions, occurring from 9 to 29 days after infection. In 14 of the early abortions (Days 9-14), EHV-1 was not demonstrated in the foetal tissues by virus isolation or immunostaining despite no other non-viral cause for the abortion being evident. Application of the polymerase chain reaction to foetal tissues from 9 of these cases also proved negative. One of the 14 mares was destroyed immediately after abortion, and post-mortem examination revealed severe and widespread vasculitis, thrombosis and secondary ischaemic damage in the endometrium with replication of EHV-1 in endothelial cells. These findings suggest that EHV-1 abortion can occur due to endometrial damage without the establishment of a foetal infection.

Studies were carried out to determine the optimum conditions for the production of equine monoclonal antibodies (MAbs). Lymphocytes from ponies immunised with influenza A equine 2 virus, isolate A/Equine/Newmarket/79 (H3N8) were fused with mouse myeloma (NSO) cells and with horse-mouse heterohybridomas made aminopterin-sensitive by selective growth in 8-azaguanine. Although all fusions initially resulted in heterohybridoma colonies that secreted equine immunoglobulin, many of these were unable to maintain secretion for longer than a few weeks. Increasing the time between immunisation and the booster injection of Newmarket/79 virus, the inclusion of Freund's incomplete adjuvant and the use of an aminopterin-sensitive primary heterohybridoma as the fusion partner, improved the production of HIg-secreting heterohybridomas. After two clonings eight cell lines were established which maintained anti-Newmarket/79 antibody secretion for over a year. FACS analysis of the cell lines provided a useful means of predicting breakdown of MAb secretion by the cell lines, thus enabling re-cloning to be carried out in time.

Shortcomings in the original methods (based on haemagglutination of erythrocytes) used to measure potency of equine influenza vaccines and antibody responses stimulated by vaccines, coupled with the lack of a reliable challenge system in the target species, has hindered progress in identifying the antigenic content required to provide protection. Reliable methods are now available for measuring the haemagglutinin (HA) content of vaccines and the antibody responses they elicit. The development of challenge systems in the target species has allowed antibody levels consistent with protection to be identified under experimental conditions. Comparison of vaccine performance under experimental conditions and in the field suggests that exposure to a nebulised virus aerosol provides a realistic challenge for the assessment of equine influenza vaccines.

An experimental ISCOM vaccine has been prepared from gradient purified equine herpes virus type 1 (EHV-1). Radiolabelling studies demonstrated that this vaccine contained all the major viral glycoproteins in relative amounts similar to those found in non-detergent disrupted viral preparations. This EHV-1 ISCOM vaccine generated fully protective responses in hamsters challenged with an otherwise lethal dose of the hamster-adapted EHV-1 strain RACH.

Infection of seronegative Welsh mountain ponies was established by intranasal instillation or exposure to nebulised aerosol of egg grown H3N8 viruses. Pyrexia and coughing were noted following intranasal instillation and high titres of virus were recovered from the nasopharynx. Exposure to aerosol resulted in more severe clinical signs characterised by high temperatures, dyspnoea, anorexia and coughing; lower levels of virus were recovered from the nasopharynx. The severity of clinical signs and the kinetics of virus shedding were dose-related with the minimal infectious dose being 10(2)EID50/ml when ponies were exposed to aerosols produced by nebulisation of 20ml allantoic fluid. Full clinical signs only developed when ponies were exposed to a dose of 10(6)EID50/ml. It was concluded that exposure to nebulised aerosols of egg grown H3N8 viruses was a more reliable method of inducing clinical influenza than intranasal inoculation and would be more suitable for challenge studies.

Equine influenza type 2 infections occurred in the Newmarket areas in January 1976. The disease did not spread to any extent and while this may have been due to recent vaccination it was found that not all vaccinated horses were fully protected. The virus involved showed some antigenic drift from the prototype strain A/equine/Miami/1/63 (Heq 2 Neq 2).

Subtype-1 isolates of Equine herpesvirus-1 (EHV-1) from a quadriplegic horse and from an aborted foetus were compared with each other and with a subtype-2 respiratory isolate. All 3 isolates were detected in the epithelium and macrophages of the respiratory tract. Both the paresis and foetal subtype-1 isolates replicated in the epithelium of the ileum and this correlated with the recovery of virus from faeces in vivo. The paresis subtype-1 isolate also had a predelection for vascular endothelial cells, particularly in the nasal mucosa, but also in the lungs, central nervous system, adrenal and thyroid. In the 9 foals inoculated with the paresis isolate two developed hind limb dysfunction, four developed diarrhoea, and one of these 4 died with an intussusception. The differences between these isolates are discussed in relation to other herpesviruses.

Serum antibody (IgGab, IgM and IgA) responses to primary and secondary infection with influenza A/equine/Newmarket/79 (H3N8) by nebulised aerosol were compared with local (nasopharyngeal and tracheal) antibody responses in ponies. Circulating IgGab antibody was of long duration after primary infection, whereas IgM responses were short-lived after both primary and secondary infections. The antigenic stimulation of secondary infection with equine influenza was sufficient to induce elevations of serum IgM and IgA in the presence of high levels of circulating IgGab. These results support the potential of virus-specific IgM measurement for the detection of recent exposure to virus in horses which have high levels of circulating IgGab. Unlike serum IgGab, nasal and tracheal wash antibody of this isotype did not show long duration after primary infection, but local antibody memory was demonstrated by anamnestic responses on rechallenge. Nasopharyngeal IgA developed later than IgGab and IgM, and was more durable.

Racehorses perform badly for many different reasons. Trainers often expect clinicians to determine the cause in individual cases and, more especially, where most of the immates of the stable are apparently affected by loss of form. Clinical examinations may reveal signs including fever, serous nasal discharge and the occasional cough. Haematology and blood biochemistry are commonly used aids to diagnosis in the field and may be helpful, but there is a need for facilities for virological investigations to be made readily available for use by clinicans as an adjunct to more commonplace laboratory techniques. This paper presents the background to a serious and widely publicised problem experienced by racing stables in the UK in recent years and, in discussing its epidemiology, lays emphasis on upper respiratory tract (URT) disease caused by viruses. It is suggested that, if the incidence of URT disease could be reduced, there would be a corresponding diminution of the "poor performance" syndrome. The inter-relationship of viral diagnosis, epidemiology and research is discussed in terms of methology, interpreting results and limits of present day knowledge.

An extensive outbreak of EHV-1 abortions occurred on a stud farm in England in 1985. Of the 67 pregnant mares present on the stud farm, 31 were challenged with EHV-1, resulting in the loss of 22 fetuses or foals. Laboratory investigations revealed that the spread of the virus closely followed movement of apparently healthy mares (during the incubation period of the infection). During the outbreak mares were challenged 1-4 months before the expected foaling date. There was no relationship between the gestational age at the time of challenge and the subsequent outcome of infection in terms of abortion. The period between the estimated date of infection and abortion varied between 9 days and 4 months. Virus was isolated from the nasopharynx of 3 apparently healthy foals born during the outbreak.

Serological responses following two and three doses of an inactivated equid herpesvirus 1 ( EHV -1) vaccine containing a subtype 1 strain were examined in yearling ponies. Complement fixing antibody responses were significantly higher against the subtype 1 vaccine strain than against a subtype 2 virus. Complement fixing antibody responses declined rapidly after the second dose of vaccine and had returned to almost pre-vaccination levels eight weeks after the second dose of vaccine. Complement fixing antibody titres to the heterologous subtype 2 strain increased after each successive dose of vaccine. The neutralising antibody responses following vaccination were weak although less strain specific than the complement fixing antibody responses. When challenged with a subtype 2 strain, 15 ponies which had received three doses of vaccine, two, eight and 12 weeks previously resisted challenge infection. However, 11 out of 15 ponies which had received two doses of vaccine eight and 12 weeks previously were susceptible to infection. While two doses of vaccine did not reduce duration of virus excretion or febrile responses in ponies challenged eight weeks after the second dose, the amount of virus excreted was significantly reduced.

An investigation was made of 3 serological tests (virus neutralization, complement fixation and indirect immunofluorescence), which are applicable to epidemiological studies of infections by Equid herpesvirus 1 (EHV-1). Sera from gnotobiotic foals inoculated intranasally with various strains of EHV-1 were unable in some cases to neutralize heterologous strains and these results were not consistent with the existence of clearly-defined subtypes of EHV-1, as previously proposed. The cross-reactions in complement-fixation tests paralleled those with neutralization but immunofluorescence tests were found to be both more sensitive and more broadly reactive than the other two. Complement-fixing antibodies declined more rapidly following experimental infection than did those measured by neutralization or immunofluorescence. The results are discussed in relation to the diagnosis of EHV-1 infection and the significance they may have for the epidemiology of this disease.

Seven monoclonal antibodies (MAbs) were produced which recognized equid herpesvirus type 1 (EHV-1). Three MAbs neutralized the subtype 1 virus (strain Army 183) in the presence of complement, but did not neutralize the subtype 2 virus (strain MD). All three MAbs immunoprecipitated an Mr 83K glycoprotein from a detergent-solubilized virion envelope preparation of the subtype 1 virus. The target antigens of the four non-neutralizing MAbs (6F11, 2A4, 1F10 and 8D9) were identified by immunoblotting against purified EHV-1 virions and had respective apparent Mr values of greater than 205K, greater than 205K, 97K and 13K.

An enzyme-linked immunosorbent assay (ELISA) was developed for the detection of serum antibodies to equine arteritis virus (EAV). Results from this assay produced a good correlation with results from virus neutralisation tests in horses which had not been regularly vaccinated with commercially available mammalian tissue culture-derived viral vaccines. Vaccination of some horses with tissue culture-derived vaccines induced the formation of antibodies to bovine serum. These antibodies reacted with the bovine protein contaminants in the EAV ELISA antigen, producing false-positive results. Non-viral protein contaminants were found to be closely associated with EAV in that they co-purified with the virus during gradient centrifugation.

Cytotoxic cell precursors and/or cytotoxic memory cells were demonstrated in the peripheral blood of ponies after aerosol infection with influenza A/equine/Newmarket/79 (H3N8). In order to reveal their cytotoxic potential, peripheral blood mononuclear cells required a secondary antigenic stimulation. In vitro induced cytotoxic cells showed activity against influenza infected target cells in a 3-4 h 51Cr-release assay. The reactivity of cytotoxic cells was markedly influenced by the conditions of the secondary induction culture. If high concentrations of exogenous crude equine IL-2 were used, virus infected target cells were susceptible to lysis by autologous or allogeneic effector cells. However, if IL-2 concentration was reduced, cytotoxic cells were generated which showed features consistent with cytotoxic T cells in that target-cell killing was genetically restricted.

The duration of immunity as measured by virological, serological and clinical responses following infection with influenza A/equine/Newmarket/79 (H3N8) was assessed in repeated challenge experiments in which ponies were infected by exposure to aerosols of infectious virus. Previous infection stimulated complete clinical protection which persisted for at least 32 weeks as demonstrated by the absence of febrile responses and coughing in two groups of ponies infected 16 weeks or 32 weeks after the first infection. Partial clinical protection persisted for over a year as demonstrated by the absence of coughing and a reduction in the number of febrile responses in a group of ponies infected 62 weeks after their first infection. These results contrasted with those observed in immunologically naive control ponies which developed pyrexia, dyspnoea and nasal discharge and coughing. The kinetics of virus specific antibody production in primary and secondary infections with equine influenza were studied by the single radial haemolysis test and a radioisotopic antiglobulin binding assay which measured virus specific IgGab antibody isotype. Antibody to the haemagglutinin, as measured by the single radial haemolysis test, declined rapidly after primary infection whereas the IgGab responses to whole virus antigens persisted for longer. The single radial haemolysis test was therefore particularly useful for the detection of antibody responses in multiple infections or exposures to influenza antigens. The radioisotopic antiglobulin binding assay was more sensitive for identifying infections which had occurred more than six months previously, as evidenced by anamnestic IgGab responses in ponies with low levels of antibody before rechallenge.

An antigen capture indirect enzyme linked immunosorbent assay (ELISA) was developed to detect influenza nucleoprotein antigen in nasal secretions from horses infected with A/equine/H3N8 viruses. Results from this assay were compared with conventional virus isolation in embryonated hens eggs.

Thirty-one ponies immunized with inactivated virus vaccine containing A/equine/Miami/63 (H3N8) virus and six seronegative ponies were experimentally challenged with the homologous virus strain. All 6 unvaccinated ponies and 11 out of 31 vaccinated ponies became infected. A clear relationship between pre-challenge antibody, measured by single radial haemolysis (SRH), and protection was demonstrated as judged by virus excretion, febrile responses and antibody responses. Those ponies with SRH antibody levels greater than 74 mm2 were completely protected against challenge infection by the intranasal route.