Subclinical mastitis (SM) is a major concern in the dairy industry owing to economic losses from decreased milk production, treatment costs and the effects on dairy product quality (Urech et al. Reference Urech, Puhan and Schallibaum1999; Tyler et al. Reference Tyler, Cullor and Smith2002; LeBlanc et al. Reference LeBlanc, Lissemore, Kelton, Duffield and Leslie2006). SM in affected mammary glands easily develops into inflammation by increase in somatic cell count (SCC) and can be intractable. The mechanisms of onset and progression remain unclear, and diagnosis is often delayed because there is no visible sign of inflammation in milk and individual cows. Although many researchers have described methods for rapid detection of SM, the sensitivities and specificities of those methods vary and are uncertain (Ball et al. Reference Ball and Greer1991; Musser et al. Reference Musser, Anderson, Caballero, Amaya and Maroto-Puga1998; Middleton et al. Reference Middleton, Hardin, Steevens, Randle and Tyler2004). Consequently, a new marker for prediction of SM is needed for dairy cow management.
During inflammation of the mammary gland, the increase in SCC produces many chemotactic factors such as cytokines (Riollet et al. Reference Riollet, Rainard and Poutrel2000). The value of pre-inflammatory cytokines such as tumour necrosis factor-α and interleukin (IL)-1 and IL-6 for the detection of bovine clinical mastitis has been described previously (Droogmans et al. Reference Droogmans, Cludts, Cleuter, Kettmann and Burny1992; Nakajima et al. Reference Nakajima, Mikami, Yoshioka, Motoi, Ito, Ishikawa, Fuse, Nakano and Yasukawa1997; Okada et al. Reference Okada, Ohtsuka, Kon Nai, Kirisawa, Yokomizo, Yoshino and Rosol1999). The present study examined whether detection of IL-6 in milk could be useful for prediction of SM.
Materials and Methods
Quarter milk samples and cows
In this study, 77 quarter milk samples were collected from 20 healthy cows that were confirmed to have had no treatment of disease, including mastitis, 3 months before the experiment. The average age was 4.2 years. All cows were in the mid-to-late stages of lactation and were reared on the same dairy farm at Rakuno Gakuen University.
We divided the cows into two groups and compared the high individual SCC group (High group) with low individual SCC group (Low group). The average age in each group was 4.7 years in High group, and 3.8 years in Low group. Each group contained 10 cows; 40 quarter milk samples were collected from the High group, and 37 quarter milk samples were collected from the Low group (3 quarters were not collected from the Low group because of blind udder tracts). Cows were placed into each group according to the means of individual SCC levels; cows with an average of 7·7×105/ml [range (1·3–33·1)×105/ml] for 3 months before the experiment were placed in the High group, and cows with an average of 3·6×104/ml [range (1·1–7·4)×104/ml] for 3 months before the experiment were placed in the Low group. The general condition of the cows’ health were monitored by blood examinations on all cows.
Bacteriological examination of milk samples
Milk samples from all the cows were examined for the detection of bacteria as follows: approximately 10 μl of the quarter milk sample (samples were collected 3 times at different time points during the experiment) was streaked on Trypticase™ Soy Agar with 5% sheep blood (Becton Dickinson, Franklin Lakes NJ, USA) and was cultured aerobically at 35°C for 24 h. In the absence of bacterial growth after 24 h, plates from milk samples were cultured for an additional 24 h. The bacterial colonies were Gram stained and identified according to the flow chart by Tyler et al. (Reference Tyler, Cullor and Smith2002). Bacterial species were identified using commercial identification systems for Streptococcus spp. (API 20 Strep, bioMerieux Japan Inc., Tokyo, Japan) and Enterobacteriaceae (Enterotube II, Japan Becton Dickinson).
Examinations of SCC in individual milk and quarter milk samples
Individual SCC before and after the experiments were recorded from monthly results of the daily herd improvement test. SCC from the quarter milk samples were determined using a fluorescence optical system (FossomaticTM 90, FOSS Electric A/S, Hillerød, Denmark) according to the manufacturer's instructions.
Flow cytometry analysis of the milk cells
Milk cells from all cows were analysed by FCM. Cells were washed with phosphate-buffered saline (PBS) solution and incubated with specific monoclonal antibodies of CD4 (CACT138A, Veterinary Medical Research Diagnostics, Pullman WA, USA) and CD8 (CACT80C, Veterinary Medical Research Diagnostics) as described previously (Hagiwara et al. Reference Hagiwara, Domi and Ando2008). After washing again with PBS, the samples were incubated with FITC-conjugated anti-mouse IgG (Immuno Tech Co. Ltd., Osaka, Japan) to detect antibody-binding cells. After washing once again with PBS, the samples were fixed with 0·5% formaldehyde PBS and analysed using a flow cytometer (EPICS XL, Beckman Coulter Inc., Brea CA, USA).
Detection of IL-6 in whey samples
Detection of IL-6 in whey samples was by a sandwich enzyme-linked immunosorbent assay (ELISA) using bovine IL-6-specific antibody, as described previously (Hagiwara et al. Reference Hagiwara, Yamanaka, Hisaeda, Taharaguchi, Kirisawa and Iwai2001). Concentrations were determined from plotted standard curves obtained using recombinant IL-6 (rbIL-6).
Statistical analysis
We used the commercial statistical analysis software SPSS 15.0 J (SPSS Japan Inc., Tokyo, Japan) for all statistical analyses. An independent Student's t test and Mann-Whitney's U test were used to compare averages for the two groups. Sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio were calculated from the cross-tabulation tables.
Results
Data from the blood examinations showed no significant difference between High group cows (4.7 years old) and Low group cows (3.8 years old). Bacterial examinations in quarter milk samples showed a bacteria-positive rate of 40% in High group cows (16/40 samples) and 27% in Low group cows (10/37 samples). The judgment assumed even once positive among inspection three times if bacteria were detected. Bacterial species isolated from the examined milk were identified as coagulase-negative staphylococci [High group: 10% (4/40), Low group: 5·4% (2/37)], Enterococcus spp. [High group: 20% (8/40), Low group: 5·4% (2/37)] and Streptococcus spp. other than Str. agalactiae [High group: 10% (4/40), Low group: 16% (6/37)]. Bacteria detection did not differ significantly between the two SCC classification groups. The ratio of neutrophils to lymphocytes in the milk samples tended to be lower in bacteria-positive quarters (1·37) than in bacteria-negative quarters (1·57) but there was no significant difference between them. The CD8-positive cell rate of milk cells tended to be higher in High group cows (42·3%) than in Low group cows (30·6%). In contrast, the CD4-positive cell rate of milk cells tended to be lower in High group cows (66·7%) than in Low group cows (73·6%).
Levels of IL-6 in whey from mastitis cow are very high, e.g 6·5–46·5 ng/ml (Hagiwara et al. Reference Hagiwara, Yamanaka, Hisaeda, Taharaguchi, Kirisawa and Iwai2001). In the present experiment, the average of IL-6 concentration in quarter milk was significantly higher in the High group (207·0±441·6 pg/ml) than in the Low group (12·6±33·4 pg/ml) (Fig. 1). IL-6 concentrations in normal milk were estimated as low as the average of IL-6 in the Low group (12·6 pg/ml) plus twice the sd (66·8 pg/ml), resulting in 80 pg/ml. If the cross table was constructed using a threshold value of 80 pg/ml of IL-6 concentration and 1·0×105/ml of SCC in at least one quarter (Ruegg, Reference Ruegg2003) specificity and the positive likelihood ratio were higher using IL-6 concentration rather than using SCC for the individual prediction of SM (Table 1). IL-6 was detected in the milk from High-SCC quarters to a high rate. Among the study cows, 5/8 (62·5%) that showed positive for IL-6 in milk were subsequently diagnosed with clinical mastitis, which had to be treated with antibiotics, over 3 months after the measurements. On the other hand, the quarters for which IL-6 was not detected did not worsen to clinical mastitis and did not need treatment afterwards even if they were from the High-SCC group.
Fig. 1. Comparison of interleukin (IL)-6 concentrations in quarter milk samples from the low- and high-somatic cell count (SCC) groups of cows. Concentrations of IL-6 in quarter milk samples were quantified. Data for IL-6 levels were indicated as low-SCC cows (Low) and high-SCC cows (High).
Table 1 Comparison results of IL-6 concentration with somatic cell count (SCC) from the cross-tabulation table
† “Individual” columns are calculated as follows: “positive individual” has at least 1 quarter over each threshold, and “negative individual” has 0 quarters over each threshold
Discussion
IL-6 concentration in quarter milk was significantly higher in the High-SCC group than in the Low group (healthy control). Many researchers have confirmed that several cytokines, including IL-6, elicit the acute phase response and allow the accumulation of leucocytes in bovine clinical mastitis (Nakajima et al. Reference Nakajima, Mikami, Yoshioka, Motoi, Ito, Ishikawa, Fuse, Nakano and Yasukawa1997; Okada et al. Reference Okada, Ohtsuka, Kon Nai, Kirisawa, Yokomizo, Yoshino and Rosol1999; Riollet et al. Reference Riollet, Rainard and Poutrel2000; Hagiwara et al. Reference Hagiwara, Yamanaka, Hisaeda, Taharaguchi, Kirisawa and Iwai2001). The increase in IL-6 concentration in quarter milk is believed to indicate the duration of mammary gland inflammation due to bacterial infection. In the present study, the detection of IL-6 in milk samples predicted inflammation of exact quarter more accurately than SCC, because the specificity and positive likelihood ratio of IL-6 concentration was significantly greater than those of SCC (Table 1). Although SCC is the standard method used in the diagnosis of clinical mastitis (Tyler et al. Reference Tyler, Cullor and Smith2002; Pyorala Reference Pyorala2003), it was mentioned that SCC may not be enough for identifying SM (Middleton et al. Reference Middleton, Hardin, Steevens, Randle and Tyler2004). There is no definitive diagnostic probe to prognose SM leading to clinical mastitis. In this experiment, the sensitivity of IL-6 in quarter milk was better than that of SCC for future prognosis of mastitis. The measurement of IL-6 has the advantage in understanding the inflammatory status of the exact regions of udder.
CD4+ T lymphocyte is the predominant phenotype in healthy mammary glands, and CD8+ T cells migrate in bacteria-infected glands or in colostrum (Riollet et al. Reference Riollet, Rainard and Poutrel2000; Hagiwara et al. Reference Hagiwara, Domi and Ando2008). Therefore, the detection of and increase in CD8+ cells may indicate SM in mammary glands and inflammation caused by bacterial infection.
In this experiment, it was shown that there was a significant difference in IL-6 levels in milk between the two groups of cows. The results imply that the concentration of IL-6 in quarter milk may correspond to the early stages of inflammation in mammary glands, resulting in a future prediction marker for SM. Future work should make comparative study of IL-6 detection in milk using a large number of samples.
The authors thank the Hokkaido Dairy Milk Recording and Testing Association for technical support and Hitomi Aoki DVM, for general assistance. We are also grateful for the grant provided by The Japanese Society for Animal Cytokine Research.
