Epidemiology, surveillance, and zoonotic aspects of human influenza.

Department of Virology, Leiden University Medical Centre, Leiden, The Netherlands

Human influenza viruses annually cause minor or major epidemics with considerable morbidity and mortality. Despite an elaborate WHO network of national influenza centres all over the world and extensive vaccination programmes in many countries, the virus still cannot be controlled. The WHO network provides us with the most recent information on the changes in the circulating influenza viruses, but protection provided by the vaccine will never be complete. One reason is that mutations rapidly accumulate in circulating influenza viruses (antigenic drift) and new variants emerge while the vaccine is being produced. Another reason is that the majority of the mortality by influenza is observed in the elderly and especially in this group of people the immunological response to the influenza vaccine is sub-optimal.

In principle, human influenza is a zoonosis. An influenza pandemic may occur as a result of the introduction of a new subtype of influenza A virus into the human population. At least parts of the segmented genome of the viruses that are being introduced in an "antigenic shift" are of animal origin. If such a virus manages to establish itself in the human population, drift variants of these viruses cause epidemics in the years to follow and they become human influenza viruses. In some cases, direct transmission of influenza viruses from animals may cause disease in humans. Most reports on interspecies transmission to humans are dealing with swine influenza viruses. But, as has been shown in the influenza A (H5N1) virus outbreak in Hong Kong in 1997, poultry can be the source of viruses that are transmitted directly to humans as well. However, in all these cases the viruses did not manage to spread efficiently amongst the human population. Interspecies transmission of animal influenza viruses to humans probably is a more frequent event than we know. Studies in human sera from rural China show antibodies to a variety of non-human influenza virus subtypes. In addition, the prevalence of antibodies to the avian-like swine influenza A (H1N1) viruses seem to increase with a higher exposure to pigs. This means that many asymptomatic interspecies transmissions from animal influenza viruses to humans may occur. If this happens during human influenza epidemics, a reassortant virus with pandemic potential may be generated. The influenza A (H2N2) and influenza A (H3N2) viruses that caused the 1957 and 1968 pandemics, both were reassortant viruses and contained both human- and avian-like gene segments. Therefore, human influenza surveillance should not only focus on the identification of antigenic drift variants of human influenza viruses. More attention should be paid to the frequency of transmission of animal influenza viruses to humans.

Department of Microbiology, Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China

Influenza is a natural infection of certain animals particularly aquatic birds in which the virus lives asymptomatically in the intestine. In contrast, the virus can cause severe infections of economic importance in domestic, land-based birds. Humans seemingly are irrelevant to the virus's survival, yet irregularly occurring zoonotic transmission is thought to give rise to human influenza pandemics. All pandemics including the H3N2 (Hong Kong) pandemic of 1968 have occurred without warning. After that pandemic, it was the aspiration to have in place an early warning system that would recognize an incipient pandemic in humans and so blunt its effect. A higher aspiration was to recognize the virus in animals, the hypothetical source of the virus, before it appeared in humans.

The occurrence of influenza viruses in certain animals, agricultural practices that facilitate transmission of influenza viruses between animals and humans, the origin of the H3N2 and the H2N2 (Asian) pandemic virus of 1957 and historical records identified in 1982 southern China as a hypothetical influenza epicentre in which this event would most likely occur.

In August 1997, influenza H5N1 virus was isolated from a child who died in May on the heels of a H5N1 influenza outbreak of high mortality in chicken in Hong Kong. In November and December, a further 17 cases of H5N1 influenza of which 5 were fatal, occurred in apparently haphazard fashion suggesting avian-to-human rather than human-to-human transmission. Virus surveillance of a wide range of animals indicated major market poultry as the source of the virus; chicken, of which about one-fifth were infected, and less so ducks and geese, giving rise to an excessive H5N1 virus 'load' upon the Hong Kong community.

Genetic studies indicate that the H5N1 virus is a reassortant 'avian' influenza virus that was rapidly evolving in Hong Kong poultry markets.

It possessed 'human specific' consensus amino acid sequences previously found in internal proteins of human strains which may be important in zoonotic transmission. This information vindicates on genetic grounds the slaughter of chicken and other poultry as the source of virus for humans.

During the H5N1 incident, H9N2 viruses were isolated from duck, geese, pigeon, quail, chicken (4.9%) and market environmental swabs. The virus appears to be widespread in chicken in Asia and may have been the donor of the 'internal' genes of the virus responsible for the H5N1 incident in Hong Kong in 1997.

Hong Kong recognized the signs of an incipient influenza pandemic in humans and its immediate source in chicken. It reacted vigorously and timely and may have averted a full-blown pandemic. This does not preclude the possibility that the virus is still present in a non-human host in the region or beyond.

In February 1999, H5N1 virus was isolated from the faeces of geese imported into Hong Kong. Its H5 gene was almost identical to that of the 1997 H5 virus, the other seven genes being more closely related to those of avian influenza viruses of the 70's and 80's.

The isolation of H5N1 viruses in 1999 upholds the policy of having segregated wholesale poultry markets in Hong Kong, one for chicken and other land-based birds, the other for ducks and geese which are killed inside the market.

The intriguing question is - why was the 1997 H5N1 virus in Hong Kong pathogenic for humans as well as chicken while elsewhere H5 viruses are apparently pathogenic for land-based poultry only?

Genetic analysis of host-specificity determinants within the transcription/replication complex of influenza A viruses

N. Naffakh, P. Massin, C. Quémin, N. Escriou, B. Crescenzo-Chaigne and S. van der Werf

Unit of Molecular Genetics of Respiratory Viruses, URA 1966 CNRS, Institut Pasteur, Paris, France

Introduction

Generally, avian influenza viruses do not replicate efficiently or cause disease in humans. However, the pandemic viruses isolated in 1957 and 1968 were found to derive their gene segments from both human and avian viruses. Avian virus segments encoded not only the surface glycoproteins (NA and/or HA) but also the RNA polymerase PB1. More recently, human infection with an avian influenza virus (A/HongKong/156/97) was reported in Hong-Kong (1,2), thus stressing the need for a better understanding of the molecular basis of host specificity among influenza viruses. We addressed the question of wether the proteins involved in the replication process were determinant in host-specificity, by making use of the genetic system for the in vivo reconstitution of active synthetic influenza virus-like ribonucleoproteins described by Pleshka et al. (3).

Material and methods

Molecular cloning of cDNAs: Viral RNA was extracted from the A/FPV/Rostock/34 (Rostock), A/Mallard/NY/78 (Mallard), and A/HongKong/156/97 (HK) viral isolates grown in 11 day-old embryonated chicken eggs (Rostock, Mallard) or on MDCK cells (HK). Full-length cDNAs were prepared by reverse transcription of viral RNA using a primer complementary to the 12bp conserved at the 3' end of each viral segment, and subsequently amplified using primers specific of each PB1, PB2, PA or NP gene. Double stranded cDNAs were then cloned into the EcoRV site of the pHMG plasmid vector (3), and their sequence was determined. The cDNAs derived from A/PR/8/34 (PR8) and A/Victoria/3/75 (Victoria) viruses were kindly provided by J. Pavlovic and A. Portela, respectively.

Transient replication assay: The pPOLI-CAT-RT plasmid (3), designed to allow expression of an artificial viral RNA-like transcript, was kindly provided by P. Palese. A mixture of pHMG-PB1, -PB2, -PA, -NP (1;1;1;2mg) and pPOLI-CAT-RT (1mg) plasmids was transfected into COS-1 cells, using the Fugen-mediated method (Boehringer Mannheim). After 48hrs of incubation at 37°C, cells were harvested and cell extracts tested for CAT activity, using the CAT ELISA Kit (Boehringer Mannheim).

Domains of PB2 conferring plaquing ability for an avian influenza virus in mammalian cell lines.

Y. Yongxiu*, J. McCauley^# , J.W. Almond* and W.S. Barclay*.

*School of Animal and Microbial Sciences, University of Reading and ^#Institute of Animal Health, Compton, U.K.

Birds as reservoirs and disseminators of influenza virus genes: an ecologic approach

C. Hannoun

Institut Pasteur , Paris, 18 Villa Prévost, 92120, Montrouge, France

Hemagglutinating viruses have been isolated by egg inoculation from several animal species: swine, horse, birds and were later found to be similar to human influenza viruses. Surface proteins are similar to those of human viruses and they can be classified in the same serotypes: H and N, whatever is the origin of the strains. This does not mean that these viruses are exactly similar: there are differences in physical properties (resistance to heat and to pH), or in their pathogenic properties (they are usually very species specific). From 1971, a universal classification system has been utilized: 15 HA and 9 NA. A virus is described by a double serotype: HxNx. In addition, isolates of the same group of viruses were obtained from other species, whether from healthy or sick animals (whale, mink, seal, dog, cat,...).

Birds play a very special role

- They host all known serotypes when other species have only a few.

- They can be highly pathogenic, but most of them are not.

- Bird viruses multiply in the intestine as well as in the respiratory tract and they are transmitted also by fecal-oral route.

A survey was undertaken since 1976 in order to investigate the presence and the circulation of influenza viruszes in wild bird populations in the North of France

Wild birds, mostly ducks and other water birds were captured by a team of experienced ornithologists in the frame of a programme of study of migrations. The ornithological reserve, located in the estuary of the Somme river hosts a mixture of resident and migratory birds belongind to several species. Cloacal swab samples were taken to birds which were individually identified to be recognized if recaptured later. Samples were put in transport medium containing antibiotics and brought to the laboratory for inoculation to eggs for the isolation of virus. After proper incubation, harvested allantoic fluids were examined for the presence of hemagglutination and positive samples were subpassaged and the agent was identified by hemagglutination and neuraminidase inhibition (1) using reference sera for all types of influenza viruses.

Studies in Northern France since 1976 have produced a large number of isolations in a single place over the years (2). In the present survey, 176 strains of hemagglutinating viruses were obtained, including 93 influenza viruses, 50 parainfluenza and 33 non fully identified influenza strains. The isolates (Table 1) include a wide range of serotypes (8 out of 15 HA and 6 out of 9 NA), especially from ducks and other water fowl. The serotypes are not always the same year after year, even from samples taken on the same day: some are dominant during one season and can disappear the next. The infected birds are migratory or sedentary healthy animals belonging to several species, with a majority of mallards, shellducks and pintails.

An interesting finding was the occurrence of strains difficult to identify and for a few of them, cloning of the viral population was performed by plaque selection. Lysis plaques on MDCK cells, on which these viruses grow easily, were picked individually from early passages and each clone was re-identified. The results with a characteristic isolate is shown in Table 1. Out of 34 clones, the identification of hemagglutinin and neuraminidase serotypes was distributed into 4 categories associating H6, H7, N2 and N4 in all 4 possible combinations. This isolate contained a mixture of two different viruses and the result is a reassortment similar as those that can be obtained experimentally by inoculation of a mixture of viruses.

Table 2. 34 clones from the isolate influenza MB81

Discussion and conclusions

The results show

1- A high rate of isolation, meaning a high incidence of feral state, up to 12% in wild animals of different species and 60% in semi-captive mallards. In the latter case, virus was also isolated repeatedly from the pond in which the ducks were living.

2- However, the presence of virus is related to the time of the year, the highest rates being obtained shortly after the return from northern migration in October. Afterwards, the incidence decreases and no isolations were obtained after January.

3- The infected animals are quite healthy and were recaptured in the following weeks. Some remain positive during 15 days at least.

4- There are possibilities of recombinations since in two instances at least, mixtures of viruses were obtained from the same sample and all combinations of serotypes could be demonstrated in the mixture.

In conclusion, these observations confirm and reinforce the concept of the role of birds as a permanent reservoir of genes maintaining the virus for long periods of time. They can transport these genes on long distances and contribute to the dissemination of influenza before transspecies reassortments and eventual transmission to man. This is an important aspect of the ecology of influenza viruses and has important consequences on the human epidemiology since the transfer of influenza virus genes from avian strains to human strains is generally recognized to be the cause of the appearance of pandemic strains in human populations (3).

(1) Fiszon,B. & Hannoun C.(1990), Comparison of neuraminidases of the same type but different species using a new method of titration.. J.Virol.Meth., 27, 79-90.

(2) Hannoun,C. & Devaux,J.M.(1980), Circulation enzootique permanente de virus grippaux dans la baie de la Somme. Comp.Immun.Microbiol.. Infect.Dis., 3, 177-183.

(3) Webster, R.G., Hinshaw V.S., Bean, W.J., Turner,B. & Shortridge,K.F. (i977), Influenza viruses from avian and porcine sources and their possible rôle in the origin of human pandemic strains. Biol Standard. 39, 461-468.

L. Englund¹ and B. Klingeborn²

1Department of Small Animals, National Veterinary Institute, Box 7073, S-750 07 Uppsala (Sweden), ²Department of Virology, National Veterinary Institute, Biomedical Centre, Box 585, S-751 23 Uppsala (Sweden)

Avian influenza virus sometimes cross the species-barrier from birds to mammalian species, including man. Occasionally this infection causes clinical disease which increases the probability of virus detection. When the influenza virus crosses over to a domestic species previously unaffected by influenza-like disease it is likely that the "new disease" will come to the attention of a diagnostic laboratory. This is what happened in 1984 when a six-week outbreak of a severe respiratory disease caused a three-fold increase of the expected mortality rate in 33 mink farms (1).

Necropsies were performed on 46 mink which died following symptoms of severe respiratory distress. Lung tissue samples were submitted for histopathology, virology and bacteriology.

The influenza virus isolates from the field were serologically suptyped by HI and NI tests. Convalescent sera during the outbreak as well as 2,891 serum samples collected five months later were analysed by HI-test or ELISA for antibodies to A/mink/Sweden/84 virus.

The entire genomes of three H10N4 viruses, isolated from mink (Sweden, 1984), mallard, and fowl (both UK, 1985) were genetically compared by oligonucleotide mapping.

Experimental infection of mink was initially used to link the isolated influenza virus to the clinical symptoms and pathological lesions observed from the field outbreak. In a subsequent study, mink were infected intranasally with mink/84, mallard/85, fowl/85, or chicken/49 (2) to compare clinical symtoms, antibody response, and possible in-contact transmission. Experimental aerosol infections of mink, using mink/84 or chicken/49, were then used to compare in more detail the pathogenesis of the two virus infections. Measurements of daily weight gain and area density of pneumonia in lung sections were compared using the two-tailed unpaired Student’s t-test.

Interstitial pneumonia was observed in a majority of the necropsied mink from the field outbreak. Six antigenically identical isolates of influenza A virus of subtype H10N4 (mink/84) were recovered. Convalescent sera during the outbreak as well as more than 50% of the sera collected five months later in the affected farms were positive for antibodies to mink/84 virus, while no antibodies or virus could be detected in samples from unaffected farms. Interstingly, all convalescent sera were also positive for antibodies to the two concomitantly circulating human H3N2 strains (1, 3).

Oligonucleotide mapping of the whole genomes showed a very high (ª 98%) sequence homology between mink/84 and mallard/85 and fowl/85, respectively (4). Together with phylogenetic studies of nucleoprotein genes of avian and mammalian influenza viruses (5) this strongly indicates the direct avian origin of mink/84.

Following intranasal infection of mink, all three H10N4 isolates, i.e. mink/84, mallard/85 and fowl/85, caused respiratory disease, interstitial pneumonia, specific antibody production and were transmitted via contact infection. The most consistent disease signs were seen with mink/84. Chicken/49 (H10N7) did not cause clinical disease or contact infection, but induced antibody production and mild lung lesions (4).

Further comparison between mink/84 and chicken/49 revealed that the infections progressed in similar patterns over the first 24 hours post infection but from 48 hours post infection obvious differences were recorded. Whereas in mink infected with chicken/49 no disease signs were observed, the inflammatory pulmonary lesions remained focal and presence of viral antigen sparse, the mink infected with mink/84 showed severe signs of respiratory disease, with inflammatory lesions spreading throughout the lung and viral antigen present in substantial numbers of cells in the lung, nasal mucosa, and trachea. The subjective observations were supported by the facts that the mean daily weight gain was significantly lower (p<0.001), while the median area densities of pneumonia were significantly higher (p=0.043) for the mink/84 group than for the chicken/49 group (6).

Our studies show that mink is a species capable of picking up and transmitting certain avian influenza virus strains. Since farmed mink are also sometimes infected by human influenza virus strains, as shown in serological surveys, the possible role of mink as mixing vessels for strains of avian and human influenza must not be ignored.

In this case the avian influenza virus (A/mink/Sweden/84, H10N4) was highly pathogenic for mink as well as contagious, but did not remain in the mink population. One reason could be the production structure which caused the annual culling of >80% of the mink population for fur production some weeks after the influenza outbreak, saving only adult breeding stock, thus diminishing the population at risk and minimizing the population density.

These results show that respiratory disease in mammalian species should be analysed for influenza virus of avian origin. An increased awareness of influenza virus as a possible cause of disease in previously unaffected species is necessary to detect any cross-species infections, which could subsequently present a threat to the human population.

1. Klingeborn, B., Englund, L., Rott, R., Juntti, N., and Rockborn, G. (1985) An avian influenza virus killing a mammalian species - the mink. Arch. Virol., 86:347-351.

2. Dinter, Z. (1949) Eine variante des virus des Geflügelpest in Bayern? Tierärtzl. Umchau, 4:185-186.

3. Englund, L., Klingeborn, B., and Mejerland, T. (1986) Avian influenza virus causing an outbreak of contageous interstitial pneumonia in mink. Acta vet. Scand., 27:497-504.

4. Berg, M., Englund, L., Abusugra, I.A., Klingeborn, B., and Linné, T. (1990) Close relationship between mink influenza (H10N4) and concomitantly circulating avian influenza viruses. Arch. Virol., 113:61-71.

5. Gammelin, M., Altmüller, A., Reinhardt, U., Mandler, J., Harley, V.R., Hudson, P.J., Fitch, W.M., and Scholtissek, C. (1990) Phylogenetic analysis of nucleoproteins suggests that human influenza A viruses emerged from a 19th-century avian ancestor. Mol. Biol. Evol., 7(2):194-200.

6. Englund, L. and Hård af Segerstad, C. (1998) Two avian H10 influenza A virus strains with different pathogenicity for mink (Mustela vison). Arch. Virol., 143:653-666.

L. Gérentes, M. Aymard, and N.Kessler

National Influenza Centre, Laboratory of virology, 8 Avenue Rockefeller 69373 Lyon cedex 08 France

The introduction of an avian A/H5N1 influenza virus (A/HK/156/97) in the human population in May 1997 in Hong-Kong, and the following identification of subsequent cases in November and December, was alarming ; indeed, the lack of protective immunity in the human beings against a new HA or a new NA can result in rapid spread of the virus, and it was reasonable to ask the question of how likely was a pandemic. Being faced with such a situation, we were interested to evaluate the serological status of the French population against the A/HK/156/97 strain

Viruses : Three different influenza viruses were used : A/PR/8/34 and A/Texas/36/91 viruses were produced by inoculation of embryonated hen’s eggs. A/HK/156/97 H5N1 strain was isolated and then produced in MDCK cells

Human sera :1) Fifty nine paired sera were collected during winter seasons 1994-1995, 1995-1996, in a home for the aged, 34 pairs (71-98 years) were obtained from elderly people (mean age 80 years) and 25 pairs from adult staff members (24-60 years). All of them were vaccinated annually from the start of their stay/work in the home, with a trivalent inactivated, containing A/Texas H1N1.

2) Eighty two sera were randomly collected from blood donors (36-69 years)

Due to the fact that A/H1N1 virus was responsible for two pandemics during the twentieth Century, (1918, 1934) all 200 sera we collected for this study, were ranged in groups of age corresponding to people that potentially met : (i) the two pandemics, 1918 and 1934 (76-98 years old people), (ii) the sole 1934 pandemic (60-75 years old people) and (iii) no pandemic (20-59 years old people)

Hemagglutinin inhibition test (HI) was done in microtiter plates as described by Palmer et al (1975) using hen red blood cells.

Neuraminidase inhibition test procedure (NI) was performed according to Aymard-Henry et al (1973) using fetuin as substrate.

Virus neutralisation test (Nt) were done in MDCK cells by Tobita et al (1975)

1)Antigenic characterisation of A/HK/156/97 H5N1 NA. The NA of A/HK/ H5N1 virus was characterized in NI test using specific post infection antisera from ferret. (i) The H5N1 NA cross reacted with the four ferrets anti sera which were directed against the swine influenza strains. (ii) No antigenic relationship could be detected between the NA of A/HK virus and that of the human A/H1N1 influenza strains (A/PR8 and A/Texas)

2) Distribution of anti-A/HK/156/97 H5N1 HI Abs in human sera. All human sera were tested in HI test using A/HK/ H5N1 virus. Among the 200 sera tested, 198 were found to be devoid of HI Abs and two of them exhibited a very low HI titre (10). These two weakly positive sera were eliminated from the serum panel we used thereafter.

3) Distribution of anti-A/HK/156/97 H5N1 NI Abs in human sera. All the 198 remaining human sera were tested in NI against the A/H5N1 strain, and two different A/H1N1 strains, the prevalence of NI positive sera was expressed as a function of both antigen and age : (i) the first profile obtained with A/HK and A/PR8 showed a very high prevalence (³ 80 %) of NI positive sera in the two age groups corresponding to persons > 60 years old, when a very low prevalence was found in the third group (< 59 years old). (ii) The second profile obtained with A/Texas differed from the one described above in terms of NI positive Ab prevalence in the two age groups corresponding to persons <76 years ; indeed the prevalence increased significantly in the youngest age group (20-59 years) while it decreased significantly in the medium age group (60-75 years old).

Such an analysis of NI Ab prevalence in human sera as a function of age of patients and A/H1N1 virus used in NI test, suggested a close antigenic relationship between NA of both A/HK and A/PR8 viruses when A/Texas NA looked significantly different. As a consequence we have no longer introduced the A/Texas strain in the following experiments.

Demonstrating in human sera the existence of NI Abs in absence of any HI Abs gave us an excellent opportunity to search for the neutralizing potential of anti-A/HK anti-NA Abs.

The prevalence of anti-A/HK Nt Abs was calculated by age groups and then was compared with that of either anti- A/HK/ NI Abs or anti A/PR8 NI Abs (figure 1). Two observations emerged from such a comparison : (i) Inside each age group, a similar ratio between anti A/HK Nt Ab frequency and either anti A/HK NI Ab or anti A/PR8 NI Ab frequencies. (ii) As a function of the age group, the frequency ratio between Nt Abs and NI Abs (anti A/HK and anti PR8 as well) was shown highly different. While a similar NI Ab prevalence (80%) (anti- A/HK and anti-A/PR8 as well) was found in both 60-75 and 76-98 years age groups, anti A/HK Nt Ab prevalence were found highly different, 26 % versus 70 %. As a consequence Nt to NI Ab ratios were found 0.33 and 0.80 in 60-75 and 76-98 age groups respectively. Due to the very low frequency of anti A/HK N1 Abs in the 20-59 year age group, results regarding Nt/NI frequency ratio could not be considered as interpretable.

Southeast Poultry Research Lab, 934 College Station Road, Athens, USA

Analysis of the structure of the avian influenza (AI) virus hemagglutinin (HA) gene and protein has yielded a wealth of information on the virulence mechanisms of influenza viruses. The AI hemagglutinin appears to be unique in its capacity to accept basic amino acids at its’ proteolytic cleavage site (PCS). The association of multiple basic (MB) amino acids, tissue spread and virulence by AI strains first proposed by Klenk, Rott, Orlich and others in the late 1970’s has held fast for over two decades now. While other structural characteristics and other genes can certainly influence virulence, the presence of the MB amino acids at the PCS has provided a hallmark structural feature which justifies continuing sequence analysis of emerging field isolates of AI strains. In addition to this structural feature, the distal tip of the HA is prone to appearance and disappearance of glycosylation sites, some of which have been associated with virulence. In our laboratory, for example it was shown that presence of a carbohydrate near the receptor binding site (RBS) of an H7 AI isolate was associated with increased virulence in chickens.

The H5N1 viruses which emerged in Hong Kong in 1997 were unique both in structure of their HA protein and pathology in animals. These chicken viruses contained multiple basic amino acids at the cleavage site and a mutable carbohydrate addition site at the tip of the receptor binding site. They were highly lethal in chickens but unlike other highly pathogenic AI strains they were also lethal in mice. Thus these viruses have provided an ideal model system with which to study the structural features of the HA of a highly virulent virus.

The importance of the structural features of the HA in the recent outbreaks in Mexico, Australia, Hong Kong and in the ongoing outbreak of H7 avian influenza in the northeast United States, as delineated by our laboratory will be summarized. The structure of the avian influenza hemagglutinin is a continually changing landscape and following and characterizing these changes remains an important challenge.

J. Banks, E. C. Speidel, D. J. Alexander.

VLA Weybridge, Addlestone, UK.

Two H5N1 viruses were isolated from a single turkey during an outbreak of highly pathogenic avian influenza (HPAI) in Norfolk, England in 1991. One (A/ty/Eng/50-92/91) is highly pathogenic for domestic poultry, has multiple basic amino acids at the haemagglutinin (HA) cleavage site and grows well in MDCK cells without the addition of trypsin, characteristics typical of HPAI viruses. The other (A/ty/Eng/87-92/91) has the same HA cleavage site, but is not pathogenic and does not grow in MDCK cells even with the addition of trypsin. The two isolates do not have any sequence differences in the haemagglutinin (HA) gene which correlate with virulence (1). This study investigates the growth and molecular characteristics of a limiting dilution passage clone of A/ty/Eng/87-92/91 (LDP3).

Two groups of ten 3-4-week old specific pathogen free leghorn chickens were inoculated with 10⁶ EID50 of freshly harvested LDP3 virus, one group intranasally and the other group by the intravenous route. One-day post infection (PI) six uninoculated birds were placed in contact. Cloacal and tracheal swabs were taken daily for 10 days, and one bird from each inoculated group was killed from days 3-11. Eight organ samples were taken from each bird at sacrifice (kidney, pancreas, lung, heart, brain, spleen, intestine, and trachea). Virus isolation was attempted from the samples by three passages in 9 or 10-day-old embryonating fowls’ eggs. Brain tissues were tested by nested RT/PCR for evidence of H5 influenza virus. The experiment was repeated using 3-4-week old turkeys.

Fourteen-day-old embryonating fowls’ eggs were inoculated with the avian influenza isolate LDP3. Allantoic fluids were harvested after 48 hours incubation at 37^oC and screened for growth on confluent MDCK monolayers without the addition of trypsin. Parental virus LDP3 was included as a control. Each culture was sampled daily and tested for HA activity. Selected cultures positive by HA were passaged in MDCK cells and 11-day-old embryonating fowls’ eggs. Two isolates (G33 and B22) were selected for sequencing of the entire HA gene. The entire PB2 gene for isolates G33, LDP3 and clone 2L was sequenced and partially sequenced for A/ty/Eng/50-92/91 clones 3L, 5L, and 87v from A/ty/Eng/87-92/91. The neuraminidase (NA) gene was sequenced for LDP3 and 2L. An IVPI was determined for clone G33 in 6-week-old white leghorn chickens

Chickens and turkeys infected with clone LDP3 by the intranasal or by the intravenous routes excreted very little virus. No clinical signs were seen and there was no transmission to ‘in contact’ birds. Virus was only recovered from the kidneys of sacrificed chickens and nested RT/PCR analysis of the brains for evidence of virus was negative. This is in contrast to A/ty/Eng/50-92/91, which can be easily isolated or detected by RT/PCR from tracheal or cloacal swabs, or from organs, including the lungs, kidney and brain. All sentinel birds become infected with this virus.

Of the 128 14-day-old embryonating fowls’ eggs inoculated with LDP3 55 (43%) gave positive HA results in the MDCK screening assay. All MDCK isolates inoculated into 11-day-old embryonating fowls’ eggs were lethal to the embryos and the allantoic fluids were positive for HA activity. The second passage of 32 MDCK isolates yielded 19 HA positive cultures, the parental virus LDP3 did not initiate a productive infection on primary or second passage. An IVPI index for isolate G30 was 0 and no clinical signs were observed. Nucleotide sequencing of the HA gene for isolates G33 and B22 showed no differences from the parental LDP3 sequence. Nucleotide sequencing of the entire PB2 gene for clones LDP3, G33 and 2L and partial sequencing of clones 5L from A/ty/Eng/50-92/91and 87v from A/ty/Eng/87-92/91 did not show any amino acid changes that correlate with pathogenicity

There are no nucleotide differences between the NA genes of LDP3 and 2L but there is a deletion in the stalk region of 23 amino acids as compared with A/parrot/NI/73.

The H5N1 isolate A/ty/Eng/87-92/91 was isolated from the brain of a turkey that had died of HPAI. However, clone LDP3 derived from this isolate is unable to disseminate throughout the body or to replicate in the brains of chickens and turkeys, as would be expected for a virus with multiple basic amino acids at the HA cleavage site. Passage in 14-day-old embryonating fowls’ eggs readily selects mutants from the avirulent clone LDP3 of A/ty/Eng/87-92 that can grow in MDCK cells in the absence of trypsin. This is in contrast to passage of LDP3 in chickens or 10 to 11-day-old embryonating fowls’ eggs, which fails to produce mutants capable of growth in MDCK. This increased measure of virulence however, did not translate to virulence in chickens, clone G30 had an IVPI of 0. No changes were seen in the HA nucleotide sequences relative to the parental LDP3 virus and therefore the selective pressure giving rise to growth in MDCK cells of mutants B22 and G33 must lie elsewhere in the genome. Reassortant experiments have led to speculation that the PB2 and Matrix genes may play a role in the modulation of pathogenicity (2). Therefore the entire PB2 gene for clones 2L G33 and LDP3 was nucleotide sequenced highlighting three deduced amino acid changes that appeared to correlate with pathogenicity. However, partial nucleotide sequencing of clones 5L, and 87v showed that there was no correlation with the deduced amino acid changes and IVPI.

Recently it has been suggested that additional glycosylation of the HA together with a shortened NA stalk are characteristic features of the H5 and H7 chicken viruses (3). Interestingly the NA genes of LDP3 and A/ty/Eng/50-92/91 have a predicted 23 amino acid deletion as compared with A/parrot/NI/73. However neither of these isolates have additional glycosylation sites on the HA thus proving to be exceptions for this rule.

1. Wood, G. W., Banks, J., McCauley, J. W. & Alexander, D. J. (1994). Deduced amino acid sequences of the haemagglutinin of H5N1 avian influenza virus isolates from an outbreak in turkeys in Norfolk, England. Archives of Virology 134, 185-194.
2. Banbura, M. W., Kawaoka, Y., Thomas, T. L. & Webster, R. G. (1991). Reassortants with equine 1 (H7N7) influenza virus haemagglutinin in an avian influenza virus genetic background are pathogenic in chickens. Virology 184, 469-71.
3. Matrosovich, M., Zhou, N., Kawaoka, Y. & Webster, R. (1999). The surface glycoproteins of H5 influenza viruses isolated from humans, chickens, and wild aquatic birds have distinguishable properties. Journal of Virology 73, 1146-1155.

Ralf Wagner¹, Thorsten Wolff¹, Astrid Herwig¹, Stephan Pleschka², and Hans-Dieter Klenk¹

1. Institut für Virologie, Robert-Koch-Str. 17, 35037 Marburg, Germany
2. Institut für Mikrobiologie, Frankfurter Str. 107, 35392 Giessen, Germany

Influenza virus infection is initiated by binding of virus particles to sialic acid-containing receptors on the surface of host cells. This binding function is provided by the influenza virus hemagglutinin (HA) glycoprotein. The hemagglutinin region responsible for receptor binding has been shown to involve a pocket of amino acids at the membrane distal tip of the molecule. With HA of fowl plague virus (A/FPV/Rostock/34) this binding pocket is flanked by two N-linked oligosaccharide side chains. In previous studies in which FPV-HA mutants lacking either one (mutants G1 and G2) or both (mutant G1,2) glycosylation sites were expressed from an SV40 vector we showed that these glycans modulate receptor binding affinity (1). In this study we were interested in whether these glycan depletions have an effect on the virus growth when mutated HA is stably inserted in recombinant viruses via a reverse genetics approach. In this system an influenza virus-like mutated HA segment is transcribed in the nucleus of transfected cells by cellular RNA polymerase I from a plasmid vector. Upon infection of the cells with a suitable helper virus mutated HA segments are incorporated into progeny virions. Two reassortants of A/WSN/33 were used as helper viruses to create two series of HA-mutant recombinant viruses only differing in carrying neuraminidase (NA) of either N1- or N2-subtype, respectively.

Here we show that the lack of one (mutant G2) or both glycans (mutant G1,2) has a severe effect on growth of the N1-NA-containing viruses which is mainly due to an impaired release of progeny virus from infected cells. For N2-NA viruses such growth restrictions were not detectable with the G2 mutant and much less apparent with the G1,2 mutant. Our results indicate that interactions between the receptor binding site and adjacent oligo-saccharides of HA are crucial for the potential of influenza viruses to grow in host cells.

CV1 cells were transfected with the plasmid containing the FPV-HA gene under RNA Pol I control along with expression plasmids encoding the proteins (PB1, PB2, PA, NP) of the influenza virus polymerase complex. 36 hours post transfection cells were infected either with the reassortant HK-WSN (H3N1) or WSN-HK (H1N2) and progeny viruses were harvested 16 hours p.i.. Selection for recombinant viruses was performed using a neutralizing anti H3 serum (HK-WSN system) or by passaging of supernatants on MDBK cells in the absence of trypsin (WSN-HK system). Recombinamnt viruses were purified by three plaque passages on MDBK cells.

For growth curves MDBK or MDCK cells were infected at a multiplicity of infection (M.O.I.) of 0.0005 and HA titers in the supernatant monitored at various times p.i.. Plaque assays were done on MDCK and MDBK cells according to common methods.

Cells were infected at an M.O.I. > 1 for 10 hours. Vibrio cholerae neuraminidase (VCNA) was added to a concentration of 25 mU/ml for 2 hours. Infected cells not treated with VCNA were used as a control. Virus titers in supernatants were determined by plaque assay (infection of cells were done on ice to prevent an inhibitory effect of VCNA on virus binding).

RNA was isolated from viruses and subjected to RT-PCR with FPV-HA specific primers. RT-PCR fragments were checked by endonuclease digestion for the presense of a restriction site which had been introduced into the plasmid based HA-gene.

HA from metabolically labelled viruses was immunoprecipitated and analysed in an SDS-PAGE. These approaches confirmed the recombinant nature and the presence of the desired mutations in the HA of the rescued viruses.

Cells infected at low multiplicity with recombinant viruses were monitored at various times p.i. for the HA-titer in the medium. With viruses of the N1-series there was a marked decrease of virus growth. Release of the G1,2 mutant was retarded and virus yields in the medium were reduced about tenfold.

Likewise, there was a reduction in plaque size that was distinct with G1,2 and less pronounced with G2.

With viruses of the N2-series we detected a plaque size reduction for the G12 virus while G1 and G2 viruses were unaffected.

Growth restrictions of the glycosylation mutated viruses as described above could be attributed to an impaired release from infected cells by means of treatment with Vibrio cholerae neuraminidase (VCNA) (Fig. 1). With G2 and G12 viruses of the N1-series VCNA treatment was an absolut requirement for the quantitativ release of progeny viruses from the cells. Without VCNA treatment G2-viruses were released to only 19% and G1,2 viruses to only 6%. In the group of N2-NA containing viruses only the release of the G1,2 mutant was dependent on VCNA treatment.

Jens Reinhardt and Thorsten Wolff

Institut für Virologie, Philipps-Universität, Marburg, Germany

Influenza viruses utilize and manipulate host cell functions. We are interested to identify cellular factors that are targeted by the virus through protein-protein interactions. We have previously identified host proteins that interact with the viral NS1 protein (1) and have extended these studies to the M1 matrix protein of the influenza A virus. The M1 protein has been reported to have both structural and regulatory roles in virion assembly and disassembly, viral RNA transcription and intracellular transport of the genomic RNP segments. In addition, M1 proteins of influenza virus subtypes A, B and C were found to be phosphorylated on serine and threonine residues in infected cells. The significance of M1 protein phosphorylation and the corresponding kinases have not been determined yet. We hypothesized that one or more of the properties of the M1 protein are mediated by interactions with unknown cellular factors. In order to identify such cellular components, we have used the yeast two-hybrid system with the M1 protein as a bait. As a result, we isolated four independent cDNA clones that interacted specifically with the M1 protein in the yeast two hybrid system and corresponded to the highly conserved RACK1 (receptor of activated C kinase) protein. RACK1 has been previously suggested to serve as an intracellular receptor protein for activated protein kinase C (PKC) (2). We discuss a possible role of the M1-RACK1 interaction for phosphorylation of the M1 protein.

M1 cDNA of the influenza A/PR/8/34 virus was subcloned into pEG202 as a translational fusion with the DNA binding domain of the bacterial LexA protein. We used the yeast two hybrid system (3) to screen 1.2 x 10⁶ independent HeLa cell cDNA clones for interaction with the LexA-M1 fusion protein as measured by their ability to activate a LEU2 and a lacZ reporter gene simultanously. A total of four cDNAs were isolated that corresponded to the human RACK1 protein and interacted specifically with LexA-M1 but not with an unrelated LexA fusion construct.

RACK1 cDNA was expressed in E. coli as a glutathione-S-transferase (GST) fusion protein. GST-RACK1 was purified and immobilized on glutathione agarose beads and used to select proteins from extracts of metabolically labelled MDCK cells infected with the mouse-adapted influenza A/WSN/33 virus. Precipitated proteins were analyzed by SDS gel electrophoresis.

GST and GST-M1 fusion proteins were expressed in E. coli and purified from bacterial lysates by adsorption to glutathione agarose. Immobilized GST proteins were incubated with purified protein kinase C (Calbiochem) and 10 mCi of g-³²P-ATP for 15 minutes at 30° C. The GST proteins were subsequently washed and analyzed by SDS gel electrophoresis.

We used the M1 protein as a bait in the yeast two hybrid system to screen a human cDNA expression library for interacting factors. 4 cDNAs corresponding to the human RACK1 (receptor of activated C kinase) protein that interacted specifically with the M1 protein were isolated. The genetic interaction was confirmed by the specific coprecipitation of the M1 protein from extracts of virus-infected cells by using a GST-RACK1 fusion protein in a a GST-pull down assay (Fig.1).

Anke Feldmann, Ralf Wagner, Astrid Herwig, Thorsten Wolff and Hans-Dieter Klenk

Insitut für Virologie, Philipps-Universität, Marburg, Germany

Cleavage of hemagglutinin into its subunits HA1 and HA2 is a prerequisite for influenza virus infectivity. Cleavage is accomplished by a variety of cellular proteases depending on the amino acid composition of the connecting peptide between HA1 and HA2. Hemagglutinins of highly pathogenic avian influenza viruses possess multibasic cleavage sites that are recognized by ubiquitously expressed proteases like furin or PC5/PC6. These viruses replicate in cell culture without addition of exogenous proteases to the medium. In chickens these strains cause systemic infections. In contrast, apathogenic avian influenza viruses and most human strains possess a single arginine residue at the cleavage site. HA of these strains is cleaved by secreted proteases of epithelial cells, e.g. tryptase Clara or Factor X. In infected animals the infection of these viruses is usually limited to the upper respiratory or intestinal tract. The addition of trypsin to the medium is therefore necessary for productive replication of these viruses in cell culture.

An exception is the influenza virus A/WSN/33 (H1N1) that possesses only a single arginine residue at the cleavage site but replicates in some cell lines without addition of exogenous proteases. In mice, infection is not only limited to the lungs but virus can be found in various tissues of infected animals. Recently it was shown that WSN-NA promotes cleavage of WSN-HA by sequestering of plasminogen.

We were interested to analyze whether WSN like pathogenic avian influenza viruses causes a systemic infection in chicken embryos and if the "gene constellation" of WSN has an influence on the spread of infection.

Influenza virus strains used: A/FPV/Rostock/34 (H7N1); A/chick/Germany/N/34 (H10N7); A/WSN/33 (H1N1); A/HK-WSN (H3N1); A/WSN-HK (H1N2); A/H7WSN-HK (H7N2); A/H7HK-WSN (H7N1).

11 day old chicken embryos were infected via the allantoic route with the virus strains listed above. 18h (FPV) or 48h (others) after infection the bodies of the embryos were frozen in isopentane cooled on dry ice. 20µm cryosections were prepared and viral RNA was detected by in situ hybridization technique using ³⁵S-UTP labeled riboprobes specific for mRNA of H7 (FPV), H1 (WSN), H10 (Virus N) NP(Aichi) and NA (WSN). Bound radioactivity was visualized by exposition of the labeled slices to X-ray screen. Cell tropism studies were performed by covering the slides by liquid nuclear emulsion followed by a two days exposure. Sections were then developed and counterstained by hematoxilin-eosin.

In earlier experiments we compared the organ and cell tropism of the highly pathogenic FPV and the apathogenic Virus N. When 11 day-old chicken embryos were infected with FPV all organs were heavily labeled by the H7-specific riboprobe reflecting the systemic infection caused by this virus. In contrast, in chicken embryos infected with Virus N only the cloaca which is connected with the allantoic cavity was stained by the H10-specific riboprobe (Fig.1).

S.J. Baigent¹, R.C. Bethell² and J.W. M^cCauley¹

1Institute for Animal Health, Compton, Newbury, UK; ²Glaxo Wellcome, Stevenage, UK

The sialic acid analogue 4-guanidino-2,4-dideoxy-2,3-dehydro-N-acetylneuraminic acid (zanamivir) is a high affinity, high specificity inhibitor of all naturally-occurring influenza A and B virus NAs (1). Zanamivir-resistant variants, having mutations in either the HA or the NA, have been selected following repeated passage in the presence of the drug in tissue culture (see 2). We previously demonstrated that naturally-occurring avian influenza viruses show variation in sensitivity of their replication to zanamivir in tissue culture (3). A/Duck/Ireland/113/83 H5N8 (Duck Ireland) is highly sensitive, A/FPV/Egypt/45 H7N1 (Egypt) and A/FPV/England/1/63 H7N3 (Langham) show intermediate sensitivity and SD17 (H7N1, a reassortant having HA and NA from A/FPV/Rostock/34) is relatively resistant. However, for each of the four viruses, the NA activity is inhibited by zanamivir, while the haemagglutinating activity is relatively insensitive to zanamivir. Reassortant viruses, produced between pairs of these four ‘parent’ avian viruses, were studied in order to identify the genes which contribute to the variation in sensitivity to zanamivir in vitro and to define the mechanism by which these genes influence sensitivity. This knowledge could be used to predict the properties of new human pandemic influenza viruses that may emerge on widespread use of zanamivir.

Reassortant viruses, produced by co-infection of chick embryo fibroblast cells, were genotyped using serological analysis, nucleic acid analysis, and electrophoretic analysis of virus polypeptides. For each parent virus, and for 54 reassortants, the effect of zanamivir on virus replication in MDCK cells was examined in multi-cycle plaque formation assays and in single-cycle virus yield assays. Virus yield was measured by either infectivity titre or by haemagglutination titre. Mean IC₅₀ values (the concentration of zanamivir required to decrease virus replication by 50%) were determined, and the effect of genotype on sensitivity to zanamivir analysed using the Wilcoxon Rank Sum Test (4). Ability of the viruses to elute from chick erythrocytes at 37⁰C was studied. The length of the stalk region of each NA was determined by sequencing of PCR products. The locations of potential glycosylation sites on the four HA trimers were determined from the amino acid sequences of these molecules.

Analysis of the reassortant viruses showed that both sensitivity of virus to zanamivir in tissue culture, and ability of virus to elute from erythrocytes, are determined by segment 4 (HA) and segment 6 (NA) and are independent of the remaining six RNA segments.

In each of the three tissue culture assays, SD17 HA was associated with lowest sensitivity to zanamivir in tissue culture and with ability to elute from erythrocytes (table 1). SD17 HA, unlike the other three HAs, possesses a glycan at position 158 (H3 numbering), close to the receptor binding site. For a given HA type, SD17 NA and Langham NA were associated with lowest sensitivity to zanamivir in tissue culture and with inability to elute from erythrocytes (table 1). These two NAs have short stalks (34 and 28 amino acids respectively). Egypt NA and Duck Ireland NA conferred greatest sensitivity in tissue culture and were associated with ability to elute from chick erythrocytes (table 1). These two NAs have long stalks (56 and 52 amino acids respectively).

Table 1

S=SD17; L=Langham; E=Egypt; D=Duck Ireland

+ virus able to elute; - virus unable to elute

The differential sensitivities of avian influenza viruses to zanamivir in tissue culture can be explained by the properties of their HA and NA. We suggest that sialic acid bound to position 158 of SD17 HA may fill the receptor binding pocket, competing with sialic acid attached to cellular receptors or to virus particles. Thus, viruses having this HA can elute from receptors with decreased dependence on NA activity and are less sensitive to zanamivir in tissue culture.

A long-stalked NA is more efficient in releasing virus from receptors than is a short-stalked (inefficient) NA, so viruses having long-stalked NAs can elute efficiently from receptors in the absence of zanamivir. However, because a long-stalked NA plays a greater role in release of virus from cells than does a short-stalked NA, viruses having a long-stalked NA are more sensitive to zanamivir in tissue culture.

1. von Itzstein, M. et al. (1993). Rational design of potent sialidase-based inhibitors of influenza virus replication. Nature 363: 418-423.
2. Mendel, D.B. and Roberts N.A. (1998). In-vitro and in-vivo efficacy of influenza neuraminidase inhibitors. Current Opinion in Infectious Diseases, 11: 727-732.
3. Thomas, G.P. et al. (1994). Inhibition of the growth of influenza viruses in vitro by 4-Guanidino-2,4-dideoxy-N-acetylneuraminic acid. Antiviral Research 24: 351-356.
4. Wilcoxon, F. (1945). Individual comparisons by ranking methods. Biom. Bulletin. 1: 80-83.

Laboratory of Virology, Faculty of Veterinary Medicine, University of Gent, Belgium

Uncomplicated influenza in man and animals is characterized by an abrupt onset of fever, chills, loss of apetite, myalgias, and respiratory signs. The target cell for the virus is the epithelial cell along the entire respiratory tract, but the pattern of infection and the relative involvement of the lower respiratory tract seem to vary in different animal species. Histological studies of the upper respiratory tract and/or lungs typically reveal necrosis of epithelial cells and infiltration with neutrophils. Despite extensive clinical and pathologic descriptions, the causes of disease symptoms and pathology remain largely unexplored. There is, however, growing evidence that so-called early cytokines produced at the site of infection mediate both symptom formation and inflammation. Among these cytokines are interferon-alpha (IFN-a), tumour necrosis factor-a (TNF-a), interleukin-1 and interleukin-6, and several chemokines. This review concentrates on several aspects of the cytokine response during acute influenza: 1) Cytokine profiles in the respiratory tract; 2) Cytokine induction at the cellular level; 3) The role of specific cytokines in influenza pathogenesis. Special emphasis is paid to studies in man and natural influenza virus hosts.

The kinetics of several cytokines have been studied during experimental influenza virus infections of humans and swine. In human volunteers, systemic disease and upper respiratory illness were most prominent (2). IFN-a, IL-6, TNF-a and IL-8 were found in nasal secretions during the period of maximal symptoms. IFN-a and IL-6 correlated directly with viral titers and illness severity. Other cytokines such as IL-1b, IL-2 or TGF-b did not show significant increases.

Using gnotobiotic pigs, an experimental model to reproduce "swine flu" has been developed in the author’s laboratory (7). The typical clinicopathological manifestations include fever, anorexia, depression, severe dyspnoea and a necrotizing bronchopneumonia. High levels of IFN-a, TNF-a and IL-1 were detected in bronchoalveolar lavage (BAL) fluids within 1 day after infection. When cytokine concentrations, lung virus titers and BAL neutrophil counts were compared, a direct correlation between the three was noted. In addition, the rise and fall of cytokine production exactly coincided with clinical symptoms.

The mechanisms of cytokine induction by the influenza virus are still unclear. There is no answer yet to several pertinent questions such as: Does cytokine production occur in infected or bystander cells? Which viral genes/proteins are involved? Do mechanisms of induction differ for different cytokines? Some in vitro IFN-a and TNF-a studies have revealed conflicting data. Cytokine production is strongly cell-type-dependent, and the use of different cell types may account for discrepancies in in vitro studies (1). So far, the in vivo cytokine producing cells during influenza have not been characterized. Therefore, we set out to determine the cellular origin of IFN-a and TNF-a in the lung of influenza virus-infected pigs. Preliminary results will be presented at the symposium.

Cytokines are part of a complex network, and the role of individual cytokines in vivo is difficult to determine. Here again, experimental studies using different approaches and animal species have given conflicting results. The effects of specific anti-cytokine strategies have been examined in murine models only (3, 4, 5). These studies have confirmed the significance of TNF-a, IL-1b, IL-1a and IFN-a in influenza pathogenesis. However, the effect of blocking individual cytokines was only partial and in some cases very modest. Studies in ferrets, on the other hand, failed to find evidence for a role of IL-1, IL-6 and TNF-a in influenza-induced fever (6).

1. Bussfeld, D., Bacher, M., Moritz, A., Gemsa, D. and Sprenger, H. (1997) Expression of transcription factor genes after influenza A virus infection. Immunobiology, 198(1-3): 291-298.

2. Hayden, F.G., Fritz, R., Lobo, M.C., Alvord, W., Strober, W. and Straus, S.E. (1998) Local and systemic cytokine responses during experimental human influenza A virus infection. Relation to symptom formation and host defense. J. Clin. Invest., 101(3): 643-649.

3. Kozak, W., Zheng, H., Conn, C.A., Soszynski, D., Van der Ploeg, L.H.T. and Kluger, M.J. (1995) Thermal and behavioral effects of lipopolysaccharide and influenza in interleukin-1b-deficient mice. Am. J. Physiol., 269: R969-977.

4. Kurokawa, M., Imakita, M., Kumeda and C.A., Shiraki, K. (1996) Cascade of fever production in mice infected with influenza virus. J. Med. Virol., 50: 152-158.

5. Peper, R.L. and Van Campen, H. (1995) Tumor necrosis factor as a mediator of inflammation in influenza A viral pneumonia. Microbial Pathogenesis, 19: 175-183.

6. Price, G.E., Fenton, R.J., Smith, H. and Sweet, C. (1997) Are known pyrogenic cytokines responsible for fever in influenza? J. Med. Virol., 52(3): 336-340.

7. Van Reeth, K., Nauwynck, H. and Pensaert, M. (1998) Bronchoalveolar interferon-alpha, tumor necrosis factor-alpha, interleukin-1, and inflammation during acute influenza in pigs: a possible model for humans? J. Infect. Dis., 177(4): 1076-1079.

S. Schultz-Cherry and D. Swayne

Southeast Poultry Research Laboratory, United States Department of Agriculture, Agriculture Research Service, 934 College Station Road, Athens, Georgia 30605, USA

Virulent strains of influenza A virus continue to cause severe losses to the poultry industry worldwide. Highly pathogenic strains of avian influenza (AI) virus causes systemic infection and rapid death in chickens. Prior to 1997 avian influenza viruses were not considered a direct threat to human health. The outbreak in Hong Kong and the death of 6 people changed the thinking. Many elegant comparative studies showed that the avian and human Hong Kong strains of virus are identical, suggesting the cross-species transmission is not directly related to changes in the viral genes.

Although a great deal is known about influenza viruses, very little is known about viral pathogenesis. Influenza viruses induce apoptosis, or programmed cell death in cells in vitro and in vivo. The role of apoptosis in pathogenesis is unclear.

Studies using the highly pathogenic strain A/Turkey/Ontario/7732/66 (Ty/Ont) suggested the lymphocyte depletion observed in infected chickens is a direct result of apoptosis. Additional work showed that transforming growth factor-b (TGF-b), a highly conserved growth regulatory protein, is directly activated by influenza virus and induces apoptosis. Increased TGF-b activity was found within 8 hrs after chickens were infected with Ty/Ont. Interestingly, chickens infected with an equivalent concentration of A/Mallard/Wisconsin, a low pathogenic strain, showed no increased TGF-b activity and very little apoptosis.

In these studies, chickens infected with A/Hong Kong/156/1997 were examined for histopathology, increased TGF-b activity, and levels of apoptosis.

A/Hong Kong/156/97 (H5N1) (10⁶ mean embryo lethal doses) was intranasally inoculated into 4-week-old chickens. Birds were euthanatized at 1, 2, 4, 8, 16, 24, 36 and 48 hours post-inoculation (PI) with sodium pentobarbital. Select tissues were fixed in neutral buffered formalin solution and routinely processed to paraffin blocks and histologic sections.

The lungs of infected and control chickens were aseptically collected and stored in saline at –70C. After homogenization, the tissue extracts were filtered through a 0.22 mm syringe and tested in the normal rat kidney colony forming soft agar assay for TGF-b activity.

Serial sections of the described tissues were specifically stained for apoptotic cell death using a TUNEL assay (Boehringer Manheim).

No lesions or viral antigen were identified at 1, 2, 4, 8 and 16 hours PI. At 24 hours, AI viral antigen was present in the nasal cavity within respiratory epithelium and inflammatory cells. Histologically, mild heterohistiocytic rhinitis was present, especially in the infraorbital sinus and ventrum of the middle nasal septum adjacent to the choanal slit. Degeneration and cell death was present in respiratory epithelium. The spleen and cecal tonsils had AI viral antigen in macrophages and capillary endothelial cells. At 36 and 48 hours, degeneration and cell death were widespread in multiple visceral organs and the brain. Viral antigen was present in many cell types, especially in vascular endothelial cells, various epithelial cells and macrophages.

Increased apoptotic cell death was observed in the lungs and nasal septum 8 hours pi. Within 24 hours, there was a systemic increase in apoptosis, even in areas that failed to stain with viral antigen. Active TGF-b levels in the lungs of HK 156 infected chickens increased slightly 8 hours pi, but surprisingly dropped to background by 24 hours and remained level at 48 hours. These results are similar to chickens infected with the low path strain A/Mallard/Wisconsin. In contrast, chickens infected with Ty/Ont showed dramatic increases in TGF-b activity throughout the 72 hour experiment. The reason for these differences is unknown.

In these studies we show that Hong Kong 156 virus initially replicates in the respiratory epithelium, followed by rapid dissemination throughout the body. Increased apoptosis is seen early after infection and increases with time. In contrast to another highly pathogenic strain of avian influenza virus, HK 156 fails to activate TGF-b. The reason for this is unclear.

P.P. Heinen, A.P. van Nieuwstadt, E.A de Boer-Luijtze and A.T.J. Bianchi

Department of Mammalian Virology, Institute for Animal Science and Health (ID-DLO), Lelystad, The Netherlands

Vaccination of pigs against influenza is currently performed by two times intra-muscular application of an inactivated " whole-virus" vaccine, containing adjuvant. Such a vaccine induces high IgG antibody titres in the blood and can protect against clinical signs of disease, but protection against an infection and antigenically different virus strains of the same subtype remains questionable. We believe that a vaccination procedure that stimulates more of a mucosal immune response and cellular immunity will provide better protection. Therefore, we started a study to compare immune responses after natural infection or vaccination. An influenza challenge model in specified pathogen free (SPF) pigs was developed, and tools to analyse the immune response.

In a first experiment, six ten-week-old pigs were inoculated into the nostrils with an aerosol of the field isolate A/Swine/Neth/St. Oedenrode/96 (H3N2). ELISAs were developed to measure influenza virus nucleoprotein specific IgM, IgG1 and IgA antibodies. These isotype-specific ELISAs were used to monitor antibody responses in serum and at the mucosa of the respiratory tract after a primary infection with influenza virus. Also, the total Ig isotype concentrations were quantified, using an ELISA and the specific activity (titre of antibody / concentration of immunoglobulin) for each isotype in BALF and NLF was compared with that in serum collected at the same time.

In a second experiment humoral and cellular immune responses after vaccination and infection were compared. A lymphocyte proliferation test (LPT) was developed and attempts were made to develop a cytotoxic lymphocyte test (CTL) to analyse the cellular immune response. FACS analysis was used to phenotype the T cell subsets that were involved in the influenza-induced cellular immune response. Protection after infection or vaccination was measured by monitoring signs of disease, virus excretion and infection of contact animals.

Experiment 1:

The aerosol infection with influenza virus caused acute respiratory disease characterised by fever, dyspnoea and anorexia. An exudative endobronchitis was observed throughout the lung and viral antigens were present in the epithelial cells of the bronchi and bronchioli on post-infection days 1 and 2. Virus was isolated from bronchioalveolar lavage fluid (BALF), nasal lavage fluid (NLF) and from pharyngeal swabs until 5 days after infection.

Kinetics of IgM, IgG and IgA responses were determined in the blood and at the respiratory mucosa, with the isotype-specific ELISAs. The influenza specific activity of IgA in NLF was 570 and in BALF it was 280 times higher than in blood. The influenza specific activity of IgG1 was 8 times higher in BALF than in blood.

Experiment 2:

Humoral and cellular immune responses after infection or vaccination with influenza virus were studied.

The results indicate an active secretion of locally produced IgA antibodies into the nose and lungs and transudation of locally produced IgG1 into the lung after infection with influenza virus. These mucosal antibodies can provide local protection against re-infection. Conventional vaccination with an inactivated whole-virus vaccine with an adjuvant provokes mainly a systemic IgG response, which will protect against pneumonia. The development of immunisation procedures that stimulate mucosal IgA production may improve efficacy of vaccination.

IgM antibodies specific for influenza virus were detectable for a short period after infection. Therefore, the IgM ELISA can be used for the diagnosis of a recent infection with influenza virus.

The challenge model in SPF pigs and our studies of humoral and cellular immune responses will improve our insight into immunity of pigs against influenza and will provide a rational basis for improvement of the vaccine.

D.L. Larsen¹*, N. Dybahal-Sissoko¹, A. Karasin¹, S. Carey¹, F. Zuckermann², and C.W. Olsen¹

1Department of Pathobiological Sciences, School of Veterinary Medicine, University of Wisconsin-Madison, Madison, WI 53706; ²Department of Veterinary Pathobiology, College of Veterinary Medicine, University of Illinois, Urbana, IL 61802

Influenza virus is a common respiratory pathogen of pigs (1). Pigs that have recovered from infection are protected from further challenge, but the currently available inactivated-virus vaccine doesn’t consistently induce complete protection (2). We have investigated DNA vaccination in pigs using the hemagglutinin (HA) gene of a prototypical swine virus. However, to assess the responses to vaccination, we need to understand the protective immune responses to natural infection, both systemically and at mucosal surfaces in the respiratory tract. To accomplish this we infected pigs with A/Swine/Indiana/1726/88 (Sw/IN) (H1N1) (3). At various times after infection we monitored antibody production and we enumerated antibody secreting cells (ASC) and IFN-g secreting cells by ELISPOT (4).

Fifteen 8-week old influenza seronegative pigs were infected intranasally on day 0 with 2x10⁵ EID₅₀ Sw/IN (H1N1). Three pigs were rechallenged with the same dose on day 42. Nasal swabs were taken for virus isolation on days 3-7 after primary infection and on days 3 and 6 after rechallenge. Three infected and one uninfected control pig were euthanized and sampled on days 7, 14, 21 and 42 post primary infection and 14 days after rechallenge. The following table outlines the samples taken and assays performed.

All 15 pigs shed virus for 6 or 7 days and showed mild clinical signs after primary infection, but the 3 pigs that were rechallenged on day 42 were completely protected and did not shed virus. Virus-specific serum IgG, IgA, and HI titers all peaked 2-3 weeks after primary infection and did not substantially increase after rechallenge. The predominant virus-specific isotype in serum was IgG. Pigs responded with virus-specific IgG and IgA in both the upper (nasal washes) and lower (BAL) airways. However, IgA was the predominant isotype in both sites. Results of the ASC ELISPOT assays demonstrated that the highest numbers of both IgG and IgA secreting cells are present in the nasal mucosa. In contrast, IFN-g secreting cells were predominately localized to the spleen and tracheobronchial lymph nodes.

These results give a detailed description of the kinetics and phenotype of the immune response to influenza virus infection in the pig. This information will be useful in the evaluation the immune responses to future influenza DNA vaccine studies in swine and help in determining which responses are important in protection from infection.

This research was supported by a USDA NRICGP grant.

1. B.C. Easterday, V.S. Hinshaw (1992) Swine Influenza. Diseases of Swine, 7^th edition, 349-357.
2. G. Brown, J.K. McMillen (1994) MaxiVac-FLU: evaluation of the safety and efficacy of a swine influenza vaccine. Proc. Am. Assoc. Swine Prac. 25^th Annual Meeting: 37-39.
3. M.G. Sheerar, B.C. Easterday, V.S. Hinshaw (1989) Antigenic conservation of H1N1 swine influenza viruses. J.Gen. Virol. 70: 3297-3303.
4. F.A. Zuckermann, R.J. Husmann, R. Schwartz, J. Brandt, E.M. deAntonio, S. Martin (1998) Veterinary Immunology and Immunopathology 36: 57-67.

Oliveira, J; Guitián, F.J.; Sanjuán, M.L.; Yus, E.

Infectious Diseases and Epidemiology, Veterinary College of Lugo, University of Santiago de Compostela, Spain

Swine influenza (SI) is a highly contagious, viral disease characterised by coughing, dyspnea, fever and prostration. In pig farms, morbidity is usually high (up to 100%) while mortality normally does not exceed 5 or 10% due mainly to secondary bacterial infections. Until now, in European countries, influenza viruses (IV) from pigs of the H1N1 and H3N2 subtypes have mainly been isolated. This infection contributes facilitating the setting-up of secondary bacterial infections, which cause important lesions in the respiratory system. Lung and pleura lesions, due to their chronic character, can be seen at the moment of the inspection at the slaughterhouse. Several studies have showed the relationships between lung and pleura lesions, recorded at slaughter, and poor productive performance, especially in fattening pigs(3). But lung and pleura lesions are influenced and associated to multiple factors both at individual as group level (another infections, environment, management, season, age of animals etc.). Some of these factors could be also associated to influenza infections and such factors could confound the true association between infection of influenza virus and respiratory system lesions recorded at slaughter(1). For example large pigs farms have a higher risk of virus infection than small farms, but often large farms have better management and environmental conditions than small farms, so if farm size is not taken into account a association high seroprevalence against some infections agents and low prevalence of lesions could be found and it would be concluded that infections be a protector factor of lesions.

Interaction virus influenza-secondary bacteria infection has been mainly reported at individual level but little research has been carried out at batch or farm level(2). The goal of this study is assessing and quantifying the association between geometric mean titre (MGT) against to two subtypes of IV and prevalence of pleurisy (PP) in batches of slaughterpigs adjusting for potential confounding factors.

During a month six period blood samples from slaughterpigs were collected and the lungs of such pigs were inspected for the presence of lung and pleura lesions. A total of 22 pig batches from fattening farms with a system all in-all out integrated in a regional Co-operative were inspected a blood sampled in a slaughterhouse in NW Spain. None of batches had been vaccinated against any subtype of IV. The number of pigs studied in each batch ranged from 10 to 50 (due to practical constraints non systematic randomization could be attempted and could not match serum samples and lesion on individual level).

Serum samples were tested using the IHA method (previously described by Palmer) against the subtypes H1N1 (strain A/NJ/8/76) and H3N2 (strain A/Victoria/1/75). Mean geometric titre (MGT) in each batch was calculated as:

Veterinary Laboratories Agency-Weybridge, New Haw, Addlestone, Surrey KT15 3NB, UK

Swine influenza (SI) was first observed at the time of the pandemic in humans in 1918 and it is known that the viruses responsible were closely related, possibly having derived from a common ancestor (1). Although the disease was described in pigs during the following years it was not until 1930 that the virus was isolated and identified. This H1N1 virus was the prototype strain of a group of viruses now known as classical viruses, which have been reported in pig populations worldwide.

Influenza A viruses of subtypes H1N1 and H3N2 have been widely reported in pigs, frequently associated with clinical disease. These include classical swine H1N1, avian-like H1N1 and human- and avian-like H3N2 viruses. Influenza is widespread and endemic in pig populations worldwide and is responsible for one of the most prevalent respiratory diseases in pigs.

SI generally appears with the introduction of new pigs into a herd, thereby being related to the movement of animals from infected to susceptible herds. Once a herd is infected the virus is likely to persist through the production of young susceptible pigs and the introduction of new stock. Infection is frequently subclinical, and typical symptoms are seen often in only 25 to 30 per cent of a herd. Swine husbandry practises influence directly the evolution of SI viruses through reduced immune pressure and constant availability of susceptible hosts, leading on the whole to reduced genetic drift in the genes encoding HA and NA, compared to those of similar viruses in the human population. However, mixing of pigs from multi-sources and the high frequency of contact with other species particularly humans provides an opportunity for cocirculation of viruses and genetic reassortment.

The classical H1N1 viruses have remained largely conserved both genetically and antigenically, although antigenic variants have been reported including strains in Canada since 1990 which have been associated with altered pathogenesis. In Europe, classical viruses have been largely replaced since 1979 by avian-like H1N1 viruses which are antigenically and genetically most closely related to H1N1 viruses isolated from ducks. Initially, following transmission to pigs, these viruses were genetically extremely unstable and it has been proposed that this was linked to a mutator mutation which may be necessary to enable influenza virus to cross the species barrier (2). Subsequently these viruses have become more stable establishing a new lineage in pigs. More recently an independent introduction of H1N1 virus from birds to pigs has occurred in Asia, and these viruses form a distinct sublineage of the Eurasian avian lineage.

Influenza A viruses of H3N2 subtype related closely to early human strains have been widely reported from pigs, particularly in Europe and Asia where they continue to circulate long after their disappearance from the human population. These viruses have evolved more slowly than their human counterparts and now form distinct lineages related to geographical location. In addition, some of the H3N2 viruses isolated from pigs in Asia are entirely avian-like, whilst other repeated introductions from humans appear to occur regularly. The maintenance of H3N2 viruses in pigs present a reservoir of virus which may in the future infect a susceptible human population.

The pig has been the leading contender for the role of intermediate host for reassortment of influenza A viruses. Pigs are the only mammalian species which are domesticated, reared in abundance and are susceptible to, and allow productive replication of avian and human influenza viruses. This susceptibility is due to the presence of both a 2,3- and a 2,6-galactose sialic acid linkages in cells lining the pig trachea which can result in modification of the receptor binding specificities of avian influenza viruses from a 2,3 to a 2,6 linkage thereby providing a potential link from birds to humans (3). The internal protein genes of human influenza viruses share a common ancestor with the genes of most swine influenza viruses. Also the pig has a broader host range concerning the compatibility of the nucleoprotein gene of viruses derived from other species. Experimental studies have shown that a wide diversity of avian influenza viruses (H1-H13) can be transmitted to pigs (4). Evidence for the pig as a mixing vessel of influenza viruses of non swine origin has been demonstrated in Europe with the detection of human-avian H3N2 reassortants (5) and human-avian H1N2 multiple reassortants (6) both of which have become established, regularly resulting in epidemics of disease in pigs. In addition, H1N2 viruses, derived from classical H1N1 and human-like H3N2 viruses have been isolated in France and Japan, where, in the latter case the virus spread and became associated with respiratory epizootics

Successful interspecies transmission of influenza viruses depends on the viral gene constellation with a progeny virus containing a specific constellation having the ability to replicate in the new host. In addition to these requirements it would appear that adaptation of a ‘newly’ transmitted influenza virus to pigs can take many years. Both human H3N2 and avian H1N1 were detected in pigs many years before they acquired the ability to spread rapidly and become associated with disease epidemics in pigs.

Pigs serve as a major reservoirs of H1N1 and H3N2 influenza viruses and are frequently involved in interspecies transmission of influenza viruses. The maintenance of these viruses in pigs and the frequent introduction of viruses from other species may be important in the generation of ‘new’ strains of influenza, some of which may have the potential to transmit to other species including humans.

1. Reid, A.H., Fanning, T.G., Hultin, J.V. & Taubenberger, J.K. (1998) Origin and evolution of the 1918 ‘Spanish’ influenza virus hemagglutinin gene. Proc.Natl.Acad.Sci., 96:1651-1656

2. Scholtissek, C. (1996) Molecular evolution of influenza viruses. Virus Genes, 11:209-215

3. Ito, T, Couceiro, J.N., Kelm, S., Baum, L.G., Krauss, S., Castrucci, M.R., Donatelli, I., Kida, H., Paulson, J.C., Webster, R.G. & Kawaoka, Y. (1998) Molecular basis for the generation in pigs of influenza A viruses with pandemic potential. J. Virol., 72:7367-7373

4. Kida, H., Ito, T., Yasuda, J., Shimizu, Y., Itakura, C., Shortridge, K. F., Kawaoka, Y., & Webster, R. G. (1994). Potential for transmission of avian influenza viruses to pigs. J. Gen.Virol., 75: 2183-2188

5. Castrucci, M. R., Donatelli, I., Sidoli, L., Barigazzi, G., Kawaoka, Y. & Webster, R. G. (1993). Genetic reassortment between avian and human influenza A viruses in Italian pigs. Virology, 193: 503-506

6. Brown, I.H., Harris, P.A., McCauley, J.W. & Alexander, D.J. (1998) Multiple genetic reassortment of avian and human influenza A viruses in European pigs, resulting in the emergence of an H1N2 virus of novel genotype. J.Gen.Virol, 79: 2947-2955

Nannan Zhou⁽¹⁾, Dennis A. Senne⁽²⁾, John G. Landgraf⁽²⁾, Sabrina L. Swenson⁽²⁾, Scott Krauss⁽¹⁾, Robert G. Webster^(1)*

1. Virology/Molecular Biology, St Jude C.R. Hospital, Memphis, TN, USA.
2. National Veterinary Services Laboratories, Ames, Iowa, USA.

* Presentor

Influenza in pigs was first observed in the United States during the catastrophic 1918 human influenza pandemic. The H1N1 swine influenza virus has remained in the swine population and has been responsible for one of the most prevalent respiratory disease in pigs in North America. Serological studies of pigs in the United States in the late 1980s demonstrated that the classic H1 influenza virus continued to circulate at high frequency (51%) among pigs. The isolation of H3N2 influenza A virus from pigs was first recorded in Taiwan during a human epidemic in 1969. The subsequent isolation of a number of H3N2 strains of influenza virus and indirect evidence from serological surveillance indicated that H3N2 influenza virus variants have been introduced into the swine population in other parts of the world.⁽¹⁾ Previous studies indicate that there have been multiple introductions of H3N2 influenza virus into pigs in Asia and Europe. However, in the USA, H3N2 virus infection in pigs is rare, compared to the prevalence of H1N1 viruses.⁽²⁾

The present study examines the characteristics of four H3N2 swine influenza viruses isolated in the USA in 1998.

Four swine viruses isolated from September to December of 1998 were used in this study. The four swine influenza viruses were derived from separate influenza outbreaks in four states including North Carolina (NC), Texas (TX), Minnesota (MN), and Iowa (IA). The viruses were isolated in chicken embryos from nasal swabs and tissues of sick and dead pigs. Antigenic, sequence analysis and phylogenetic analysis was done as previously described.⁽³⁾

In August 1998, a flu-like outbreak occurred on a pig farm in North Carolina, USA. The outbreak was severe with high fever, respiratory signs and abortion in the breeding females with 10% mortality in the breeding females. In November and December 1998, influenza outbreaks occurred in pigs in Texas, Minnesota and Iowa. These outbreaks were less severe with no mortality.

Antigenic characterization established that all of the isolates were H3N2 influenza viruses related to the strains circulating in humans in 1997-98.

Sequence analysis of each of the gene segments of the H3N2 isolates revealed that they fall into two groups; the North Carolina isolate being distinguishable from the viruses from the other three states. Phylogenetic analysis indicates that the North Carolina isolate is a reassortant possessing genes encoding human-like H3N2 glycoproteins with the remaining gene segments from classical swine influenza virus. The isolates from TX, MN, IA are also reassortants possessing genes encoding similar human-like H3N2 glycoproteins but the "internal genes" were derived from classical swine and avian influenza viruses.

The emergence of reassortant influenza viruses in pigs in the USA possessing genome segments originating from swine, human and avian influenza viruses illustrates continuing evolution of influenza viruses and the role of reassortment in nature. It remains to be determined whether these viruses will become established in the pig population of North America and whether they will represent an economic burden.

1. McFerran, J.B., Sandford, J.C., Connor, T.J., and Clarke J.K. (1972) Influenza-A virus in pigs. Lancet 2, 536.
2. Bikour M.H., Frost E.H., Deslandes S., Talbot B., Weber J.M., Elazhary Y. 1985. Recent H3N2 swine influenza virus with hemagglutinin and nucleoprotein genes similar to 1975 human strains. J. Gen. Virol. 76:697-703.
3. Guan, Y., Shortridge, K.F., Krauss, S., Li, P.H., Kawaoka, Y., Webster, R.G. 1996. Emergence of avian H1N1 influenza viruses in pigs in China. J. Virol. 70:8041-8046.

The H3N2 viruses were kindly provided by Kyoung-Jin Yoon, Iowa state veterinary Diagnostic Laboratory and Kurt Rossow, U of Minnesota Veterinary Diagnostic Laboratory and in North Carolina. The studies were supported in part by Public Health research grant AI29680 and by the American Lebanese Syrian Associated Christies.

Y.P. Lin, K.R. Cameron, M.S. Bennett, V. Gregory, A. Douglas, A.J. Hay, V. Regazzoli, *

P. Lenihan*

National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA.

* Department of Agriculture and Food, Central Veterinary Research Laboratory, Abbotstown, Castleknock, Dublin, 15, Ireland

The predominant influenza A H1N1 viruses currently circulating in pigs in Europe have evolved from an avian virus introduced into pigs in 1979 (1). Genetic comparisons of avian-like H1N1 viruses isolated from pigs in southern China in 1993 indicated that they resulted from an independent introduction of an avian virus into pigs in Asia (2). H1N1 viruses isolated from pigs in Ireland since 1991, although antigenically related to the early European avian-like H1N1 viruses, are genetically distinct from those isolated during the same period in continental Europe. Phylogenetic analyses of the eight genes have indicated that they form a separate lineage from those of other European isolates and the H1N1 viruses isolated in southern China in 1993 and suggest that these viruses may also be the result of an independent introduction of an avian H1N1 virus into pigs in Ireland.

Haemagglutination-inhibition (HI) tests using post-infection ferret antisera and monoclonal antibodies showed that the H1N1 viruses isolated in Ireland between 1992 and 1998 were antigenically related to, but distinguishable from, H1N1 reference viruses isolated in other European countries since 1979.

Comparison of the nucleotide sequences of HA genes showed that whereas the sequence encoding HA1s of recent 1997 Italian isolates differed by approximately 7 per cent from that of the early isolate, sw/Netherlands/3/80, the differences in sequence between the HA of the latter virus and those of recent Irish isolates were some two fold greater, approximately 14%, and similar to the divergence observed for the Hong Kong isolates such as sw/Hong Kong/168/93 (2). The divergence between the HAs of recent Irish and Italian viruses of approximately 19% was greater than between the HAs of Hong Kong and Irish viruses, isolated at about the same time (approximately 11%).

The sequences of the neuraminidase genes of the Irish viruses also differed to a greater extent (approximately 9%) than those of recent Italian viruses (approximately 6%) from the sequence a 1980 isolate, sw/OMS/2898/80, and have subsequently diverged further.

Comparative analyses of the other six genes indicated similar differences (approximately 2 fold) and divergence in the relationships between the early avian-like H1N1 viruses and the recent H1N1 isolates from Ireland or continental European countries. The two groups of viruses, however, share an unusual common property, amantadine resistance, due to substitution of serine 31 by asparagine in the M2 proteins of viruses isolated since about 1987.

On the one hand, the H1N1 viruses isolated from pigs in Ireland are antigenically related to H1N1 viruses circulating in other European countries and both groups of viruses are resistant to amantadine, an unusual characteristic for natural influenza A virus isolates. On the other hand, the Irish viruses are genetically as divergent from those circulating in continental Europe as the viruses isolated from pigs in southern China. The available evidence is thus consistent with the independent introduction of avian H1N1 viruses into pigs in these three different geographical locations.

1. Pensaert, M., Ottis, K., Vandeputte, J., Kaplan, M.M. and Bachmann, P.A. (1981). Evidence for the natural transmission of influenza A virus from wild ducks to swine and its potential importance for man. Bull. WHO, 59, 75-78.
2. Guan, Y., Shortridge, K.F., Krauss, S., Li, P.H., Kawaoka, Y. and Webster, R.G. (1996). Emergence of avian H1N1 influenza viruses in pigs in China. J. Virol. 70, 8041-8046.

Oliveira,J.; Guitián,J.;Yus, E.; Sanjuán,M.L.

Infectious Diseases and Epidemiology, Veterinary College of Lugo, University of Santiago de Compostela, Spain

Influenza virus (IV) type A has the capacity to infect man, birds, swine and some other animals. Birds are the most susceptible type of animals and which the virus is better adapted to. In birds the infection sometimes causes great disasters while in mammals the infection is usually less severe.

Since several years ago it has been proposed that IV would be able to shift from one species to another. Genetic and antigenic analysis carried out in virus isolated from several animals have showed important links within each other. The pig has been proposed as an animal that could play a key part in interspecies transmission especially from birds to man (1,2). This possibility would be greater if any variant of virus coming from birds could adapt and infect extensively to large pig populations. In this sense little research has been carried out to investigate the dissemination of the seroreactors in a large pig population against antigenic variants isolated from another species.

The goal of this study is to describe, at individual and at batch level, the extension of seropositivity against strains isolated from birds, humans and pigs in a large swine population.

During a six month period blood samples from slaughterpigs were collected in a slaughterhouse plant in NW Spain (Matadero Frigolouro, Pontevedra). The serum samples were kept at -20º C. until they were tested for the presence of antibodies against IV by IHA(3). Before testing by this serological technique and in order to eliminate the non-specific inhibitors against H1N1 and H3N2 subtypes all serum samples were heated at 56ºC. in a waterbath, adsorbed in chicken erythrocytes and treated with R.D.E.

A preliminary screening was made with the strains A/NJ/8/76 (H1N1) (NJ) and A/Victoria/1/75 (H3N2) (VIC). Every batch with one or more serum samples with a titre above 1:40 against such strains was considered as a seropositive batch. All samples of these seropositive batches were tested by IHA against A/sw/Netherlands/85 (H3N2) (NET), A/sw/Germany/91 (H1N1) (SW) an A/turkey/Germany/91 (H1N1) (TK ). IHA titres were converted to a logarithmic scale on base 2. A serum sample with a transformed titre >=3 (>=1:80) was considered positive.

The prevalence in each batch was computed as the number of seropositive samples divided by the total number of samples tested in a batch. Descriptive statistics for seroprevalence insight the batch against different strains were calculated only in batches with >=1 seropositive pigs.

A total of 47 batches and 897 individual samples were serotested against all strains. The total number of seropositive samples against all H1N1 strains was higher than against H3N2 strains in samples tested by IHA (Table 1). In figure 1 we show the distribution of transformed titres against all strains in the samples tested. Table 2 shows descriptive statistics of the seroprevalence into pig batch. Mean and quartiles of the distribution of seroprevalences insight the batch against H1N1 variants was higher than against the strains H3N2.

W.L.A. Loeffen^*, G. Nodelijk^†, P.P. Heinen^‡, L.A.M.G. van Leengoed^†, W.A. Hunneman^*, J.H.M. Verheijden^†

*) Animal Health Service, Boxtel, The Netherlands

†) Faculty of Veterinary Medicine, University of Utrecht, The Netherlands

‡) Institute for Animal Science and Health (ID-DLO), Lelystad, The Netherlands

During the winter of 1995/96 sixteen outbreaks of acute respiratory disease in finishing pigs were investigated¹. Influenza turned out to be the major cause of these outbreaks. Seven outbreaks were due to primary influenza infections, while two others were concurrent with an infection with Actinobacillus pleuropneumoniae. Infections with influenza virus, resulting in outbreaks with clinical symptoms, occurred mostly at the age of 15 to 20 weeks. In acute sera, collected during the early stages of disease, no antibodies were present. In other groups of the same age antibodies were already present, indicating that infections with influenza virus may often occur at an earlier stage in life, possibly with little or no clinical signs.

Several studies were set up to determine the moments of infection in piglets and finishing pigs. This paper describes a longitudinal study into the course of maternal antibodies up to the age of ten weeks and a cross-sectional study into seroprevalences in weaning piglets. The aim of these studies was to determine a decay function for maternal antibodies and to estimate the incidence of influenza infections during the weaning period.

Two herds were included in the longitudinal study in piglets. In one herd, also participating in the cross-sectional study, blood samples were collected weekly in eight litters (five piglets each) from 1 to 10 weeks of age. In another herd, not participating in the cross-sectional study, blood samples were collected in ten litters (four piglets each) at the age of 3 days and 1, 2, 4, 6, 8 and 10 weeks. In both herds blood samples from the mothers of the piglets were also collected within one week after farrowing.

Thirty-two breeding herds with at least 100 sows were randomly selected from a Dutch database, containing all pig herds. All herds were visited twice, in October 1995 and March 1996. During each visit one compartment with piglets of 4-5 weeks old and one compartment with piglets of 8-9 weeks old were sampled. In each compartment approximately 17 to 22 blood samples were taken (dependent on the number of piglets in the compartment)².

Since maternal antibodies turned out to be of major influence on the seroprevalence of eight to nine week old piglets, a mathematical method was used to correct for seroprevalence due to maternal antibodies. In short: based on the titers in the 4-5 week old group an estimate of the seroprevalence due to maternal antibodies in an 8-9 week old group in the same herd was calculated. This ‘expected’ seroprevalence was compared to the true seroprevalence in the 8-9 weeks old group. Since the true seroprevalence was based on a random sample, a 95% confidence interval could be calculated also.

All blood samples were tested at the same time for antibodies against influenza H1N1 and H3N2, using a haemagglutination inhibition (HI) test. Virus strains used in the test were A/swine/Best/96 (H1N1) and A/Swine/St Oedenrode/96 (H3N2), both isolated during the above mentioned survey in 1995/96¹. All titers were log-transformed.

Antibody titers of the sows and their piglets at the age of one week were positively correlated (Pearson correlation coefficient of 0.80-0.86 with a p<0.01 for influenza subtype H1N1 in both studies and H3N2 in the second study. For H3N2 in the first study, the correlation coefficient was only 0.54 with a p=0.17).

The linear model of the decay of maternal antibodies fitted very good for both subtypes H1N1 and H3N2. Explained variance for separate litters ranged from 0.93 to 0.99 for H1N1 and 0.78 to 0.99 for H3N2. Figure 1 shows an example of the decay for antibodies against H1N1 in the first longitudinal study. Both studies together resulted in a half-life of 12 days for both maternal H1N1 and H3N2 antibodies.

A.P. van Nieuwstadt, N. Stockhofe-Zurwieden, W. Loeffen, E. Kamp, J.C.de Jong

Institute for Animal Science and Health (ID-DLO), P.O.Box 65, NL-8200 AB LELYSTAD, The Netherlands

Animal Health Service, BOXTEL, The Netherlands

National Influenza Centre, Erasmus University, Rotterdam

Introduction.

Swine influenza is an acute respiratory disease caused by influenza virus type A, subtype H1N1 or H3N2. Outbreaks of respiratory disease due to influenza virus have occurred in the swine population of most West European countries since 1979. European H1N1 isolates are distinct from American strains, which are named classical swine influenza virus, and are more closely related to the virus that caused the severe influenza epidemic in humans in 1918, called the Spanish flu. Antgenically and genetically, European swine influenza virus isolates are closely related to the H1N1 isolates from ducks. Since 1984, infections with influenza virus subtype H3N2 also have been associated with outbreaks of respiratory disease in pigs. These H3N2 virus isolates from swine were antigenically very similar to the human Port Chalmers/1/73 strain. Serological surveys revealed that the H3N2 virus circulated in the swine population already at least ten years before clinical influenza in swine due to H3N2 virus infections was reported.

During the winter of 1995/96 we conducted a survey in 40 holdings of fattening pigs with a history of recurring respiratory disease for diagnosis of causative infectious agents. Sixteen outbreaks of respiratory disease were reported. Results showed that 7 outbreaks were caused by influenza virus and 5 were caused by Actinobacillus pleuropneumoniae. No causative infectious agent could be unambiguously diagnosed in the other outbreaks. The results indicated that H1N1- and H3N2 virus infections were a major cause of acute respiratory disease in fattening pigs, which had not been expected. The results did raise questions about (1) extensiveness of circulation of influenza virus among fattening pigs; (2) whether influenza virus infections caused disease on their own without complicating secondary bacterial infections; (3) whether an antigenic drift of H1N1- or H3N2 virus had occurred, which would require an updating of virus strains in the vaccine.

Forty herds with a history of recurring respiratory disease in fattening pigs were monitored during the winter of 1995/96. In these herds pigs were housed together at an age of 10 weeks in groups of 60 to 100 pigs. Stockholders were requested to notify early signs of respiratory disease such as increase of coughing, laboured breathing, and decrease of food intake. In this way 16 outbreaks of acute respiratory disease were reported to the veterinarian. From each outbreak 4 diseased pigs from a group with clear clinical signs were slaughtered for post mortem examination of the lungs, and 2 clinically healthy pigs of comparative age and origin from another group with no signs were slaughtered and served as controls. Lungs were examined for gross pathological lesions and tissue specimens were collected for histological examination and isolation of bacteria, viruses, and mycoplasmas. Acute and convalescent blood samples were collected from 10 groupmates in each outbreak. A 3rd blood sample was collected from the same pigs at the end of the fattening period. In addition, one or two other groups from the 16 holdings with a reported outbreak of respiratory disease were monitored serologically for influenza virus infections (27 groups in total). Blood from 10 pigs from each group was collected at the beginning and the end of the fattening period. Farmers who had reported a first outbreak of respiratory disease in their livestock were not requested to notify further outbreaks, so no clinical observations were available from those latter groups.

Seven out of 16 outbreaks of acute respiratory disease were diagnosed as influenza. These influenza outbreaks occurred 5 to 10 weeks after pigs were grouped together for fattening. In 5 outbreaks H1N1 virus was isolated from pneumonic lung lesions of all 4 diseased pigs that were necropsied and diagnosis was confirmed by a seroconversion of all 10 groupmates in a HAI test with H1N1 virus. In 2 other groups a H3N2 virus infection was diagnosed by virus isolation from the lungs and diagnosis was confirmed serologically. First signs of an outbreak were fever, prostration, apathy, and anorexia in 10 to 30% of pigs in the affected group. At necropsy a multifocal catarrhal pneumonia was observed, which was frequently associated with parenchymal haemorrhages and affected 10 to 50% of the lung tissue. Histological examination revealed necrotizing endobronchitis/bronchiolitis accompanied by broncho-interstitial inflammation and sero-fibrinous exudation and haemorrhages into the alveolar space. Similar lesions were observed in 7 of 14 control pigs, but lesions were much less severe. Influenza virus was isolated from lungs of 5 control pigs. Pathogenic bacteria were isolated from 15 out of 28 lungs of diseased pigs, but we never isolated the same species from all 4 pigs of the same outbreak.. Species isolated were Haemophilus parasuis, Bordetella bronchiseptica, Pasteurella multocida, Streptococcus suis, and (from one pig only) Actinobacillus pleuropneumoniae. In 13 pigs influenza was not complicated by a secondary bacterial infection, as no pathogenic bacteria could be isolated from the lungs.

Serological examination revealed extensive circulation of H1N1 virus during the fattening period in 20 of 43 groups (46.5%; the 16 groups with a reported outbreak were included) and of H3N2 virus in 9 groups (21%). Because no HAI antibodies were detected in blood samples collected at the beginning and the end of the fattening period, we concluded that no H1N1 virus circulated in 4 groups, and no H3N2 virus circulated in 11 groups. In the remaining groups, a few pigs or not any showed a significant rise of HAI antibody titre, but most pigs had antibodies in the first blood sample either indicating an earlier infection during the breeding period or a subclinical infection in the first few weeks of the fattening period.

Recent isolates of H1N1 virus [Infl A/swine/Neth/Best/96 (H1N1)] and H3N2 virus [Infl A/swine/Neth/St.Oedenrode/96 (H3N2)] were used as a HA antigen in the HAI test. A few convalescent sera were also tested with older isolates of H1N1- and H3N2 virus from swine, and HAI antibody titres were not much different. However, with the human Port Chalmers/1/73 (H3N2) strain antibody titres were significantly lower than with the St. Oedenrode/96 strain.

Influenza virus infections are a major cause of respiratory disease in fattening pigs. Influenza virus infections are often complicated by secondary bacterial infections, but an influenza virus infection can be the primary and single cause of a pneumonia.

Clinical signs of influenza are not pathognomic, and an infection with Actinobacillus pleuropneumoniae has to be considered in clinical differential diagnosis. At gross pathological examination, the two infections can be distinguished, however. An infection with Actinobacillus pleuropneumoniae causes a necrotizing haemorrhagic fibrinous pleuropneumonia.

There are indications of an antigenic drift of porcine H3N2 virus, by which antigenic relatedness with the human Port Chalmers/1/73 strain was lost.

This study was financially supported by Intervet International, Boxmeer, The Netherlands

Iwona Markowska-Daniel, Zygmunt Pejsak

National Veterinary Research Institute, Swine Diseases Department, 57 Partyzantow Str., 24-100 Pulawy, Poland

Swine influenza is an acute infectious, respiratory diseases of swine caused by type A influenza viruses belonging to the Family Orthomyxoviridae, genus Influenzavirus. The disease is characterized by sudden onset, coughing, dyspnea, fever and rapid recovery.

The aim of the study was to determine the prevalence of porcine and human Influenza A virus (subtypes H1N1 and H3N2) among domestic pigs and wild boar populations in Poland.

3945 blood samples from animals raised in large farms located in 48 provinces (out of 49) of Poland were tested. Also blood samples from heart or pleural cavity, were collected from 440 hunted wild boars of unknown age, during the period January-November 1998 in 42 provinces of Poland. Samples of sera were stored at –20^oC prior to testing.

Two virus strains of different subtypes were used in this study: H1N1 (strain A/Sw/Bel/1/83, at a titer EID₅₀ 10^7.5/ml) and H3N2 (strain A/Sw/Gent/1/84, at a titer EID₅₀ 10^6.5/ml). The viruses were grown in the allantoic and amniotic cavities of 10-day-old embryonated SPF chicken eggs.

As a control of haemagglutination inhibition assay the reference sera for H1N1 subtype, at a titer 512 and for H3N2 subtype, at a titer 2560 were included.

RDE (receptor destroying enzyme) (Sigma) of Vibrio comma was used for elimination of non-specific inhibitors of haemagglutination.

Serum antibodies to influenza virus were detected by the haemagglutination inhibition assay (HI) according to standard procedure. The sera were heat - treated at 56^oC for 30 min. During the next step the sera were adsorbed with 50% (v/v) chicken erytrocytes (RBC) to remove non-specific haemagglutinins for 1 hr at 4^oC. Sera tested for seroconversion for H3N2 subtype were additionally treated with 100 U/ml RDE, overnight at 37^oC, followed by the inactivation at 56^oC for 30 min. In each test the necessary serum, virus and red blood cells controls were included. HI tests were done using 4 haemagglutinating units (HAU) of each virus and 0.5% (v/v) RBC. Owing to the different treatment of the sera, the initial serum dilutions were 1:4 against H1N1 influenza virus and 1:20 against H3N2 influenza virus. In microassay HI titer was taken as the reciprocal of the highest mean antibody dilution inhibiting 4 HAU of virus. To facilitate reading exact end-points the plates were slanted so that non-agglutinated cells flew freely as a tear drop. As a positive results the haemagglutynin titer „ 1:8 and „ 1:40, respectively for H1N1 and H3N2 strains was estimated.

The monitoring study showed the occurrence of influenza virus among pigs and wild boars. Antibodies specific to subtype H1N1 were detected in 35% of sera from pigs and 24.1% of sera from wild boars. A serological survey for the subtype H3N2 demonstrated 29.3% and 6.8% of seropositives, respectively among pigs and wild boars. It should be stressed that the subtype H1N1 of SIV circulate at high frequency among swine as well as wild boars in the West of Poland while H3N2 subtype is circulating particularly in the East of Poland.

The titer against H1N1 strain of influenza virus in the most sera was below 1:8; 395 out of 3945 tested sera (10%) have the titer >1:16, and 258 sera (6.5%) the titer above >1:32.

The seroconversion for H3N2 virus strain and the estimated titers were similar. 675 sera (17.1%) was characterized by the titer >1:40, a titer > 1:160 was estimated only in 158 cases (4%). 484 of pigs sera (12.3%) and 17 wild boar’s sera (3.9%) reacted with both tested antigens.

Presence of antibodies against influenza virus in domestic pigs and wild boars

Infections of domestic swine with H1N1 virus were described in France, England, Germany, Thailand and Japan. In our study we reported that epizootiological situation of pig herds in Poland in field of swine influenza is similar to the situation in neighboring countries. It was shown by the German authors that for different strains of influenza virus with antigenic pattern of H1N1 the percentage of seroreagents among tested swine ranged from 24% to 32 %, with the exception for the A/Singapore/6/86 strain, which caused the seroconversion in 3% of the animals only. A slightly less intense spreading was noted in humans for the strains SIV of H3N2 pattern, with the seroconversion with various studied strains estimated for 10-23% by the authors mentioned above. Moreover, these authors found that of the animals examined approx. 10% exhibited the antibodies against both antigenic types of the influenza virus. Similar studies were performed by Brown et al. in Great Britain. Positive results of serological investigation in regard to strains H1N1 were obtained in approx. 26% of the sera evaluated. A marked spreading of the infection with subtype H3N2 was found, with the presence of seroreagents at the level of 39%. The study performed by Madec et al. involved reproductive swine herds in the northern region of France (Bretany), where all the major centers of pig husbandry are located. The above mentioned authors found that until 1981 year the swine population was infected with the influenza virus to the minimal degree. In the forthcoming year the number of infected herds increased up to approx. 50%.

Summarizing, it might be stated that epizootiological situation concerning SI in Poland and other European countries intensively producing pigs is similar.

1. Brown IH, Ludwig S, Olsen CW, et al. Antigenic and genetic analyses of H1N1 influenza A viruses from European pigs. J Gen Virol. 1997;78 :553-62.
2. Madec F. et al. A retrospective serological study about influenza infections in the pic in France. Proc. IPVS Congress, Lousanne, 1990.
3. Zhang X.M. et al. Detection of Antibodies Against Human Influenza A Virus (H1N1) in Swine Sera in the Federal Republic of Germany J .Vet. Med. B 35, 474-476, 1988.

K. Gillis, K. Van Reeth and M. Pensaert

Laboratory of Virology, Faculty of Veterinary Medicine, University of Gent, Belgium

Influenza viruses are subject to antigenic variation. This has two important implications. First, immune subjects may be incompletely protected against more recent drift variants. Second, antibodies against recent strains may react to a lower degree when tested against older strains in haemagglutination inhibition (HI) tests.

The evolution of swine influenza viruses has been followed less intensively than that of human and equine viruses. In this study we examined whether recent swine sera from the field show different reactivity against strains from '83-'84 and strains from '98 in HI tests.

A/New Jersey/8/76, Sw/Belgium/1/83 ¹, and Sw/Belgium/1/98. The H3N2 viruses included were: A/Port Chalmers/1/73, Sw/Ukkel/1/84 ², and Sw/Flanders/1/98. While A/New Jersey/8/76 and A/Port Chalmers/1/73 are included in most vaccines for swine, Sw/Belgium/1/83 and Sw/Ukkel/1/84 are used in routine haemagglutination inhibition test in the authors' laboratory. Sw/Belgium/1/98 and Sw/Flanders/1/98 were isolated in '98. All viruses were at the fourth passage in embryonated eggs.

All viruses were characterized in HI tests with monospecific swine sera. To obtain a monospecific antiserum for each specific isolate, gnotobiotic piglets were inoculated intratracheally with 10^7.5 EID₅₀ of the respective strain. Sera were collected 62 days post inoculation. Haemagglutination inhibition tests were performed as described earlier ³.

The '83-'84 and '98 isolates were included in HI-test against recent field sera collected from a total of 30 pigs on 3 different fattening farms (10 pigs per farm).

Results

Cross-HI tests with H1N1 strains demonstrated that the recent isolate Sw/Belgium/1/98 reacts with antisera against the strains from '76 and '83 (Table 1). On the other hand, the serum against the '98 strain revealed at least 32-fold lower antibody titers against the older isolates than against the homologous strain.

For the H3N2 strains, the recent Sw/Flanders/1/98 isolate reacts with antisera against both older isolates (Table 1). In the antiserum against the '98 strain, the antibody titers against the isolates from '73 and '84 are at least 8-fold lower than those against the homologous strain.

In the comparative HI tests with the H1N1 isolates 100%, 100%, and 90% of the animals of farms A, B, and C respectively, test positive with the recent Sw/Belgium/1/98 strain. These percentages are 80%, 50% and 30% when the tests are performed with Sw/Belgium/1/83 (Table 2). Sera which tested positive for both strains show 8 to 16 times higher titers against the recent strain.

For the H3N2 strains 30%, 60%, and 90% of the animals are found positive on the respective farms when tested with the recent Sw/Flanders/1/98 strain (Table 3). On these farms 0%, 40%, and 50% of the sera are positive against the Sw/Ukkel/1/84 strain. If a serum is positive for both H3N2 strains, antibody titers are up to 8 times higher for the '98 strain compared to the Sw/Ukkel/1/84 strain.

Table 3: H3N2 antibody titers in field sera from '98 in HI tests with isolates from '84 and from '98

These results clearly demonstrate that the number of animals detected depends on the virus strain used in the HI test. Although the underlying cause was not determined in this study, antigenic drift can be put forward.

In conclusion, these preliminary results indicate that the virus strain used in haemagglutination inhibition tests for serology should frequently be tested for its ability to detect seropositive animals and, if needed, be updated. In the future more swine influenza virus strains will be compared in HI tests to validate these findings.

1. HAESEBROUCK, F. and PENSAERT, M.B. (1986). Veterinary Microbiology 11, 239.
1. HAESEBROUCK, F., BIRONT, P., PENSAERT, M.B. and LEUNEN, J. (1985). American Journal of Veterinary Research 46, 1926.
2. PALMER, D.F., COLEMAN, M.T., DOWDLE, W.R. and SCHILD, G.C. (1975). Advanced Laboratory Techniques for Influenza Diagnostics. US Department of Health, Education and Welfare, Immunology Series 6.

Z. Pospísil, P. Lány, D. Zendulková, B. Tumová^§, P. Jahn, P. Cíhal

University of Veterinary and Pharmaceutical Sciences, Faculty of Veterinary Medicine, Palackého 1-3, 612 42 Brno, Czech Republic