- Email: [email protected]
Comparison of Milk Somatic Cell Counts by Coulter and Fossomatic Counters1
Comparison of Milk Somatic Cell Counts by Coulter and Fossomatic Counters R. H. M I L L E R , M. J. PAAPE, and J. C. ACTON Milk Secretion and Mastitis Laboratory Agricultural Research Service US Department of Agriculture Beltsville, MD 20705 ABSTRACT
means of original milk samples from 14 of the quarters obtained by direct microscopic, Fossomatic, and Coulter counts were 199,300, 311,000, and 399,300, respectively.
Unpreserved milk samples from 28 quarters of 18 cows were used to compare milk somatic cell counts obtained by Fossomatic and Coulter Counter and to determine effect of temperature and sample age on Fossomatic counts. Samples represented high and low cell count milk (16 cows) and colostrum (2 cows). Fifteen milliliters of both foremilk (after milking preparation) and strippings were obtained; one-third was used for Coulter and two-thirds for Fossomatic. Milk for Fossomatic was subdivided once for fresh and 24-h determinations and once again for heating to 40 and 60°C for 15 min. Analysis of log10 count included effects for quarter, Fossomatic versus Coulter, and (for Fossomatic) sample age, incubation temperature, and agetemperature interaction. For foremilk, geometric means of Coulter milk somatic cell counts and Fossomatic counts were not different. For Fossomatic, milk samples incubated at 60°C counted higher than those at 40°C (230,096 versus 173,638); 24-h samples counted higher than fresh (201,679 versus 192,380). For strippings, Coulter counted higher than Fossomatie (700,521 versus 570,033). Interaction of time and temperature was significant for Fossomatic. Counts from samples held 24 h and heated to 60°C were highest (553,291). Fossomatic counts from flesh samples at 40°C were lowest (447,729). Geometric
INTRODUCTION
Received October 23, 1985. 1Mention of product trade names does not constitute approval by the United States Department of Agriculture. 1986 J Dairy Sci 69:1942--1946
Counting of somatic cells in milk (MSCC) is important in mastitis research programs. Two machines commonly used for cell counting are the Fossomatic 1 (Foss) and Coulter Counter. 1 Schmidt-Madsen (12) reported that Foss and Coulter counts had equally high correlations with direct microscopic somatic cell counts (DMSCC). Heald et al. (5) compared Foss with DMSCC and concluded that Foss gave satisfactory results for central laboratory use in Dairy Herd Improvement records. Other work (3, 6, 7) has shown that Foss and Coulter results may differ significantly, because the operating principles of the two instruments are different. Foss counting is based on fluoro-opfical principles specific for nuclear material; Coulter counts particles above a fixed size. Hoare et al. (7) showed that Coulter counts were consistently higher than Foss, although heating Coulter samples to 55°C for 15 rain prior to fixation reduced the difference, presumably due to the removal o f noncellular particles, rather than gas bubbles, as suggested previously (2). Hill et al. (6) reported that samples to be counted on the Coulter Counter, which were heated to 55°C for 30 min prior to fixation, gave counts equivalent to those from Foss. Heald et al. (5) studied effects of sample preservation and holding temperature on Foss cell count results. Further comparison of the results of Foss and Coulter under different conditions is needed to determine whether counts made by the two machines can be used interchangeably. In our laboratory, we recently converted from Coulter to Foss; we wished to determine optimum incubation procedure for counting
1942
COMPARISON OF COULTER AND FOSSOMATIC cells in fresh milk by Foss within 24 h after collection. Further, we wished to know how counts made by Foss compared with counts by Coulter. MATERIALS AND METHODS
Collections of milk samples were made during two different trials. Milk samples from 14 different udder quarters (28 total) were obtained in each trial, representing 18 different cows. Cows were chosen to represent a wide range of cell counts, based on California Mastitis Test results. Fifteen milliliters of foremilk (following premilking preparation) and 15 ml stripping milk (after machine removal) were obtained. In each trial, all were obtained at morning milking on the same day. No preservative was used, because we were concerned only with counts on fresh milk. After collection, each original milk sample (foremilk or strippings) was divided into three parts. One part was used for counting by Coulter; a second part was used for counting by Foss on the same day as the sample was taken; the third part was used for counting by Foss on the day after the sample was obtained. Coulter counts were made according to Newbould (9). The instrument was calibrated to count all particles 4.4/~m in diameter and larger. Cells in milk were fixed by adding 3 drops of somafix to 10 ml milk and incubating at 60°C for 5 min. For stripping samples, a 1:500 dilution of milk to Somoton, rather than the customary 1:100 concentration, was made to disperse their elevated fat content. The rationale for this revised procedure was discussed by Schuhze and Paape (13). For Foss counts, milk was further subdivided for two alternative treatments: incubation at 40 or 60°C for 15 rain prior to counting. Heating at 40°C for 15 rain is the current recommended procedure (4). The 60°C temperature was selected to damage the somatic cells in milk, thus making the cell membranes more permeable to the fluorescent dye, ethidium bromide. This dye penetrates the wall and protoplasm of the cell and forms a complex with the deoxyribonucleic acid (DNA)
2 Heald, C. W. Somatic cell count samples. Department of Animal Science, Pennsylvania State University, University Park 16802.
1943
in the cell nucleus. Duplicate readings were made on each milk sample fraction by Foss and by Coulter. Machines used were a Coulter model TA-II and a Foss model 215. Machine counts were checked at beginning of study using reference milk samples of Heald. 2 Geometric mean DMSCC for duplicate determinations on six reference samples was 707,900, whereas geometric mean MSCC were 691,800 and 660,700 for Foss and Coulter, respectively. To test further the hypothesis that heating cells to 60°C is necessary to allow penetration of ethidium bromide, the percentage of milk somatic ceils fluorescing after incubation at 40 and 60°C was determined microscopically. Milk was collected from an uninfected mammary quarter from each of 2 cows. The MSCC averaged .7 x 106 cells/ml. Milk was centrifuged at 1800 rpm (5°C, 15 min) and the cream layer and approximately one-third of the supernatant were discarded. Pellet was resuspended in the skimmed milk and the sample divided into two aliquots and heated at either 40 or 60°C for 15 rain. Samples were counted on the Foss electronic cell counter and a sample of milk containing ethidium bromide dye and buffer was removed after counting for microscopic examination. Cells within a microscopic field were first counted with visible light using a 40x objective and again using fluorescent light. Ten fields, approximately 50 cells, were examined. In the second trial, milk samples were also prepared for DMSCC (8) and stained with the DNA-ribonucleic acid (RNA)-specific stain pyronin Y-methyl green (10). Analysis of Data
Analyses of variance were conducted to compare results of Foss and Coulter electronic ceil counters and to determine how sample age and incubation temperature affected Foss results. Separate analyses were conducted for DMSCC on foremilk and stripping milk. The statistical model included effects of quarter, machine, interaction o f quarter by machine and within Foss results, effects for sample age, incubation temperature, interaction of age by temperature, quarter by age, quarter by temperature, and quarter by age by temperature. Quarters were assumed random ;thus, interaction of quarter by machine was used to test significance of machine differences. Also, interactions Journal of Dairy Science Vol. 69, No. 7, 1986
1944
MILLER ET AL.
TABLE 1. Analysis of variance of somatic cell count (logl0) determinations on foremilk samples.
23 1
172o** 12.43"*
were n o t different. F o r Foss counts, incubation t e m p e r a t u r e had a large e f f e c t (P<.01), whereas the effect of sample age was small (P<.05). There was no interaction o f age and temperature. Analysis of variance for counts on stripping milk samples is in Table 2. Quarter variation (random) was very large. Foss and Coulter counts were different (P<.01). There was an interaction of age by t e m p e r a t u r e for Foss counts (P<.05), which was in contrast to the counts for foremilk samples. Table 3 gives g e o m e t r i c mean MSCC for machine type for f o r e m i l k and strippings samples. Coulter counts were higher than Foss for b o t h types of milk samples, although significantly so o n l y for strippings (700,521 versus 570,033). These results indicate that MSCC m a d e on f o r e m i l k by Foss and by Coulter are comparable, b u t counts made on stripping milk by the different machines c a n n o t be compared. Table 4 gives g e o m e t r i c m e a n MSCC by sample age and incubation t e m p e r a t u r e for the Foss results. For foremilk, there was no interaction o f age and temperature, so that MSCC for 6 0 ° C were m a r k e d l y higher than those for 40°C, regardless of w h e t h e r MSCC were run on fresh samples or on samples stored for 24 h (at least 50,000 higher in each case). A t a constant incubation temperature, 24-h samples counted higher than same-day samples, but the difference was small and nonsignificant. The interaction o f sample age and incubation t e m p e r a t u r e for Foss counts on stripping milk is shown by the means in Table 4. Incubation at 6 0 ° C rather than at 4 0 ° C resulted in a greater MSCC increase for same-day samples than it did for samples stored 24 h. This is probably due to greater damage to cells in fresh samples due to heating to 60°C, thus allowing dye to reach ceil nuclei.
1 1 1 23 23 23 23 120
4.35* 25.86* * 4.78*
TABLE 3. Geometric mean somatic cell counts (MSCC) for Fossomatic and Coulter Counter results.
df Quarter (qtr) Fossomatic (Foss) vs. Coulter Within Foss Fresh vs. 24 h 40 vs. 60°C Age X temperature (temp) Qtr X age Qtr X temp Qtr X age X temp Qtr x Machine Residual
27 1 1 1 1 27 27 27 27 140
F-Ratio 1197"* <1 4.59* 75.89** 1.33
*P<.05. * *P<.O1.
with quarter were used to test age, t e m p e r a t u r e , and age by t e m p e r a t u r e effects. Because the only unbalanced aspect of the data structure was the greater n u m b e r of MSCC for Foss c o m p a r e d with Coulter, ordinary analysis of variance was used rather than least squares. The MSCC were transformed to logi0. RESULTS A N D DISCUSSION
Analysis of variance results are in Table 1 for foremilk. Quarter was the largest source of variation (random). Foss and Coulter counts
TABLE 2. Analysis of variance of somatic cell count (logl0) determinations on stripping samples. df Quarter (qtr) Fossomatic (Foss) vs. Coulter Within Foss Fresh vs. 24 h 40 vs. 60°C Age × temperature (temp) Qtr X age Qtr X temp Qtr x age × temp Qtr x Machine Residual
F-Ratio
*P<.05.
**P<.Ol. Journal of Dairy Science Vol. 69, No. 7, 1986
Geometric mean MSCC
Fossomatic Coulter
Foremilk
Strippings
199,883 211,549
570,033 700,521
COMPARISON OF COULTER AND FOSSOMATIC TABLE 4. Effect of age and incubation temperature and sample age on Fossomatic somatic cell count (MSCC) results.
Age
Temperature
Geometric mean MSCC Foremilk Strippings
Fresh Fresh 24 h 24 h
40°C 60°C 40°C 60°C
164,821 224,547 182,926 235,783
447,729 547,729 490,196 553,291
Tables 3 and 4 show that the higher MSCC from heating Foss samples to 60°C correspond more closely to the Coulter Counter results, especially for stripping milk samples. However, Foss counts obtained for all conditions of sample age and incubation temperature were lower than the corresponding results by Coulter Counter. These results indicate that fresh, unpreserved samples can be successfully counted by Foss if first incubated at 60°C to allow for penetration of the ethidium bromide dye into the cell. The small auxiliary study made to determine how the different incubation temperatures affected uptake of ethidium bromide by milk somatic cells showed that for samples heated at 40 and 60°C, DMSCC averaged 3.1 and 3.4 x 106/ml. Percentage o f fluorescing cells averaged 78 and 99%, respectively. Results indicate that a greater percentage of cells absorb the dye at 60°C than at 40°C. In trial 2 (second set of 14 quarter samples), each original milk sample was also counted by DMSCC. The geometric mean DMSCC were lower than means of both Foss and Coulter. Geometric means for foremilk were 91,600, 148,500, and 139,200 for DMSCC, Foss, and Coulter, respectively. Corresponding means for stripping samples were 199,300, 311,000, and 399,300. (Foss MSCC were averaged over the four sample age and temperature combinations.) Because MSCC by Coulter are not DNAspecific, elevated Coulter MSCC are probably caused by counting cell fragments, protein aggregates, and various artifacts contained in milk, as reported by other workers. Paape and Tucker (11) found large numbers of RNA-positire anucleate fragments in all fractions of milk. Hoare et al. (7) attributed falsely elevated Coulter counts to the presence of noncellular
1945
particles rather than to air bubbles. Hill et al. (6) suggested that Coulter counts included casein micelle aggregates. Brooker (1) examined pellets from centrifuged milk and concluded that cell fragments were small in size, possessed microvilli, and contained fat droplets and casein micelles. Similar results have been previously demonstrated for goat's milk, which contains many cell fragments thought to originate from alveolar secretory cells (3). In the latter study, particles containing protein and lipid were counted as cells by the Coulter Counter. Cell fragments are more frequent in stripping milk compared with foremilk. Paape and Tucker (11 ) showed that RNA-positive anucleate particles were much more numerous in stripping milk than in primary milk. Brooker (1) formed similar conclusions. Heald et al. (5) argued that Foss may count large nuclear fragments as cells, whereas DMSCC would omit such particles from counts. If this is true, even Foss counts may be too high depending on frequency of such particles. Our study agreed with this conjecture because Foss counts were higher than DMSCC. In summary, our results indicate that MSCC obtained from the Coulter electronic cell counter are higher than counts obtained from the Foss electronic cell counter when counting stripping or fresh milk with high cell count. The DMSCC were lower than counts obtained by Foss. Unpreserved milk samples counted on the Foss incubated at 60°C for 15 min before counting gives a higher count than heating to 40°C, probably due to greater dye penetration. Fresh milk samples can be counted successfully by Foss.
ACKNOWLEDGMENTS
The technical assistance of Anne M. Dulin and Loren A. Fulton is gratefully acknowledged. REFERENCES
1 Brooker, B. E. 1978. Characteristic cell fragments in bovine milk. J, Dairy Res. 45:21. 2 Dijkman, A. J., H. Breunissen, J.M.J. Van Der Leer, J. Perriens, H. Verberne, and F.H.J. Jaartsveld. 197% Entrapped air as a cause of erroneous milk cell counts (Coulter Counter). Neth. Milk Dairy J. 33:155. 3 Dulin, A. M.~ M. J. Paape, and W. P. Wergin. 1982. Differentiation and enumeration of somatic cells in goat milk. J. Food Prot. 45:435. Journal of Dairy Science Vol. 69, No. 7, 1986
1946
MILLER ET AL.
4 Foss Electric A/S. 1979, Page 2 in Fossomatic 15600 operating instructions. 5 Heald, C. W., G. M. Jones, S. C. Nickerson, W. N. Patterson, and W. E. Vinson. 1977. Preliminary evaluation o f the Fossomatic cell counter for analysis of individual cow samples in a central testing laboratory. J. Food Prot. 40:523. 6 Hill, A. W., K. G. Hibbitt, and J. Davies. 1982. Particles in bulk milk capable o f causing falsely high electronic cell counts. J. Dairy Res. 49:171. 7 Hoare, R.J.T., P. J. Nicholls, and R. F. Sheldrake. 1982. Investigations into falsely elevated somatic cell counts of bulked herd milk. J. Dairy Res. 49:559. 8 National Mastitis Council S u b c o m m i t t e e on Screening Tests. 1968. Direct microscopic somatic cell count in milk. J. Milk Food Technol. 31:350.
Journal o f Dairy Science Vol. 69, No. 7, 1986
9 Newbould, F.H.S. 1974. Electronic counting o f somatic cells in farm bulk tank milk. J. Milk Food Technol. 37:504. 10 Paape, M. J., H. D. Hafs, and W. W. Snyder. 1963. Variation of estimated n u m b e r s of milk somatic cells stained with Wright's stain or pyronin Ym e t h y l green stain. J. Dairy S ci. 46 : 1211. 11 Paape, M. J., and H. A. Tucker. 1966. Somatic cell c o n t e n t variation in fraction-collected milk. J. Dairy Sci. 49:265. 12 Schmidt-Madsen, P. 1975. Fluoro-opto-electronic cell counting on milk. J. Dairy Res. 42:227. 13 Schultze, W..D., and M. J. Paape. 1984. Effect on o u t c o m e of i n t r a m a m m a r y challenge exposure with Staphylococcus aureus of somatic cell concentration and presence o f an i n t r a m a m m a r y device. Am. J. Vet. Res. 45:420.
means of original milk samples from 14 of the quarters obtained by direct microscopic, Fossomatic, and Coulter counts were 199,300, 311,000, and 399,300, respectively.
Unpreserved milk samples from 28 quarters of 18 cows were used to compare milk somatic cell counts obtained by Fossomatic and Coulter Counter and to determine effect of temperature and sample age on Fossomatic counts. Samples represented high and low cell count milk (16 cows) and colostrum (2 cows). Fifteen milliliters of both foremilk (after milking preparation) and strippings were obtained; one-third was used for Coulter and two-thirds for Fossomatic. Milk for Fossomatic was subdivided once for fresh and 24-h determinations and once again for heating to 40 and 60°C for 15 min. Analysis of log10 count included effects for quarter, Fossomatic versus Coulter, and (for Fossomatic) sample age, incubation temperature, and agetemperature interaction. For foremilk, geometric means of Coulter milk somatic cell counts and Fossomatic counts were not different. For Fossomatic, milk samples incubated at 60°C counted higher than those at 40°C (230,096 versus 173,638); 24-h samples counted higher than fresh (201,679 versus 192,380). For strippings, Coulter counted higher than Fossomatie (700,521 versus 570,033). Interaction of time and temperature was significant for Fossomatic. Counts from samples held 24 h and heated to 60°C were highest (553,291). Fossomatic counts from flesh samples at 40°C were lowest (447,729). Geometric
INTRODUCTION
Received October 23, 1985. 1Mention of product trade names does not constitute approval by the United States Department of Agriculture. 1986 J Dairy Sci 69:1942--1946
Counting of somatic cells in milk (MSCC) is important in mastitis research programs. Two machines commonly used for cell counting are the Fossomatic 1 (Foss) and Coulter Counter. 1 Schmidt-Madsen (12) reported that Foss and Coulter counts had equally high correlations with direct microscopic somatic cell counts (DMSCC). Heald et al. (5) compared Foss with DMSCC and concluded that Foss gave satisfactory results for central laboratory use in Dairy Herd Improvement records. Other work (3, 6, 7) has shown that Foss and Coulter results may differ significantly, because the operating principles of the two instruments are different. Foss counting is based on fluoro-opfical principles specific for nuclear material; Coulter counts particles above a fixed size. Hoare et al. (7) showed that Coulter counts were consistently higher than Foss, although heating Coulter samples to 55°C for 15 rain prior to fixation reduced the difference, presumably due to the removal o f noncellular particles, rather than gas bubbles, as suggested previously (2). Hill et al. (6) reported that samples to be counted on the Coulter Counter, which were heated to 55°C for 30 min prior to fixation, gave counts equivalent to those from Foss. Heald et al. (5) studied effects of sample preservation and holding temperature on Foss cell count results. Further comparison of the results of Foss and Coulter under different conditions is needed to determine whether counts made by the two machines can be used interchangeably. In our laboratory, we recently converted from Coulter to Foss; we wished to determine optimum incubation procedure for counting
1942
COMPARISON OF COULTER AND FOSSOMATIC cells in fresh milk by Foss within 24 h after collection. Further, we wished to know how counts made by Foss compared with counts by Coulter. MATERIALS AND METHODS
Collections of milk samples were made during two different trials. Milk samples from 14 different udder quarters (28 total) were obtained in each trial, representing 18 different cows. Cows were chosen to represent a wide range of cell counts, based on California Mastitis Test results. Fifteen milliliters of foremilk (following premilking preparation) and 15 ml stripping milk (after machine removal) were obtained. In each trial, all were obtained at morning milking on the same day. No preservative was used, because we were concerned only with counts on fresh milk. After collection, each original milk sample (foremilk or strippings) was divided into three parts. One part was used for counting by Coulter; a second part was used for counting by Foss on the same day as the sample was taken; the third part was used for counting by Foss on the day after the sample was obtained. Coulter counts were made according to Newbould (9). The instrument was calibrated to count all particles 4.4/~m in diameter and larger. Cells in milk were fixed by adding 3 drops of somafix to 10 ml milk and incubating at 60°C for 5 min. For stripping samples, a 1:500 dilution of milk to Somoton, rather than the customary 1:100 concentration, was made to disperse their elevated fat content. The rationale for this revised procedure was discussed by Schuhze and Paape (13). For Foss counts, milk was further subdivided for two alternative treatments: incubation at 40 or 60°C for 15 rain prior to counting. Heating at 40°C for 15 rain is the current recommended procedure (4). The 60°C temperature was selected to damage the somatic cells in milk, thus making the cell membranes more permeable to the fluorescent dye, ethidium bromide. This dye penetrates the wall and protoplasm of the cell and forms a complex with the deoxyribonucleic acid (DNA)
2 Heald, C. W. Somatic cell count samples. Department of Animal Science, Pennsylvania State University, University Park 16802.
1943
in the cell nucleus. Duplicate readings were made on each milk sample fraction by Foss and by Coulter. Machines used were a Coulter model TA-II and a Foss model 215. Machine counts were checked at beginning of study using reference milk samples of Heald. 2 Geometric mean DMSCC for duplicate determinations on six reference samples was 707,900, whereas geometric mean MSCC were 691,800 and 660,700 for Foss and Coulter, respectively. To test further the hypothesis that heating cells to 60°C is necessary to allow penetration of ethidium bromide, the percentage of milk somatic ceils fluorescing after incubation at 40 and 60°C was determined microscopically. Milk was collected from an uninfected mammary quarter from each of 2 cows. The MSCC averaged .7 x 106 cells/ml. Milk was centrifuged at 1800 rpm (5°C, 15 min) and the cream layer and approximately one-third of the supernatant were discarded. Pellet was resuspended in the skimmed milk and the sample divided into two aliquots and heated at either 40 or 60°C for 15 rain. Samples were counted on the Foss electronic cell counter and a sample of milk containing ethidium bromide dye and buffer was removed after counting for microscopic examination. Cells within a microscopic field were first counted with visible light using a 40x objective and again using fluorescent light. Ten fields, approximately 50 cells, were examined. In the second trial, milk samples were also prepared for DMSCC (8) and stained with the DNA-ribonucleic acid (RNA)-specific stain pyronin Y-methyl green (10). Analysis of Data
Analyses of variance were conducted to compare results of Foss and Coulter electronic ceil counters and to determine how sample age and incubation temperature affected Foss results. Separate analyses were conducted for DMSCC on foremilk and stripping milk. The statistical model included effects of quarter, machine, interaction o f quarter by machine and within Foss results, effects for sample age, incubation temperature, interaction of age by temperature, quarter by age, quarter by temperature, and quarter by age by temperature. Quarters were assumed random ;thus, interaction of quarter by machine was used to test significance of machine differences. Also, interactions Journal of Dairy Science Vol. 69, No. 7, 1986
1944
MILLER ET AL.
TABLE 1. Analysis of variance of somatic cell count (logl0) determinations on foremilk samples.
23 1
172o** 12.43"*
were n o t different. F o r Foss counts, incubation t e m p e r a t u r e had a large e f f e c t (P<.01), whereas the effect of sample age was small (P<.05). There was no interaction o f age and temperature. Analysis of variance for counts on stripping milk samples is in Table 2. Quarter variation (random) was very large. Foss and Coulter counts were different (P<.01). There was an interaction of age by t e m p e r a t u r e for Foss counts (P<.05), which was in contrast to the counts for foremilk samples. Table 3 gives g e o m e t r i c mean MSCC for machine type for f o r e m i l k and strippings samples. Coulter counts were higher than Foss for b o t h types of milk samples, although significantly so o n l y for strippings (700,521 versus 570,033). These results indicate that MSCC m a d e on f o r e m i l k by Foss and by Coulter are comparable, b u t counts made on stripping milk by the different machines c a n n o t be compared. Table 4 gives g e o m e t r i c m e a n MSCC by sample age and incubation t e m p e r a t u r e for the Foss results. For foremilk, there was no interaction o f age and temperature, so that MSCC for 6 0 ° C were m a r k e d l y higher than those for 40°C, regardless of w h e t h e r MSCC were run on fresh samples or on samples stored for 24 h (at least 50,000 higher in each case). A t a constant incubation temperature, 24-h samples counted higher than same-day samples, but the difference was small and nonsignificant. The interaction o f sample age and incubation t e m p e r a t u r e for Foss counts on stripping milk is shown by the means in Table 4. Incubation at 6 0 ° C rather than at 4 0 ° C resulted in a greater MSCC increase for same-day samples than it did for samples stored 24 h. This is probably due to greater damage to cells in fresh samples due to heating to 60°C, thus allowing dye to reach ceil nuclei.
1 1 1 23 23 23 23 120
4.35* 25.86* * 4.78*
TABLE 3. Geometric mean somatic cell counts (MSCC) for Fossomatic and Coulter Counter results.
df Quarter (qtr) Fossomatic (Foss) vs. Coulter Within Foss Fresh vs. 24 h 40 vs. 60°C Age X temperature (temp) Qtr X age Qtr X temp Qtr X age X temp Qtr x Machine Residual
27 1 1 1 1 27 27 27 27 140
F-Ratio 1197"* <1 4.59* 75.89** 1.33
*P<.05. * *P<.O1.
with quarter were used to test age, t e m p e r a t u r e , and age by t e m p e r a t u r e effects. Because the only unbalanced aspect of the data structure was the greater n u m b e r of MSCC for Foss c o m p a r e d with Coulter, ordinary analysis of variance was used rather than least squares. The MSCC were transformed to logi0. RESULTS A N D DISCUSSION
Analysis of variance results are in Table 1 for foremilk. Quarter was the largest source of variation (random). Foss and Coulter counts
TABLE 2. Analysis of variance of somatic cell count (logl0) determinations on stripping samples. df Quarter (qtr) Fossomatic (Foss) vs. Coulter Within Foss Fresh vs. 24 h 40 vs. 60°C Age × temperature (temp) Qtr X age Qtr X temp Qtr x age × temp Qtr x Machine Residual
F-Ratio
*P<.05.
**P<.Ol. Journal of Dairy Science Vol. 69, No. 7, 1986
Geometric mean MSCC
Fossomatic Coulter
Foremilk
Strippings
199,883 211,549
570,033 700,521
COMPARISON OF COULTER AND FOSSOMATIC TABLE 4. Effect of age and incubation temperature and sample age on Fossomatic somatic cell count (MSCC) results.
Age
Temperature
Geometric mean MSCC Foremilk Strippings
Fresh Fresh 24 h 24 h
40°C 60°C 40°C 60°C
164,821 224,547 182,926 235,783
447,729 547,729 490,196 553,291
Tables 3 and 4 show that the higher MSCC from heating Foss samples to 60°C correspond more closely to the Coulter Counter results, especially for stripping milk samples. However, Foss counts obtained for all conditions of sample age and incubation temperature were lower than the corresponding results by Coulter Counter. These results indicate that fresh, unpreserved samples can be successfully counted by Foss if first incubated at 60°C to allow for penetration of the ethidium bromide dye into the cell. The small auxiliary study made to determine how the different incubation temperatures affected uptake of ethidium bromide by milk somatic cells showed that for samples heated at 40 and 60°C, DMSCC averaged 3.1 and 3.4 x 106/ml. Percentage o f fluorescing cells averaged 78 and 99%, respectively. Results indicate that a greater percentage of cells absorb the dye at 60°C than at 40°C. In trial 2 (second set of 14 quarter samples), each original milk sample was also counted by DMSCC. The geometric mean DMSCC were lower than means of both Foss and Coulter. Geometric means for foremilk were 91,600, 148,500, and 139,200 for DMSCC, Foss, and Coulter, respectively. Corresponding means for stripping samples were 199,300, 311,000, and 399,300. (Foss MSCC were averaged over the four sample age and temperature combinations.) Because MSCC by Coulter are not DNAspecific, elevated Coulter MSCC are probably caused by counting cell fragments, protein aggregates, and various artifacts contained in milk, as reported by other workers. Paape and Tucker (11) found large numbers of RNA-positire anucleate fragments in all fractions of milk. Hoare et al. (7) attributed falsely elevated Coulter counts to the presence of noncellular
1945
particles rather than to air bubbles. Hill et al. (6) suggested that Coulter counts included casein micelle aggregates. Brooker (1) examined pellets from centrifuged milk and concluded that cell fragments were small in size, possessed microvilli, and contained fat droplets and casein micelles. Similar results have been previously demonstrated for goat's milk, which contains many cell fragments thought to originate from alveolar secretory cells (3). In the latter study, particles containing protein and lipid were counted as cells by the Coulter Counter. Cell fragments are more frequent in stripping milk compared with foremilk. Paape and Tucker (11 ) showed that RNA-positive anucleate particles were much more numerous in stripping milk than in primary milk. Brooker (1) formed similar conclusions. Heald et al. (5) argued that Foss may count large nuclear fragments as cells, whereas DMSCC would omit such particles from counts. If this is true, even Foss counts may be too high depending on frequency of such particles. Our study agreed with this conjecture because Foss counts were higher than DMSCC. In summary, our results indicate that MSCC obtained from the Coulter electronic cell counter are higher than counts obtained from the Foss electronic cell counter when counting stripping or fresh milk with high cell count. The DMSCC were lower than counts obtained by Foss. Unpreserved milk samples counted on the Foss incubated at 60°C for 15 min before counting gives a higher count than heating to 40°C, probably due to greater dye penetration. Fresh milk samples can be counted successfully by Foss.
ACKNOWLEDGMENTS
The technical assistance of Anne M. Dulin and Loren A. Fulton is gratefully acknowledged. REFERENCES
1 Brooker, B. E. 1978. Characteristic cell fragments in bovine milk. J, Dairy Res. 45:21. 2 Dijkman, A. J., H. Breunissen, J.M.J. Van Der Leer, J. Perriens, H. Verberne, and F.H.J. Jaartsveld. 197% Entrapped air as a cause of erroneous milk cell counts (Coulter Counter). Neth. Milk Dairy J. 33:155. 3 Dulin, A. M.~ M. J. Paape, and W. P. Wergin. 1982. Differentiation and enumeration of somatic cells in goat milk. J. Food Prot. 45:435. Journal of Dairy Science Vol. 69, No. 7, 1986
1946
MILLER ET AL.
4 Foss Electric A/S. 1979, Page 2 in Fossomatic 15600 operating instructions. 5 Heald, C. W., G. M. Jones, S. C. Nickerson, W. N. Patterson, and W. E. Vinson. 1977. Preliminary evaluation o f the Fossomatic cell counter for analysis of individual cow samples in a central testing laboratory. J. Food Prot. 40:523. 6 Hill, A. W., K. G. Hibbitt, and J. Davies. 1982. Particles in bulk milk capable o f causing falsely high electronic cell counts. J. Dairy Res. 49:171. 7 Hoare, R.J.T., P. J. Nicholls, and R. F. Sheldrake. 1982. Investigations into falsely elevated somatic cell counts of bulked herd milk. J. Dairy Res. 49:559. 8 National Mastitis Council S u b c o m m i t t e e on Screening Tests. 1968. Direct microscopic somatic cell count in milk. J. Milk Food Technol. 31:350.
Journal o f Dairy Science Vol. 69, No. 7, 1986
9 Newbould, F.H.S. 1974. Electronic counting o f somatic cells in farm bulk tank milk. J. Milk Food Technol. 37:504. 10 Paape, M. J., H. D. Hafs, and W. W. Snyder. 1963. Variation of estimated n u m b e r s of milk somatic cells stained with Wright's stain or pyronin Ym e t h y l green stain. J. Dairy S ci. 46 : 1211. 11 Paape, M. J., and H. A. Tucker. 1966. Somatic cell c o n t e n t variation in fraction-collected milk. J. Dairy Sci. 49:265. 12 Schmidt-Madsen, P. 1975. Fluoro-opto-electronic cell counting on milk. J. Dairy Res. 42:227. 13 Schultze, W..D., and M. J. Paape. 1984. Effect on o u t c o m e of i n t r a m a m m a r y challenge exposure with Staphylococcus aureus of somatic cell concentration and presence o f an i n t r a m a m m a r y device. Am. J. Vet. Res. 45:420.