CLINICAL mastitis has the largest economic impact on the dairy cattle industry. Despite intensive bacteriological research, 20 to 35 per cent of clinical cases of bovine mastitis have an unknown aetiology (Miltenburg and others 1996, Barkema and others 1998). Although viral infections have occasionally been associated with bovine mastitis (Siegler and others 1984, Yoshikawa and others 1997), it is generally considered that viruses do not play a role in the aetiology of bovine mastitis (Watts

more »

... 8, Radostits and others 1994). This study was undertaken to gain more insight into the possible role of viruses in bovine clinical mastitis, due to the high percentage of unknown causes of clinical mastitis. In a case/control study, serum and milk samples were collected from 58 dairy cows with clinical mastitis in 10 different Dutch herds at the acute (day 0) and the convalescent phase (day 21) of the disease. Serum and milk samples were also taken from 58 healthy, matched control cows. The control cows were from the same herd, did not show mastitis symptoms, were of the same age as the mastitis cows and were in the same stage of lactation as the mastitis cows. Milk samples from the matched control cows were collected on the same day from the same quarters as the affected quarters of the corresponding clinical mastitis cases. The milk samples were collected as described by the National Mastitis Council (Harmon and others 1990). Samples for virus isolation were stored directly at -70°C, while milk samples for the screening of bacterial agents were stored at 2 to 4°C, and usually cultured within 24 hours. Blood samples were obtained from the median sacral vein of the tail, and centrifuged at 2000 g for 10 minutes. Sera were stored at -20°C. After sampling on day 0, cows with clinical mastitis were treated with antibiotics; the matched controls were not treated. Convalescent samples were taken at least two weeks after the last antibiotic medication. The number of lactating cows in the 10 different herds varied between 31 and 57, and the number of case/control pairs varied from one to 22 per herd (Table 1) . Milk samples were examined for the presence of viruses, using four different types of cell cultures: embryonic bovine trachea cells (EBTr; a semipermanent cell line developed in the authors' laboratory, ID-Lelystad); bovine epithelial udder cells (including fibroblasts) (Schmid and others 1983); bovine umbilical cord endothelial (BUE) cells (Van de Wiel and others 1989); and bovine alveolar lung macrophages obtained from specific pathogen free (SPF) cattle (Schrijver and others 1995). The milk samples were thawed, defatted by centrifugation at 1500 g for 10 minutes, and 0-5 ml of the defatted milk was used for virus isolation. For the virus isolation on macrophages, 100 1d of defatted milk samples were pipetted into the wells of a 96-well cell culture plate, containing 3 to 5 x 104 cells per well. The inoculated cells were incubated for five to seven days at 37°C, with 5 per cent carbon dioxide for the macrophages. After a freeze/thaw cycle, a second passage was performed. During the first and the second passages, the cell cultures were observed every day for a cytopathogenic effect (cpe). Four controls were included in each cell culture run. Two controls, containing 10'and 103 median tissue culture infective dose (TCID50) of bovine herpesvirus 1 (BHV-1) per millilitre of milk, served as positive controls. A milk sample without viruses and a plain cell culture control, served as negative controls. After the second passage, a haemadsorption reaction, for the detection of, for example, Orthomyxoviridae and Paramyxoviridae, was performed on EBTr cells with 0-2 per cent guinea pig erythrocytes, and incubated at 370C for one hour. An EBTr cell culture inoculated with parainfluenza virus 3 was used as a positive control. Electron microscopy (EM) was performed after the second passages for all four cell types. A 400 mesh carbon-coated nickel grid with a collodion film was floated on a drop of the inoculated cell cultures for five minutes, drained onto filter paper and stained with 2 per cent phosphotungstic acid (pH 6-8). The grids were examined by transmission EM after drying. Serum samples collected from mastitis cows were examined for antibodies against BHV-1 by ELISA (Kramps and others 1994), against bovine herpesvirus 2 (BHV-2) by a 24-hour virus neutralisation test (Bushnell and Edwards 1988), against bovine herpesvirus 4 (BHV-4) by ELISA (Wellenberg and others 1999), against bovine respiratory syncytial virus (BRSV) by ELISA (Westenbrink and others 1985), against bovine viral diarrhoea virus (BVDV) by ELISA (Westenbrink and others 1986), against bovine leukaemia virus (BLV) by ELISA (Pourquier), and against adenovirus type 3 by ELISA (BIO-X). In case blocking percentages, ELISA coefficients or optical density values indicated a significant increase in antibody titre, the serum samples were titrated by serial two-fold dilution steps. The control cows were only examined for antibodies against BHV-4, because only a few cows with clinical mastitis seroconverted (where seroconversion is defined as a seronegative acute serum and a seropositive convalescent serum) against viruses other than BHV-4. Serum samples containing antibodies against BHV-4, were titrated by serial two-fold dilution steps. A four-fold (two dilution steps) higher antibody titre in convalescent serum compared with acute serum is defined as a significant increase. Bacteriological culture of the milk samples was performed according to standards of the National Mastitis Council (Harmon and others 1990). Milk samples (0-01 ml) were inoculated on 6 per cent blood agar plates (both aerobically and anaerobically), on TCT medium (Thallium sulphate, Crystal Violet, Staphylococcus f-toxin; Merck) and on MacConkey number 3 agar (Oxoid). The plates were incubated at 370C and bacterial growth was evaluated after both 24 and 48 hours. Bacterial colonies were identified as described by the National Mastitis Council (Harmon and others 1990). Bacteria were considered to be pathogenic or nonpathogenic, on the basis of the description by Barkema and others (1998) which, in some cases, depended on the number of colonies isolated. No cpe was observed during the first and second passages in EBTr cell cultures, bovine udder epithelial cells, or macrophages that were inoculated with milk samples from the cows with clinical mastitis. No virus particles were detectable in these cell cultures by EM. The haemadsorption reaction, performed on EBTr cell cultures after the second passage, was negative for all these samples. The BUE cell cultures inoculated with milk samples from cow 49 (day 0), and cow 400 (day 21) of herd 8 and herd 10, respectively, showed cpe six to seven days after inoculation. Herpesvirus particles were detected by EM in the BUE cell cultures inoculated with milk from cow 12 of herd 9 (day 0 and day 21). No virus particles were detected in all the other milk samples from cows with clinical mastitis and the matched control cows by virus isolation on BUE cells, or by EM. The results of virus isolation on herd level and on an individual level are given in Tables 1 and 2, respectively. The three virus isolates were characterised as herpesviruses by EM. The virus isolates and control BHV-4 reference strains DN-599 and LVR 140, were partly neutralised with monospecific antiserum against BHV-4, and not with mono-