Milk differential leukocyte counts in relation to somatic cell count in Alpine × Beetal and Saanen × Beetal crossbred does

Introduction

Somatic cell count (SCC) is the standard indirect measure of udder health in dairy ruminants. It underpins milk payment schemes and udder health programmes worldwide (Contreras et al., 2007; Podhorecka et al., 2021) [4,17] . Moving it from cattle to goats is not simple, and the reason is physiological rather than numerical. Goats secrete milk by an apocrine process. This sheds cytoplasmic particles and cell fragments that contain DNA, and automated counters read these as somatic cells. Apparent SCC therefore rises without any inflammation (Paape and Capuco, 1997; Kaskous et al., 2023) [9,14] . As a result, normal SCC in goat milk averages several hundred thousand cells per millilitre and can reach 1.5 × 10⁶ cells/mL in clearly healthy animals. Most of the variation comes from lactation stage, parity, oestrus and season rather than from infection (Haenlein, 2002; Luengo et al., 2004; Jimenez-Granado et al., 2014; Smistad et al., 2021) [7,8,11], 21 . The SCC threshold for detecting intramammary infection in goats lies near 1.5 × 10⁶ cells/mL (Koop et al., 2011) [10] , about ten times the usual cattle figure. The cells themselves also differ. In a healthy cow quarter, macrophages are the main cell type, and neutrophils rise sharply only when infection occurs. In the goat, neutrophils are already the main cell type in milk from healthy, uninfected udders. They make up 40-80% of the somatic cell population, and they stay dominant in infected udders as well (Sierra et al., 1999; Paape et al., 2001; Bagnicka et al., 2011; Kaskous et al., 2023) [2,9,19] . The cattle rule macrophages mean health, neutrophils mean disease therefore does not hold in goats. Any diagnostic use of differential cytology in this species must start from goat baselines. Differential leukocyte counting (DLC) has been proposed as a way to get more from the goat somatic cell population than the total count gives (Souza et al., 2012; Alhussien and Dang, 2018; Kaskous et al., 2023) [1,9] . It divides the cell pool instead of only measuring its size, so it should be less affected by the particle problem. The approach has precedent in small ruminants: Cuccuru et al. (1997) [5] related differential cell counts to total counts in Sardinian ewes. Its interpretation, however, requires knowledge of how the leukocyte profile behaves across the normal SCC range in the population concerned. Such reference dgoatsarce for goats in general, and for Indian crossbred dairy goats they do not appear to exist at all. Alpine × Beetal and Saanen × Beetal are the main exotic × indigenous crosses bred for milk production in India, and no published work describes their milk leukocyte composition. The milk DLC was therefore measured across six SCC groups in clinically healthy does of both crosses under uniform management. The objectives were threefold: to characterise how the leukocyte profile changes across the normal SCC range; to determine whether the neutrophil-tomacrophage (N:M) ratio tracks that change more clearly than any single cell type; and to test whether the two crosses differ in their baseline composition, in their response to rising SCC, or in both.

Materials and methods

Study site

The study was conducted at the Livestock Research Centre, ICAR-National Dairy Research Institute, Karnal, Haryana, India (29°43′14.4″ N, 76°58′55.2″ E; 250 m above mean sea level). The region experiences summer maxima up to 45 °C and winter minima of 4-5 °C, with a diurnal range of 15-20 °C and mean annual rainfall of approximately 700 mm, falling mainly between early July and mid-September.

Experimental animals and management

Seventy-two crossbred does were enrolled: 36 Alpine × Beetal (AB) and 36 Saanen × Beetal (SB). All were in early lactation, between the third and sixth parity, and carried a body condition score of 3-4. Only clinically healthy does with no history of clinical mastitis were included; animals showing any abnormality of the udder or milk on physical examination were excluded at enrolment. The herd was housed in a covered pen with an attached open area and was stall-fed with ad libitum green fodder (maize) and a concentrate mixture (18% crude protein, 70% total digestible nutrients) formulated to NRC (2001) recommendations, with free access to water. Does were hand-milked twice daily at 04:00 and 16:00 h in a dedicated milking shed. Animals were groomed and washed before milking; udders were washed with clean water (lukewarm during winter months) and milkers washed their hands in 1:1000 chlorhexidine-cetrimide solution before milking. All does were managed identically throughout.

Milk sampling

Milk was collected at the evening milking. Udders were wiped with single-use aseptic wipes and the first jets discarded. Approximately 150 mL of pooled milk was collected from each doe by manual full-hand milking into sterile 50 mL tubes following National Mastitis Council guidelines and transported to the laboratory in insulated icepack boxes at approximately 4 °C for immediate processing. Each doe was sampled on three consecutive days, and every determination reported here somatic cell count and differential leukocyte count alike was performed on each of the three samples. The three values were then averaged to give a single figure per doe. The doe, not the sample, was therefore the experimental unit throughout, and all sample sizes given below refer to numbers of does.

Somatic cell counts and group allocation

Somatic cell count was determined by two independent methods on every sample. Automated counts were obtained on a Lactoscan SCC (Milkotronic Ltd., Stara Zagora, Bulgaria), which uses fluorescent image cytometry. Fresh milk (100 µL) was added to an Eppendorf tube pre-coated with lyophilised Sofia Green DNA-binding dye and vortexed 10-15 times. The tube was held for 2 min and vortexed again. A volume of 8.5 µL of the stained sample was then loaded onto a Lactochip cassette and counted. Direct microscopic counts were made in parallel on the Pyronin Y-methyl green stained smears described in Section 2.6, which also provided the differential count. Automated counters of this type have been evaluated against direct microscopic counts as indicators of intramammary infection in dairy goats (Persson and Olofsson, 2011) [16] . Group allocation used the mean of the three daily automated determinations. Does within each cross were assigned to six groups: <4, 4-6, 6-8, 8-10, 10-12 and >12 × 10⁵ cells/mL (n = 6 per group per cross). The two counting methods differ in what they enumerate. The automated method registers the DNA-bearing cytoplasmic particles characteristic of apocrine secretion; the microscopic count is restricted to nucleated cells (Section 2.6). The automated value is therefore the more conservative basis for allocation, and is the one reported throughout.

Differential leukocyte count

Differential counts were performed microscopically on stained milk smears, independently of the automated SCC. One smear was prepared from each of the three daily samples, giving three smears per doe; all three were read and the resulting differential counts averaged to a single value per animal. A thin smear was prepared by spreading 5-10 µL of fresh milk over 1 cm² of a clean glass slide and drying at 30-40 °C. Fat globules were removed by immersion in xylene for 20 min; slides were fixed in methanol for 1 min, flooded with distilled water for 1 min, and dried. Smears were stained for 5 min under subdued light with freshly prepared Pyronin Y-methyl green. The stain was prepared by dissolving 1 g Pyronin Y and 0.56 g methyl green in 196 mL distilled water, filtering through 10-12 µm filter paper and storing in an actinic bottle; it is light-sensitive and was prepared fresh. Excess stain was rinsed off with tap water. Slides were examined under oil immersion on an Olympus IX51 microscope (Olympus, Tokyo, Japan). Neutrophils were identified by their multilobed, bridged nucleus; lymphocytes by a densely stained, often eccentric nucleus with scant cytoplasm; and macrophages by their large size and abundant cytoplasm. Epithelial cells, identified by their smaller size and spherical nucleus, were excluded from the differential. Only round, nucleated cells with nuclei stained dark red by Pyronin Y-methyl green were enumerated. Cells smaller than 4 µm were disregarded; fragments were counted only when more than half of the nuclear material was visible; cell clusters were scored as a single unit unless nuclei were clearly separate; and doubtful objects were not counted. Counts are expressed as the percentage of each cell type within the total leukocyte population.

Neutrophil-to-macrophage ratio

The N:M ratio was computed for each doe individually as the quotient of her neutrophil and macrophage percentages, and these individual ratios were then averaged within group. Ratios were not derived from group mean percentages, since the mean of a ratio is not the ratio of means.

Statistical analysis

Analyses were performed in SPSS v22.0 (SPSS Inc., Chicago, IL, USA). Data are presented as mean ± standard error. Each response variable was examined by two-way analysis of variance with cross (2 levels), SCC group (6 levels) and their interaction as fixed effects. The cross × SCC interaction was the formal test of whether the two crosses differed in their response to rising SCC; main effects were interpreted only in the light of that term. Effect sizes are reported as partial η². The six SCC groups are ordered rather than nominal. The dose-response question was therefore addressed by a linear polynomial contrast within each cross (coefficients -5, -3, -1, +1, +3, +5). Pairwise post hoc comparison was not used for this purpose: it is underpowered for monotone trends and does not test the hypothesis of interest. Pairwise differences among groups were additionally examined by Tukey's honest significant difference test and are indicated by superscript letters in Tables 1 and2; where the pairwise and trend analyses diverge, both are reported. Significance was declared at P < 0.05. Neutrophil, macrophage and lymphocyte percentages are compositional: they are constrained to sum to 100 and therefore cannot vary independently. A decline in one fraction is in part the arithmetic complement of a rise in another. This constraint is acknowledged throughout the interpretation, and the reciprocal neutrophil-macrophage relationship is not treated as two independent findings.

Results

Tables 1 and2 give the differential leukocyte counts for both crosses across the six SCC groups. Table 4 summarises the percentage changes relative to the lowest group, and Table 3 gives the two-way analysis of variance. .33 a ± 1.52 Means within a column bearing different superscript letters differ significantly (Tukey's HSD, P < 0.05). N:M, neutrophil-to-macrophage ratio, computed per doe and then averaged. Linear trend across groups: neutrophil, macrophage and N:M, all P < 0.001; lymphocyte, P = 0.003. .83 e ± 0.17 Means within a column bearing different superscript letters differ significantly (Tukey's HSD, P < 0.05). N:M, neutrophil-to-macrophage ratio, computed per doe and then averaged. Linear trend across groups: all variables P < 0.001.

Neutrophils

Neutrophil percentage rose steadily across the SCC gradient in both crosses (Tables 1 and2; Fig. 1A). In AB does the mean rose from 52.00 ± 0.82% in the <4 × 10⁵ cells/mL group to 72.50 ± 0.76% in the >12 × 10⁵ group, an increase of 39.4%. In SB does it rose from 54.33 ± 0.49% to 75.33 ± 0.49%, an increase of 38.7% (Table 4). The linear trend was highly significant in both crosses (AB, t(30) = 19.56, P < 0.001; SB, t(30) = 42.37, P < 0.001). SCC group explained most of the variance (partial η² = 0.956; Table 3). From 6-8 × 10⁵ cells/mL upward, every group differed from the lowest by Tukey's HSD. Even the lowest-SCC does were neutrophil-dominant. At <4 × 10⁵ cells/mL, more than half of all milk leukocytes were neutrophils in both crosses. Bars are means ± SE (n = 6). Letters denote Tukey's HSD groupings within cross. The cross × SCC interaction P-value from two-way ANOVA is shown in each panel.

Macrophages

Macrophage percentage fell in the opposite direction: from 33.00 ± 0.93% to 18.17 ± 0.95% in AB (-44.9%) and from 36.33 ± 0.49% to 17.50 ± 0.43% in SB (-51.8%). Both linear trends were significant (P < 0.001; Fig. 1B; Table 4).

In SB, all six groups separated cleanly and carried distinct superscripts. Fig. 2 shows how the leukocyte pool redistributes across the gradient. As Section 2.8 notes, this fall is partly the arithmetic mirror of the neutrophil rise and is not independent of it.

Neutrophil-to-macrophage ratio

The N:M ratio rose from 1.58 ± 0.02 to 4.04 ± 0.20 in AB (+155.7%) and from 1.50 ± 0.03 to 4.32 ± 0.13 in SB (+188.0%). Both linear trends were highly significant (P < 0.001; Fig. 1C). An exponential function of SCC fitted the data closely in both crosses (AB, N:M = 1.18e^0.091•SCC, R² = 0.984; SB, N:M = 1.07e^0.105•SCC, R² = 0.995; Fig. 3). Of all the measures examined, the ratio changed the most in proportional terms, and at low SCC its standard errors were the smallest relative to the mean. The ratio more than doubles across a somatic cell count range that lies within the reported physiological range for healthy goats, which is the principal caution against using it as a standalone diagnostic index.

Lymphocytes

Lymphocyte percentage fell across the gradient in both crosses: from 15.00 ± 1.73% to 9.33 ± 1.52% in AB (-37.8%) and from 10.17 ± 0.21% to 6.83 ± 0.17% in SB (-32.8%). Pairwise post hoc testing found no significant differences among the AB groups, all six-sharing superscript a (Table 1). The SB groups separated clearly (Table 2). This difference comes from the precision of the two datasets, not from the underlying response. The mean standard error was 1.48 percentage points for AB lymphocytes against 0.18 for SB about eight times larger. A linear contrast uses the ordered structure of the groups, and on this test the decline was significant in both crosses (AB, t(30) = -3.28, P = 0.003; SB, t(30) = -14.18, P < 0.001). The proportional decline was in fact slightly larger in AB than in SB (Table 4). The apparent stability of AB lymphocytes under pairwise testing is therefore an artefact of low precision, not evidence that the response is absent (Fig. 1D).

Comparison between crosses

Two-way analysis of variance (Table 3) showed significant main effects of cross for neutrophils (F(1,60) = 70.45, P < 0.001), macrophages (F(1,60) = 7.26, P = 0.009) and lymphocytes (F(1,60) = 50.13, P < 0.001). Across all groups, SB does carried on average 3.3 percentage points more neutrophils and 4.4 percentage points fewer lymphocytes than AB does. The N:M ratio did not differ between the crosses overall (P = 0.127). The cross × SCC interaction was not significant for neutrophils (F(5,60) = 1.02, P = 0.411), N:M (F(5,60) = 2.31, P = 0.055) or lymphocytes (F(5,60) = 0.84, P = 0.528).

For macrophages it was only marginally significant (F(5,60) = 2.84, P = 0.023). The two crosses therefore differ in their baseline leukocyte composition, but they respond to rising SCC in much the same way. 1 and2) and two-way ANOVA (Table 3).

Discussion

The milk leukocyte profile of clinically healthy crossbred does change sharply and steadily across the SCC range normally regarded as healthy for goats. Neutrophils rose by about two-fifths, macrophages fell by almost half, lymphocytes declined, and the N:M ratio more than doubled. All of this happened within a gradient of less than 400 to more than 1200 × 10³ cells/mL. Where that gradient sits is what governs how it should be read. Published work puts normal goat milk SCC at a mean of about 764 × 10³ cells/mL, ranging from 200 to 1500 × 10³, and reports uninfected goats between 270 and 2000 × 10³ cells/mL (Kaskous et al., 2023) [9] . Latent class analysis suggests an infection cutoff near 1500 × 10³ cells/mL (Koop et al., 2011) [10] , and the median of more than a million Norwegian goat milk records was 440 × 10³ cells/mL (Smistad et al., 2021). The range studied therefore lies almost entirely within what a healthy, uninfected goat udder can produce. Finite mixture modelling of goat somatic cell scores points the same way, suggesting that the caprine SCC distribution contains more than one underlying population rather than a single graded one (Boettcher et al., 2005) [3] . It would be tempting to read the rising neutrophil fraction as subclinical mastitis, but that reading does not hold. No bacteriological culture was performed, the animals were clinically sound, and the SCC range never reaches the level at which infection becomes likely in this species. Goat udders with high SCC have been examined histologically and found free of mastitis (Kaskous et al., 2023) [9] . More than 90% of the variation in goat SCC has been traced to physiological rather than bacteriological causes (Luengo et al., 2004; Jimenez-Granado et al., 2014) [11] . Min et al. (2007) [8,12] found that SCC in goat milk does not correlate strongly with intramammary infection at all, and stage of lactation interacts with subclinical mastitis in shaping goat milk quality (Silanikove et al., 2014) [20] . The shift described here is better read as the cellular signature of graded cell mobilisation in a healthy gland. The baseline values reported here support this. In the lowest group, neutrophils already made up 52.0% (AB) and 54.3% (SB) of milk leukocytes. That sits squarely inside the 40-80% reported for healthy uninfected goats (Sierra et [9,19] , and it is nothing like the macrophage-dominant profile of a healthy cow quarter. Any framework that treats neutrophil dominance in goat milk as inherently abnormal would misclassify every healthy doe in this study. These data also refine the findings of Bagnicka et al. (2011) [2] , who reported that mastitis raises neutrophil content in ewe milk but not in goat milk, where neutrophils dominate regardless. Dominance did persist here, but its degree was far from fixed: it ranged from roughly 3:2 to better than 4:1 across the gradient. The N:M ratio responded most steeply and most consistently, following SCC as a near-exponential curve (Fig. 3). Its size comes from combining a rising numerator with a falling denominator. That is both its strength and its weakness, because the underlying percentages are constrained to sum to 100, so part of its sensitivity is arithmetic rather than biological. It is still useful as a summary measure, because both numbers come from the same slide. This makes it immune to the particle problem that inflates automated SCC in goats. Fluorescent cytometry counts the roughly 150 × 10³ DNA-bearing particles per mL of goat milk as if they were cells (Paape and Capuco, 1997; Kaskous et al., 2023) [9,14] . A microscopic count limited to nucleated cells above 8 µm does not. The practical message is a warning rather than a recommendation. Because N:M more than doubles across the normal range in healthy animals, any threshold built on it must be judged against that gradient. A doe with an N:M of 4.0 is not therefore inflamed; healthy does here reached that value routinely. The two crosses differed in their baselines SB does were consistently richer in neutrophils and poorer in lymphocytes but showed no significant cross × SCC interaction for neutrophils, N:M or lymphocytes. The response to rising cell load therefore looks the same in both, with the crosses separated by starting point rather than slope. This agrees with reports that SCC does not differ between goat breeds once management is held constant (Kaskous et al., 2023) [9] . None of this means the trait is uninformative: divergent selection on milk SCC in goats improves udder health and milk quality (Rupp et al., 2019) [ 18] , so SCC clearly carries real signal despite its physiological noise. This comparison should not be read more strongly. Treating a significant result in one cross and a non-significant result in the other as a real difference between them is a known error: the difference between significant and not significant is not itself statistically significant (Gelman and Stern, 2006) [6] . The lymphocyte data show this clearly. Pairwise testing suggested a difference between the crosses, but the interaction term (P = 0.528) rejects it and the effect sizes contradict it, since the AB decline was actually the larger of the two. The eight-fold gap in lymphocyte standard errors between crosses most likely reflects differences between observers or reading sessions rather than biology. Three constraints limit these conclusions, and all three point to the same next experiment. No bacteriology was performed, so udder health rests on clinical examination alone. The cell fractions sum to 100, so the neutrophil rise and macrophage fall are not independent. And the does came from one herd, in early lactation, at parities 3-6, chosen to span a wide gradient. That design controls the largest non-infectious sources of goat SCC variation (Luengo et al., 2004; Smistad et al., 2021) [11] , but it yields no information on prevalence and limits how far the results can be extended. Pairing differential cytology with bacteriological culture would turn this baseline into a diagnostic threshold. Expressing the counts as absolute concentrations (SCC × percentage) would free the fractions from the sum-to-100 constraint and test whether macrophages fall in absolute number rather than only in share. Together, these two steps would show whether the N:M ratio can separate infection from normal variation in a species where the total somatic cell count clearly cannot.

Conclusions

In clinically healthy Alpine × Beetal and Saanen × Beetal crossbred does, the milk differential leukocyte profile changes steadily and predictably across the normal somatic cell count range. Neutrophils rise, macrophages fall in step, lymphocytes decline, and the neutrophil-to-macrophage ratio more than doubles. These does were neutrophildominant even at the lowest cell counts, which confirms that goat milk cytology follows its own pattern rather than the cow's. The two crosses start from different baselines but respond in the same way, and there is no evidence that either reacts more strongly than the other. No bacteriology was performed, and the somatic cell range studied lies within the range reported for uninfected goats. These results are therefore reference data on normal physiology for two Indian crossbred dairy goat populations that had none before not a diagnostic test for subclinical mastitis. That is exactly where their value lies. They show how much the goat milk leukocyte profile varies in health, and therefore how much variation any future diagnostic threshold must be able to see past. Setting that threshold will need differential cytology paired with bacteriological culture and absolute cell counts across a wider somatic cell range.

Declarations

Acknowledgement

Acknowledgements:The authors would like to thank the Director of the Institute, National Dairy Research Institute (NDRI), Karnal, Haryana, for providing all the facilities necessary to carry out this research work.

References

1. Alhussien MN, Dang AK. Milk somatic cells, factors influencing their release, future prospects, and practical utility in dairy animals-an overview. Vet World. 2018;11:562-577.
2. Bagnicka E, Winnicka A, Jozwik A, Rzewuska M, Strzalkowska N, Kosciuczuk E, et al. Relationship between somatic cell count and bacterial pathogens in goat milk. Small Rumin Res. 2011;100:72-77.
3. Boettcher PJ, Moroni P, Pisoni G, Gianola D. Application of a finite mixture model to somatic cell scores of Italian goats. J Dairy Sci. 2005;88:2209-2216.
4. Contreras A, Sierra D, Sanchez A, Corrales JC, Marco JC, Paape MJ, et al. Mastitis in small ruminants. Small Rumin Res. 2007;68:145-153.
5. Cuccuru C, Moroni P, Zecconi A, Casu S, Caria A, Contini A. Milk differential cell counts in relation to total counts in Sardinian ewes. Small Rumin Res. 1997;25:169-173.
6. Gelman A, Stern H. The difference between "significant" and "not significant" is not itself statistically significant. Am Stat. 2006;60:328-331.
7. Haenlein GFW. Relationship of somatic cell counts in goat milk to mastitis and productivity. Small Rumin Res. 2002;45:163-178.
8. Jimenez-Granado R, Sanchez-Rodriguez M, Arce C, Rodriguez-Estevez V. Factors affecting somatic cell count in dairy goats: a review. Span J Agric Res. 2014;12:133-150.
9. Kaskous S, Farschtschi S, Pfaffl MW. Physiological aspects of milk somatic cell count in small ruminantsa review. Dairy. 2023;4:26-42.
10. Koop G, van Werven T, Toft N, Nielen M. Estimating test characteristics of somatic cell count to detect Staphylococcus aureus-infected dairy goats using latent class analysis. J Dairy Sci. 2011;94:2902-2911.
11. Luengo C, Sanchez A, Corrales JC, Fernandez C, Contreras A. Influence of intramammary infection and non-infection factors on somatic cell counts in dairy goats. J Dairy Res. 2004;71:169-174.
12. Min BR, Tomita G, Hart SP. Effect of subclinical intramammary infection on somatic cell counts and chemical composition of goats' milk. J Dairy Res. 2007;74:204-210.
13. National Research Council. Nutrient Requirements of Dairy Cattle. 7th rev ed. Washington, DC: National Academies Press; 2001.
14. Paape MJ, Capuco AV. Cellular defense mechanisms in the udder and lactation of goats. J Anim Sci. 1997;75:556-565.
15. Paape MJ, Poutrel B, Contreras A, Marco JC, Capuco AV. Milk somatic cell count and lactation in small ruminants. J Dairy Sci. 2001;84:E237-E244.
16. Persson Y, Olofsson I. Direct and indirect measurement of somatic cell count as indicator of intramammary infection in dairy goats. Acta Vet Scand. 2011;53:15.
17. Podhorecka K, Borkova M, Sulc M, Seydlova R, Dragounova H, Svejcarova M, et al. Somatic cell count in goat milk: an indirect quality indicator. Foods. 2021;10:1046.
18. Rupp R, Huau C, Caillat H, Fassier T, Bouvier F, Pampouille E, et al. Divergent selection on milk somatic cell count in goats improves udder health and milk quality with no effect on nematode resistance. J Dairy Sci. 2019;102:5242-5253.
19. Sierra D, Sanchez A, Corrales JC, Contreras A. Differential cell counts in goats' milk. In: Milking and Milk Production of Dairy Sheep and Goats. Wageningen: Wageningen Pers; 1999. p. 178-180.
20. Silanikove N, Merin U, Leitner G. On effects of subclinical mastitis and stage of lactation on milk quality in goats. Small Rumin Res. 2014;122:76-82.
21. Smistad M, Sølverød L, Inglingstad RA, Østerås O. Distribution of somatic cell count and udder pathogens in Norwegian dairy goats. J Dairy Sci. 2021;104:11878-11888.
22. Souza FN, Blagitz MG, Penna C, Della Libera AMMP, Heinemann MB, Cerqueira MMOP. Somatic cell count in small ruminants: friend or foe? Small Rumin Res. 2012;107:65-75.