Flow cytometric analysis of neutrophils in cows' milk

Procedures were developed to count neutrophils in milk using a flow cytometer. Milk samples from 2 experiments were counted: 1 with 4 noninfected cows and a second with 5 noninfected cows that were injected with endotoxin in 2 mammary quarters. Thus, the procedures were evaluated on normal milk and on that with high somatic cell count. Flow cytometric procedures involved fluorescence detection (from the dye carboxydimethylfluorescein diacetate) to distinguish intact and viable from fragmented cells, forward light scatter to detect cell size differences, and right-angle side scatter to detect cellular granularity. High fluorescence, large size, and high degree of granularity identified viable neutrophils. For all samples, neutrophils were also counted manually, using the cytologic centrifugation approach to create the slides; manual counts were used as the standard for comparison. In experiment 1 (normal milk), mean values for percentage of viable neutrophils estimated by manual and flow cytometry procedures agreed closely (26% vs 25.8% for foremilk and 28.8% vs 26.6% for bucket milk). Sources of variation in manual and flow cytometric estimates of percentage of neutrophils were examined. Cow variation was significant (P < 0.01) for manual and flow cytometric counts, but was larger for flow cytometric counts. Day-to-day variation in counts on milk from the same cow was negligible for manual counts, but was significant (P < 0.01) for flow cytometric counts. Coefficients of variation were considerably Larger for manual counts than for flow cytometry. In experiment 2 (milk with high cell count, foremilk), agreement between mean values obtained by flow cytometry and by manual counting was somewhat less. However, predicting manual percentage of neutrophils, using the flow cytometric estimate, had R2 of 0.77. Regression of the manual percentage of neutrophils value on the flow cytometric percentage of neutrophil value was close to 1.0, with only a small negative intercept. Some additional refinement of flow cytometric procedures may be required before flow cytometric estimates of percentage of neutrophils can be accepted without simultaneous validation by use of manual counting. In particular, causes of day-to-day variation in flow cytometric results should be identified and reduced.