Drying off is a very peculiar and important period for dairy cows. In this period, milk production ceases, secreting epithelial tissue renovates and the mammary gland prepares for the next lactation. In the last part of the dry period, the so-called ‘transition’ period starts. This period is defined as the three weeks before and three weeks after parturition, and represents a time of physiological stress for the dairy cows (Goff & Horst, Reference Goff and Horst1997). During drying off and the transition period, cows are subject to major changes in endocrine status, metabolic stressors and altered immune function, that could entail an increased risk of disease (Jonsson et al. Reference Jonsson, Fortes, Piper, Vankan, de Cisneros and Wittek2013). Considerable research has been conducted into the changes in immune function that occur around the time of calving, mostly from the perspective that these changes might increase susceptibility to periparturient infectious diseases (Goff & Horst, Reference Goff and Horst1997; Karcher et al. Reference Karcher, Beitz and Stabel2008) such as rumen acidosis, milk fever, lameness, metritis, and mastitis (Drackley, Reference Drackley1999). Moreover, multiple stressors can reroute resources like amino acids and energy, causing immune system depression, especially if they are severe and prolonged (Drackley et al. Reference Drackley, Dann, Douglas, Guretzky, Litherland, Underwood and Loor2005). Together with the immune competence reduction, transition cows were also shown to display an overt inflammatory response related to pregnancy and lactation, even without signs of infection and/or other pathology (Bionaz et al. Reference Bionaz, Trevisi, Calamari, Librandi, Ferrari and Bertoni2007; Bertoni et al. Reference Bertoni, Trevisi, Han and Bionaz2008; Trevisi et al. Reference Trevisi, Zecconi, Bertoni and Piccinini2010, Reference Trevisi, Amadori, Cogrossi, Razzuoli and Bertoni2012). This can increase the metabolic stress through alteration of the host's immune defenses, thus causing a vicious circle. However, previous field studies of milk and blood immune and inflammatory response parameters during periparturient period are influenced by herd (Piccinini et al. Reference Piccinini, Binda, Belotti, Casirani and Zecconi2004, Reference Piccinini, Binda, Belotti, Casirani and Zecconi2005), suggesting that metabolic adaptation in the animal as well as management of the herd could have an influence on cow immune and inflammatory response. For all these reasons, identifying accurate markers and their patterns to define the immune and inflammatory status of the healthy cow during drying-off and transition period would be very useful to increase our knowledge on cow response during these important parts of its productive life. Therefore, the purpose of this study was to determine several humoral and cellular parameters at specific time points during the drying off and transition period in 15 healthy primiparous cows from three dairy herds in Lombardy, and identifying the effects of physiological status and herd on the immune and inflammatory response of these cows.
Materials and methods
Herd and cow selection
Herds were selected based on their proximity to Milan to assure the delivery of samples in the shortest time. All herds kept cows indoors and applied total mixed rations based on corn silage all year-round. In these herds, all cows entering drying off period received local antibiotic treatment with a dry-cow product. The inclusion criteria for cows in their first lactation were to be pregnant, with parturition expected in 3 months-time, to be of the same breed (Italian Holstein-Friesian), and to be clinically healthy, without receiving any treatments or vaccines after breeding. In each herd, five cows were selected at random among the eligible ones.
Sampling
Cows were sampled from 7 d before drying off up to 4 weeks post-calving. A weekly sampling frequency was adopted with the exception of the period between 7d before expected date of calving and 7d after calving, when samples were taken every 3 d. For each cow and sampling time, three blood samples were collected from the jugular vein and put in either in EDTA or in sterile tubes to analyse respectively granulocyte population, enzymes, and cytokines; in this latter case, after addition of 10 parts of RNAlater solution (Life Technologies, USA) for total RNA isolation.
Cellular markers
White blood cells count (WBC), polymorphonuclear neutrophils (PMN), lymphocytes (LYM) and monocytes (MON) counts and proportion were assessed by an emocytometer (Sysmex XT-2000Iv, J). Proportion of CD4+, CD8+, and WC1+γδ T cell (WC1) were assessed by flow cytometric analysis following protocols routinely applied in our laboratories (Comazzi et al. Reference Comazzi, Gelain, Martini, Riondato, Miniscalco, Marconato, Stefanello and Mortarino2011). Cells were labelled with these antibodies: ILA11 (VMRD, USA), BAQ111A (VRMD, USA) and CC15 (Serotec, UK) for detection respectively of CD4+, CD8+ and WC1 subsets. Apoptosis (APO) and necrosis (NEC) were assessed by flow cytometric analysis, adding APC-conjugated Annexin V in conjunction with propidium iodide (PI) for identification of early and late apoptotic cells. All these assessments were performed using a FACScan flow cytometer and Cell Quest software (Becton Dickinson, USA).
Cytokine analysis
Total RNA was extracted from lysed blood cells using the RNeasy mini kit (RNeasy® mini kit; Qiagen, Germany). The cDNA was synthesised in 20 µl RT mix containing 50 U MuLV reverse transcriptase, 5 mm MgCl2, 10 mm Tris-HCl (pH 8·3), 50 mm KCl, 1·25 µm random hexadeoxyribonucleotide (pd (N)6) primers, 0·5 U/μl RNase inhibitor (GeneAmp® RNA PCR kit, ThermoFisher, USA), 1 mm dNTPs (Promega, USA). The mixture was subjected to 42 °C for 60 min and inactivated at 95 °C for 5 min. The final volume was adjusted to 100 µl with RNase free water. The cDNA was stored at −80 °C until used. Real-time RT-PCR was performed using QuantiTect SYBR Green RT-PCR (Qiagen, D) by a standard protocol recommended by the manufacturer in a iCycler iQ® (BioRad, USA). The primers used for interleukin 6 (IL6), interleukin 8 (IL8) and tumour necrosis factor-α (TNFα), as well as the reference gene (glyceraldehyde 3-phosphate dehydrogenase (GAPDH) derived from the paper of Leutenegger et al. (Reference Leutenegger, Alluwaimi, Smith, Perani and Cullor2000). Whereas interleukin 1 beta (IL1) primers derived from Prgomet et al. (Reference Prgomet, Sarikaya, Bruckmaier and Pfaffl2005) and interferon-γ (IFN) primers from Waldvogel et al. (Reference Waldvogel, Hediger-Weithaler, Eicher, Zakher, Zarlenga, Gasbarre and Heussler2000). The cytokines transcriptional level was calculated according to the comparative threshold cycle (CT) method (Leutenegger et al. Reference Leutenegger, Alluwaimi, Smith, Perani and Cullor2000).
Cellular ratios
Three commonly used cellular ratios were also calculated: (a) CD4+/CD8+ defined as the ratio of the proportion of CD4+ to the proportion of CD8+ as a marker of immune homeostasis (Mehrzad & Zhao, Reference Mehrzad and Zhao2008); (b) PMN/LIN, and (c) PMN/MON,defined as the ratio of proportion of PMN to, respectively, the proportion of LIN (P/L) or MON (P/M). These latter two ratios are considered as markers of inflammation (Rivas et al. Reference Rivas, Quimby, Blue and Coksaygan2001).
Enzyme analysis
Lysozyme (LYS) in blood serum was assessed in duplicate by the procedure described by Piccinini et al. (Reference Piccinini, Binda, Belotti, Casirani and Zecconi2004). The concentration of unknown samples, in μg/ml, was calculated by a standard curve obtained by adding a standard amount of lysozyme in each plate on a microplate spectrophotometer (Spectramax 340, Molecular Devices, USA). N-Acetyl-β-glucosaminidase (NAG) was assessed in duplicate by the procedure described by (Kitchen et al. Reference Kitchen, Middleton and Salmon1978), and expressed as units (pmol of 4-methylubelliferon released per min at 25 °C catalysed by 1 µl of blood serum) on a Fluoroskan Ascent microplate fluorimeter (Thermo Labsystems, Finland) at 355 exc and 460 em.
Statistical analysis
All data were collected in a database and statistically analysed by mixed model procedure for repeated measurements (MIXED) on SAS software (SAS 9·4, Sas Institute, USA) (Hatcher & Stepanski, Reference Hatcher and Stepanski1994). The between-subjects factor was represented by herds (3 levels), the within-subjects factor was represented by sampling time (12 levels) and the model applied was a full factorial, with polynomial contrasts for within-subjects factor. Cluster analysis was performed on SPSS 2·4 (IBM Corp, USA), while difference in cluster frequencies and markers’ mean values were analysed respectively with FREQ and GLM procedures of on SAS software (SAS 9·4, Sas Institute, USA).
Results
The pattern of 9 different cellular markers (APT, CD4+, CD8+, PMN, LYM, MON, NEC, WBC, WC1), five cytokines (IFN-γ, IL-1β, IL-6, IL-8, and TNF-α), two enzymes (LYS and NAG), and three cell ratios (CD4+/CD8+, PMN/LIN, PMN/MON) were assessed during drying off and transition periods in 15 primiparous cows (online Supplementary File Tables S1–S4). Our data showed that the variation of almost all parameters was influenced by the physiological period in which the samples were collected, except for APO, IL1, IL6, LYS and P/M (Table 1). Several markers were directly correlated to the herd, either alone like IL1, IL6, LYS, P/L and P/M (Table 1), or in association with the sampling time like WBC, NEC, CD4+, LYM, while NAG variability was affected by both the factors and their interaction (Table 1). Finally, three markers, PMN, MON and P/L varied depending on both time and herd, but not on the interaction between the two factors. Herd, sampling time or their interaction (Table 1), did not influence APO variation.
Cellular and humoral markers. Results of variance analysis of factors considered and their interaction by general linear model for repeated measurements
† n.s.: not significant at α = 0·05.
‡ RE: relative expression.
Among the parameters in which variation was significantly influenced by the interaction between herd and time, WBC showed large differences among herds (Fig. 1). Values observed in herd A were higher than the ones observed in the other two herds until calving was approaching. After calving, values dropped, and increased slightly 14 d after calving. On the other hand, values in herd B slightly increased close to calving to drop 7 d after calving.
Pattern of variation during the periparturient period for white blood cells (a), lymphocytes (b), CD4+ lymphocytes (c) and N-Acetyl-β-glucosaminidase (d) in the three herds considered (herd A •, herd B ○, herd C ▼). Vertical line at day 7 represents the start of drying off period, while vertical line at day 56 represents calving.
In addition, the pattern of LYM showed large differences among herds (Fig. 1). Indeed, herd A showed the lowest values until calving when a peak was observed. Values for herd B showed an opposite trend with higher values during drying-off, and a large drop around calving. Finally, herd C values, as for WBC pattern, had a lesser degree of variation, when compare to the other two herds.
Within the lymphocyte population, the CD4+ proportion showed a general increase during the observed period (Fig. 1). Values for herd B were lower than the other two herds until calving. Then, after calving, an increase was observed in all the three herds, but at different sampling times.
When NAG pattern was considered (Figure 1), large differences among herds were already observed during dry-off period. Furthermore, herd A showed two peaks in NAG levels, one seven days after drying off and one other, higher, 3 d after calving. After calving, the levels of NAG remained high in herd A, low in herd C, while values from herd B were always between the ones observed for the other two herds.
Data were evaluated by hierarchical cluster analysis to identify relative homogeneous groups based on the values of the sixteen parameters observed in each sampling time. This analysis allowed identifying 3 clusters, including, respectively, 19·1, 33·5 and 47·4% of the samples. These three clusters had different profiles along the follow-up period (Figure 2). Moreover, there is a significant association (P < 0·05 at χ2 test) between clusters and herds. Indeed, cluster 1 was significantly associated with herd C (62·2% of the samples); cluster 2 with herd A (66·7% of the samples), while cluster 3 was numerically, but not statistically, associated with herd A (50·6% of the samples).
Frequency of sample clusters during the periparturient period in the three herds considered (herd A •, herd B ○, herd C ▼). Vertical line at day 7 represents the start of drying off period, while vertical line at day 56 represents calving.
Table 2 reports the least-square mean values (±s.e.) observed in the three clusters, calculated by GLM procedure. A significant difference among clusters was observed for all the cellular markers considered out of WC1, NEC, APO and P/M ratio. When humoral markers were considered, only two of them showed a significant difference among clusters (IL1 and IL8) (Table 3).
Cellular markers. Least square means and standard errors observed in the three cluster defined by hierchical cluster analysis
† Within a row, values with different superscript statistically differ (α = 0·05).
‡ n.s., not significant.
Humoral markers. Least square means and standard errors observed in the three cluster defined by hierchical cluster analysis
† RE, Relative expression.
‡ Within a row, values with different superscript statistically differ (α = 0·05).
§ n.s., not significant.
Discussion
Despite clear evidence of changes in immune and inflammatory markers during drying-off and periparturient period in absence of any apparent disease status, the attempts to identify markers or indexes to define cows at risk during this period in a consistent way are still unsatisfactory (Trevisi et al. Reference Trevisi, Amadori, Cogrossi, Razzuoli and Bertoni2012; Hailemariam et al. Reference Hailemariam, Mandal, Saleem, Dunn, Wishart and Ametaj2014). Moreover, a range of normality for several of the parameters considered was not set or, when such ranges are available, they are not relevant to this peculiar period of cow life. Therefore, with the aim of contributing to filling this gap, we screened several parameters related to innate immunity and inflammation to identify their pattern during drying off and periparturient period, and to assess the effects of physiological status and herd on parameters’ values. Despite the relatively small number of individuals, and the expected biological variations, the statistical analysis showed the presence of a significant influence of these periods on several markers.
During drying-off and transition period significant variations among immune and inflammatory markers are expected (Goff & Horst, Reference Goff and Horst1997; Trevisi et al. Reference Trevisi, Zecconi, Bertoni and Piccinini2010) even in healthy animal, as in this study. Six markers (WC1, CD8+, INF, IL8 and TNF CD4+/CD8+), were influenced only by time. The absence of significant effects due to herd and its interaction with time suggests that these parameters could be useful to assess changes in inflammation or immune competence only in relation to physiological status, in the absence of an overt disease status. More interestingly, there are other four markers (IL1, IL6, LYS and P/L), that varied only depending by the herd. Therefore, they could be good candidate markers of the general status of the herd, as previously suggested (Amadori et al. Reference Amadori, Fusi, Bilato, Archetti, Lorenzi and Bertocchi2015).
The observed changes of PMN and MON were depending on time and herd, but not on their interaction, thus, for each specific point of time, it is not possible to know whether to attribute any differences to the physiological status or to the herd. The variability of remaining markers (WBC, LYM, CD4+, NAG, NEC) was influenced by the interaction between time and herd. This latter evidence, despite the small sample size, further supports the role of these factors as potential marker candidates to assess immunological and inflammatory status in cows in apparent health status during drying-off and periparturient period.
These observations are supported by the results of hierarchical cluster analysis, which allowed us to identify three clusters of cellular parameters with significant differences in their mean values, their frequencies during the follow-up period (Fig. 2), and with a significant association with the three herds considered.
Cluster 1 showed significantly low values for PMN, PMN/LIN, PMN/MON and WBC, while slightly higher mean values for CD4+/CD8+ and IL8 compared to the other clusters. This pattern suggests the absence of an inflammatory reaction (Rivas et al. Reference Rivas, Quimby, Blue and Coksaygan2001). Moreover, the low variation in frequencies observed during the follow-up period, suggests an association to cows with a sufficient immunologic homeostasis (Mehrzad & Zhao, Reference Mehrzad and Zhao2008).
Cluster 2 is characterised by high level of PMN, P/L, IL1 and low level of CD8+ and LYM, suggesting that samples derived from cows with some impairments of immune response and/or a subclinical inflammatory reaction (Rivas et al. Reference Rivas, Quimby, Blue and Coksaygan2001; Mehrzad & Zhao, Reference Mehrzad and Zhao2008).
Cluster 3 showed the highest frequency during drying-off and its characterised by significant high levels of LYM, WBC, and significant low levels for CD4+/CD8+, IL-1 and IL8. This pattern suggests that the cluster includes samples from cows with a good homeostasis (Danicke et al. Reference Danicke, Meyer, Winkler, Ulrich, Frahm, Kersten, Valenta, Rehage, Haussler, Sauerwein and Locher2016).
Conclusions
Drying off and periparturient period was confirmed to have a large and significant influence on the variability of several, but not all, immune and inflammatory parameters. The results of this study under field conditions, despite the relatively small sample size, demonstrated several markers of innate immunity and inflammation that varied during drying-off and transition period due to the effects of herd, as observed previously (Piccinini et al. Reference Piccinini, Binda, Belotti, Casirani and Zecconi2005; Reference Piccinini, Binda, Belotti, Casirani and Zecconi2004).
Among the sixteen markers and the three ratios considered, most of the parameters related to cellular immunity were significantly influenced by herd and its interaction with physiological status of the cow. These results support previous evidence on the role of herd factor on immune/inflammatory response, despite the apparent similarity among herds.
The presence of a significant effect of the interaction between herd and physiological status for most of the cellular parameters suggests these as suitable markers to define cow status during drying-off and peripartum period, when the aim is to assess differences among herds, or when data from a ‘reference’ herd are available.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022029918000316.
