Metabolic and inflammatory responses reveal different adaptation to the transition period challenges in Holstein, Brown Swiss, and Simmental dairy cows

Abstract

Twelve Holstein (HOL), 9 Brown Swiss (BRO), and 9 Simmental (SIM) dairy cows raised in the same barn and managed identically were enrolled to characterise the responses to the transition period. From −21 to 28 d from calving (DFC), body condition score (BCS) and milk yield were measured, and blood samples were collected to assess plasma biomarkers of metabolism, minerals, liver function, inflammation, and oxidative status. Compared with SIM, HOL and BRO had higher milk yield during the first week of lactation. HOL had the highest milk yield from the second to the fourth week and the lowest BCS at 28 DFC. SIM had the highest BCS and the highest plasma creatinine and P, reflecting a greater attitude to gain muscle mass. Compared with SIM, BRO and HOL had lower plasma fructosamine, indicating reduced glucose availability driven by lactose synthesis. SIM had the mildest acute phase response at the onset of lactation, as suggested by the lowest ceruloplasmin concentration. HOL had the highest cholesterol and paraoxonase concentrations, and the greatest interleukin-1β production by leukocytes following ex-vivo stimulation, suggesting that they had the quickest resolution of the acute phase response due to the fastest immune cell activation. BRO had the highest concentration of reactive oxygen metabolites, ceruloplasmin, bilirubin, glutamate-oxaloacetate transaminase, γ-glutamyl transferase, and β-hydroxybutyrate, and the lowest concentration of paraoxonase, reflecting the strongest activation of leukocytes and the most severe acute phase response. Together, these results highlight different metabolic and inflammatory conditions around calving in the three explored breeds.

HIGHLIGHTS

• Holstein, Brown Swiss, and Simmental are dairy breeds selected for different purposes.

• These three breeds have different productivity, and metabolic and inflammatory responses during the transition period.

• Understanding the differences among breeds might improve their management, nutrition, and productivity in commercial dairy farms.

Keywords:

• Acute phase response
• liver function
• metabolism
• milk yield
• peripartum

Introduction

The physiological adaptation of dairy cows to the transition period challenges has become a widely debated topic over the last decades (Grummer 1995; Drackley 1999; LeBlanc 2010). In this context, the pivotal role played by the immune system in affecting the success of such adaptation has been recently recognised (Bertoni and Trevisi 2013; Mezzetti et al. 2020; Cattaneo et al. 2021). Acute phase response has been documented as a physiological condition in dairy cows during the transition period (Van Knegsel et al. 2014; Premi et al. 2021), although the magnitude and duration of this process vary markedly among individuals, deeply affecting performances at the onset of lactation. During the acute phase, the liver synthesises positive acute phase proteins (APP; i.e. haptoglobin and ceruloplasmin) (Ceciliani et al. 2012), decreasing the production of other proteins referred to as negative APP (i.e. albumin, lipoproteins, retinol binding protein and paraoxonase) (Bionaz et al. 2007; Bertoni et al. 2008).

Most of the research exploring this topic focuses on Holstein Friesian as a model for the modern high-yielding cow. Despite that, dairy breeders all over the world manage various breeds selected to maximise different production traits (Back et al. 2006; De Marchi et al. 2008; Gustavsson et al. 2014; Bland et al. 2015) or adapted to face different environmental conditions (White et al. 2002; Piccand et al. 2013; Curone et al. 2018). Besides Holstein (HOL), Brown Swiss (BRO) and Simmental (SIM) are among the most widespread cosmopolitan dairy breeds, raised particularly for their milk composition traits and muscle gain, respectively. Literature documents large variability among these breeds in their adaptive response to environmental stressors (Mylostyvyi et al. 2021), reproductive performances (Piccand et al. 2013), feed efficiency (Dickinson et al. 1969), metabolism and milk composition (Benedet et al. 2020; Franzoi et al. 2020). Conversely, little is known about the immune system of the different breeds (Curone et al. 2019), and existing results were mostly obtained through ex-vivo approaches. According to Lacetera et al. (2006), BRO peripheral blood mononuclear cells are less tolerant to chronic heat exposure than those from HOL, suggesting a different decline in immune functions between these two breeds when exposed to high temperatures. Compared with HOL, BRO cows have greater resistance to bacterial infections, and their macrophages produce more reactive oxygen species and less interleukin (IL)-1β following stimulation with bacterial lipopolysaccharides (LPS) (Gibson et al. 2016).

To the best of our knowledge, studies comparing the adaptive metabolic responses of different dairy breeds at calving are lacking. We hypothesised that dairy breeds could have different haematochemical profile during the transition period, likely driven by the different breeding objectives and milk yield among them. Thus, this observational study aimed at exploring the metabolic and inflammatory adaptations during the transition period in HOL, BRO, and SIM dairy cows managed identically.

Materials and methods

The study was carried out in a commercial dairy farm located in the Parmigiano Reggiano production area (Terenzo, Parma, Italy), from November 2018 to May 2019, in accordance with Italian laws on animal experimentation (DL n. 26, 04/03/2014) and ethics (authorization of the Italian Ministry of Health n. 851/2018).

Experimental design and animal management

The herd included HOL, BRO, and SIM dairy cows raised together, under the same management practices. Cows were dried off 60 days before the expected calving and housed in a bedded-pack pen until calving. Thereafter, cows were moved to a postpartum bedded-pack pen for two weeks and finally moved to a lactation freestall pen, with cubicles bedded with straw. Lactating cows were milked twice daily at 4.00 am and pm.

All the cows were fed the same TMR diet formulated to satisfy the average herd requirements, according to the ‘Parmigiano Reggiano’ production regulation, and distributed twice daily for lactating cows (at 6.00 am and pm) and twice weekly for dry cows. A 3–5% refusal was guaranteed to ensure that cows had ad libitum access to feed. Representative samples of TMR were collected monthly. Samples were ground and characterised by a Foss NIR systems 5000 spectrophotometer (Hillerød, Denmark) equipped with a monochromator and transport module, scanning over the wavelength range 1100–2500 nm every 2 nm. The calibrations used to obtain the forage characterisations were produced by a commercial forage testing laboratory (CRPA Lab, Reggio Emilia, Italy), and samples were analysed for crude protein, NDF, ether extract, starch, and sugar. Analysis results were used to calculate the nutritional value of the feed, in accordance with NRC (2001) guidelines. Diet composition is reported in Table 1.

Table 1. Diets offered to the cows during the study.

Item^a	Lactating cows	Dry cows
Diet, % DM
Grass hay	25.10	50.00
Alfalfa hay	25.00	‒
Straw	‒	20.70
Corn grain ground dry	19.10	‒
Corn grain flaked	5.12	3.51
Barley flaked	4.16	‒
Soybean meal	‒	4.97
Soybean flaked	3.50	‒
Beet molasses	2.73	2.94
Flaxseed	1.81	‒
Flaxseed meal	3.54	3.64
Beet sugar pulp dry	1.39	‒
Soy hulls	1.41	10.80
Supplement^b	1.88	3.00
Chemical Composition
NEL, Mcal kg of DM^−1	1.59	1.30
Crude protein, % DM	15.10	12.30
Starch + sugar, % DM	27.80	17.90
Ether extract, % DM	3.28	2.50
NDF, % DM	34.80	49.60
MP, % CP	10.70	8.50
RUP, % DM	5.50	3.80

^aDM is dry matter; NE_L is net energy for lactation; NDF is neutral detergent fibre; CP is crude protein; MP is metabolisable protein and RUP is rumen undegradable protein (both estimated using NRC 2001).

^bLactating cows’ supplement contained 15.5% Ca, 2.8% P, 9% Na, 7.5% Mg, 400,000 IU/kg of vitamin A, 100,000 IU/kg of vitamin D, 2.000 mg/kg of vitamin E, 50 mg/kg of vitamin K, 110 mg/kg of vitamin B1, 1.2 mg/kg of vitamin B12, 4.200 mg/kg of vitamin PP, 5.200 mg/kg of choline, 3.350 mg/kg of Mn, 7.500 mg/kg of Zn, 2.400 mg/kg of Cu, 400 mg/kg of I, 30 mg/kg Se.

Dry cows’ supplement contained 3.5% Ca, 4.0% P, 1,5% Na, 5.0% Mg, 200,000 IU/kg of vitamin A, 50,000 IU/kg of vitamin D, 2.800 mg/kg of vitamin E, 100 mg/kg of vitamin K, 350 mg/kg of vitamin B1, 3 mg/kg of vitamin B12, 3.500 mg/kg of vitamin PP, 4.500 mg/kg of choline, 1.000 mg/kg of Mn, 2.800 mg/kg of Zn, 150 mg/kg of Cu, 100 mg/kg of I, 12 mg/kg Se.

A group of 30 parous dairy cows was enrolled in the experiment after the dry-off as follows: 12 HOL (parity: 2.7 ± 0.8, BCS: 3.25 ± 0.30, average milk production previous lactation: 11,000 ± 585 kg [mean ± SD]); 9 BRO (parity: 3.0 ± 0.5, BCS: 3.25 ± 0.31, average milk production previous lactation: 9,484 ± 354 kg); and 9 SIM (parity: 3.5 ± 2.1, BCS: 4.0 ± 0.38 average milk production previous lactation: 9,858 ± 405 kg). Subjects in each group were chosen to represent the proportion of each breed within the herd, according also to production relative to herd and breed average. Periodical checks were performed between −21 and 28 d from calving (DFC), as described in the following sections.

Body condition score and milk yield

The BCS was determined by the same operator using a 1–5 scale (Edmonson et al. 1989) at −21 ± 2, 3, 7, and 28 DFC. The ΔBCS was calculated as the difference between BCS values measured at 3 and 28 DFC. The milk yield was automatically measured in the milking parlour using the Afikim system (SAE Afikim, Kibbutz Afikim, Israel) at each milking between 1 and 30 DFC. Data about milk yield were expressed as average weekly values.

Health status

Health status was monitored daily from −21 to 30 DFC. Hoof health was assessed daily by the breeder. Mastitis was diagnosed by visual evaluation of abnormal milk from each quarter. The retained placenta was recorded when foetal membranes were not expelled within 24 h after calving. Endometritis and metritis were diagnosed by a veterinary practitioner through rectal ultrasonography according to Sheldon et al. (2006). Milk fever, displacement of the abomasum, and ketosis were diagnosed by veterinarian examination.

Blood sample collection and analysis

At −21 ± 2, 3, 7, and 28 DFC, before the morning feeding, blood was collected from the jugular vein into 10-mL evacuated heparinised tubes (BD Vacutainer, BD Diagnostics, Franklin Lakes, New Jersey, United States). After collection, blood was processed and analysed as described by Calamari et al. (2016). Briefly, blood was centrifuged (3500 × g, 15 min at 4 °C), and the packed cell volume (PCV) was directly measured after centrifugation through a capillary column (ALC Centrifugette 4203). A clinical autoanalyzer (ILAB-650, Werfen – Instrumentation Laboratory, Bedford, MA, United States) was used to measure the concentration of glucose, nonesterified fatty acids (NEFA), β-hydroxybutyrate (BHB), urea, creatinine, Ca, P, Mg, Na, K, Cl, Zn, haptoglobin, ceruloplasmin, total protein, albumin, globulin, cholesterol, total bilirubin, aspartate aminotransferase (AST-GOT), γ-glutamyl transferase (GGT), and alkaline phosphatase according to Calamari et al. (2016). Furthermore, the same instrument was used to determine reactive oxygen metabolites (ROMt) and ferric ion reducing antioxidant power (FRAP) according to Jacometo et al. (2015), paraoxonase according to Bionaz et al. (2007), thiol groups according to Minuti et al. (2014), myeloperoxidase according to Bradley et al. (1982), fructosamine according to Caré et al. (2018). Moreover, IL-1β and IL-6 according to Mezzetti et al. (2019). Further details on the analytical procedures adopted for blood analyses are reported in Table S1.

Whole blood stimulation assay. At 3 DFC, an additional sample was collected in heparinised tubes under aseptic conditions and processed according to Jahan et al. (2015), with some modifications. Briefly, a 5 µg/mL solution of Escherichia coli LPS (O111:B4, Sigma Aldrich Company Ltd., UK, Cat. No. L3012) and Dulbecco’s modified Eagle medium (Sigma Aldrich Company Ltd., UK, Cat. No. D6046) was injected through the rubber cap of the tube using a 1-mL syringe. After injection, the tubes were gently inverted 10 times and placed in a water bath at 38 °C for 3.5 h. After the whole blood stimulation assay (WBA), the plasma was collected by centrifugation at 3500 × g for 16 min at 4 °C and stored at −80 °C for IL-1β and IL-6 measurements. After the WBA, variations in the plasma cytokines were expressed as the fold change relative to the baseline.

Statistical analysis

Data were analysed using SAS software, version 9.4 (SAS Inst. Inc., Cary, NC, USA) and are presented in graphs and tables as the least squares mean and pooled standard error for individual means of breeds over time. Data underwent ANOVA testing using a mixed model for repeated measures (Glimmix Procedure, SAS Inst. Inc.). The statistical models included the fixed effect of breed (Br), time (DFC for BCS and plasma analytes, weeks from calving for milk yield), their interaction (Br × Time), and the random effect of the individual cow. For WBA, the model included only Br as a fixed effect. Parity was firstly included in the models as a covariate but was not retained because of the lack of significance for the great majority of the variables investigated. The analysis was carried out using two covariance structures: autoregressive order and spatial power, with their heterogeneous counterparts. The covariance structures were ranked according to their Akaike information criterion, with the one having the lowest criterion being chosen (Littell et al. 1998). The distribution of residuals was visually assessed. The pairwise comparison was done using the least significant difference test with the Tukey adjustment for multiple comparisons. Significance was declared p ≤ .05, and differences for p ≤ .1 were discussed in the context of tendencies.

Results

Body condition score, milk yield, and health status

The BCS was higher in SIM compared with the other breeds (3.56 vs 2.95 and 3.08 ± 0.10 for SIM, HOL, and BRO, respectively; p < .05 for both HOL and BRO), and a Br x Time interaction was also observed (p = .04, Figure 1(a)). Compared with the other breeds, SIM had the highest BCS at −21, 3 (3.61 vs 3.00 vs 2.99 ± 0.12 for SIM, HOL, and BRO, respectively), and 7 DFC (p < .01 for both HOL and BRO), and HOL had the lowest BCS at 28 DFC (3.22 vs 2.71 vs 3.02 ± 0.12 for SIM, HOL, and BRO, respectively; p < .05 and p < .01 for BRO and SIM, respectively). After calving, BCS remained unchanged in BRO, thus the ΔBCS was lower in BRO compared with the other breeds (0.00 vs. 0.29 and 0.39 ± 0.09 for BRO, HOL, and SIM, respectively; p < .05 for both HOL and SIM). Milk yield tended to be higher in HOL than in the other breeds (37.7 vs 32.6 and 31.9 ± 1.85 L/d in HOL, BRO, and SIM, respectively; p < .1 for both BRO and SIM), and a Br x Time interaction was also observed (p < .01, Figure 1(b)). Compared with other breeds, SIM had the lowest milk yield during the first week of lactation (p < .05 for both HOL and BRO), whereas HOL had the highest milk yield during the second, third, and fourth weeks of lactation (p < .05 for both BRO and SIM). Disease incidence was similar among breeds (Table S2).

Figure 1. Values of body condition score (BCS; a) and milk yield (b) from −21 to 28 d from calving (DFC) in Holstein (HOL), Brown Swiss (BRO), and Simmental (SIM) dairy cows managed identically. Br is the breed effect, Time is the time effect, and Br × Time is the interaction effect (superscript letters denote differences for p < .05 at each time point in pairwise comparisons). Time points consider single days for BCS and weekly averages for milk yield.

Table 2. Plasma concentrations from −21 to 28 d from calving of biomarkers of metabolism, minerals, liver function, inflammation, and oxidative status in Holstein (HOL), Brown Swiss (BRO), and Simmental (SIM) dairy cows managed identically.

	Breed^b				p-value^d
Item^a, unit	HOL	BRO	SIM	SEM^c	Br	Time	Br × Time
PCV, L/L	0.34	0.33	0.36	0.01	.22	<.01	.38
Metabolism
Glucose, mmol/L	4.14	4.05	4.21	0.08	.39	<.01	.16
Fructosamine, µmol/L	295^b	299^a^,^b	311^a	4.30	.03	<.01	.32
NEFA, mmol/L	0.45	0.51	0.53	0.06	.60	<.01	.88
BHB, mmol/L	0.44^b	0.56^a	0.49^a^,^b	0.03	.04	<.01	.14
Urea, mmol/L	4.06	4.38	4.33	0.33	.72	<.01	.18
Creatinine, µmol/L	95.50^b	99.40^a^,^b	108.30^a	2.31	<.010	<.01	.87
Minerals
Ca, mmol/L	2.49	2.49	2.47	0.03	.88	<.01	.68
P, mmol/L	1.73^b	1.61^b	1.92^a	0.07	.01	<.01	.02
Mg, mmol/L	0.96	0.92	0.97	0.02	.25	<.01	.18
Na, mmol/L	146	147	146	0.73	.72	<.01	.68
K, mmol/L	4.31	4.21	4.26	0.07	.56	.34	.85
Cl, mmol/L	106	107	106	0.71	.36	<.01	.81
Zn, µmol/L	11.40	12.10	11.30	0.49	.44	<.01	.22
Liver function
Bilirubin, µmol/L	3.38^b	5.28^a	3.95^a^,^b	0.56	.04	<.01	<.01
AST-GOT, U/L	101.60^a	150.80^b	90.30^a	14.60	.01	<.01	.09
GGT, U/L	22.00^a	37.20^b	19.60^a	5.12	.04	.39	.70
Alkaline phosphatase, U/L	59.40	47.90	56.20	8.42	.58	.02	.11
Inflammation
Total protein, g/L	79.70^a	76.30^b	82.50^a	1.16	<.01	<.01	.88
Albumin, g/L	35.00	34.70	36.20	0.61	.20	.01	.38
Globulin, g/L	44.60	41.60	46.20	1.49	.10	<.01	.91
Haptoglobin, g/L	0.40	0.44	0.39	0.05	.76	<.01	.30
Ceruloplasmin, µmol/L	2.65^a	3.33^b	2.37^c	0.10	<.01	<.01	<.01
IL-1β, pg/mL	309	1212	247	488	.30	.06	.23
IL-6, pg/mL	592	1094	458	242	.16	.02	.22
Myeloperoxidase, U/L	431	478	448	19.40	.21	.02	.41
Cholesterol, mmol/L	3.91^a	3.40^a^,^b	3.20^b	0.22	.04	<.01	.86
Paraoxonase, U/mL	84.20^a	67.40^b	73.20^a^,^b	5.20	.05	<.01	.06
Oxidative status
Thiol groups, µmol/L	313	318	3340	13.90	.51	.31	.72
FRAP, µmol/L	176	155	1530	20.90	.64	.29	.74
ROMt, mg H2O2/100 mL	20.70^b	23.90^a	20.10^b	1.06	.03	<.01	.02

^aPCV is packed cell volume, NEFA is nonesterified fatty acids, BHB is β-hydroxybutyrate, AST-GOT is glutamate-oxaloacetate transaminase, GGT is γ-glutamyl transferase, IL-1β is interleukin-1β, IL-6 is interleukin-6, FRAP is ferric ion reducing antioxidant power, ROMt is total reactive oxygen metabolites.

^bWithin a row, means without a common superscript differ for p < .05.

^cStandard error = greatest standard error of the means among breeds.

^dOverall effects of Breed (Br), Time, and their interaction (Br × Time).

Metabolic profile

Energy and protein metabolism biomarkers, kidney function indicators, and mineral metabolism biomarkers. Among energy metabolism biomarkers, fructosamine concentration was higher in SIM than in HOL and tended to be higher in SIM than in BRO (p < .05 and p = .1, respectively; Table 2). Plasma BHB concentration was higher in BRO than in HOL (p = .03, Table 2). No effect was detected for PCV, glucose, and NEFA (Figure S1(a–e)). Among the protein metabolism and kidney function biomarkers, creatinine concentration was higher in SIM than in HOL and tended to be higher in SIM than in BRO (p < .01 and p = .1 for HOL and BRO, respectively; Table 2). No effect was detected for urea (Figure S1(f)). Among the mineral metabolism biomarkers, P concentration was highest in SIM (p < .05 and p < .01 for HOL and BRO, respectively; Table 2), and a Br × Time interaction was also observed (p = .02, Figure 2(a)). Compared with the other breeds, SIM had the highest P concentration at 3 (p = .02 and p = .03 for BRO and HOL, respectively) and 7 DFC (p < .01 for both BRO and HOL). The P concentration was higher in HOL than in BRO at 28 DFC (p < .05). No effect was detected for the other minerals (Figure S2).

Figure 2. Trends from −21 to 28 d from calving (DFC) of plasma concentrations of P (a), bilirubin (b), glutamate-oxalacetate transaminase (AST-GOT) (c), ceruloplasmin (d), paraoxonase (e), and reactive oxygen metabolites (f) in Holstein (HOL), Brown Swiss (BRO), and Simmental (SIM) dairy cows managed identically. Br is the breed effect, Time is the time effect, and Br × Time is the interaction effect (superscript letters denote differences for p < .05 at each time point in pairwise comparisons).

Liver function and inflammation biomarkers. Among the liver function biomarkers, total bilirubin concentration was higher in BRO than in HOL cows (p = .04; Table 2), and a Br x Time interaction was also observed (p < .01; Figure 2(b)). Compared with the other breeds, BRO had the highest bilirubin concentration at 3 DFC (p < .01). The AST-GOT and GGT concentration was the highest in BRO (p = .01 for AST-GOT and p < .05 for GGT for both HOL and SIM, Table 2), and a tendency towards a Br × Time interaction was also observed for AST-GOT (p = .09; Figure 2(c)). Compared with the other breeds, BRO had the highest AST-GOT concentration at 3 (p < .01 for both HOL and SIM) and 7 DFC (p < .05 for both HOL and SIM). No effect was detected for the alkaline phosphatase concentration (Figure S3(b)). Among the inflammation biomarkers, total protein concentration was the lowest in BRO (p < .05 and p < .01 for HOL and SIM, respectively) and tended to be higher in SIM than in HOL (p < .1; Table 2). No effect was detected for myeloperoxidase, IL-1β, IL-6, and globulin (Figures S3(e) and S4(a,b)). Among the positive APP, ceruloplasmin concentration was the highest in BRO (p < .01 for both HOL and SIM) and was higher in HOL than in SIM (p = .04; Table 2). A Br x Time interaction was also observed (p < .01; Figure 2(d)). Compared with BRO, SIM had lower ceruloplasmin at −21 DFC (p = .04), and both SIM and HOL had lower ceruloplasmin at 3, 7, and 28 DFC (p ≤ .01). Compared with SIM, HOL had higher ceruloplasmin at 3 DFC (p = .01). No difference was observed for haptoglobin concentration (Figure S3(f)). Among the negative APPs, cholesterol concentration was higher in HOL than in SIM and tended to be higher in HOL than in BRO (p < .05 and p = .08, respectively; Table 2). Plasma paraoxonase concentration was higher in HOL than in BRO (p = .05; Table 2), and a tendency towards a Br x Time interaction was also observed (p = .06; Figure 2(e)). At 7 DFC, paraoxonase concentration was higher in HOL than in BRO (p < .01) and tended to be higher in HOL than in SIM (p = .09) and in SIM than in BRO (p = .07). No effect was noted for albumin concentration (Figure S2(d)).

Redox Balance Biomarkers. Among the antioxidant systems biomarkers, no effect was detected for thiol groups and FRAP concentrations (Figure S3(e,f)). Among the oxidant species biomarkers, ROMt concentration was the highest in BRO (p = .04 for both HOL and SIM; Table 2), and a Br x Time interaction was also observed (p = .02; Figure 2(f)). Compared with the other breeds, BRO had the highest ROMt concentration at 3 (p < .01 for both HOL and SIM) and 7 DFC (p < .05 and p < .01 for HOL and SIM, respectively).

Whole blood stimulation assay

After WBA, the fold change of IL-1β was the highest in HOL cows (p < .05 for both BRO and SIM; Table 3). No effect was detected in IL-6 fold changes.

Table 3. Fold changes of cytokines following a whole blood stimulation assay from blood collected 3 d after calving in Holstein (HOL), Brown Swiss (BRO), and Simmental (SIM) dairy cows managed identically.

	Breed^a				p-value
Item, unit	HOL	BRO	SIM	SEM^b
Interleukin-6, pg/mL	3.63	2.86	2.95	0.64	.55
Interleukin-1β, pg/mL	45.30^a	16.90^b	15.20^b	9.30	.01

^aWithin a row, means without a common superscript differ for p < .05.

^bSEM: greatest standard error of the means among breeds.

Discussion

The transition period is characterised by an overt systemic inflammatory response which occurs immediately after calving even without any sign of infections (Trevisi and Minuti 2018), as confirmed in the present study by the dramatic increase in the positive APP (i.e. haptoglobin and ceruloplasmin) paired with the decrease in the negative APP (i.e. cholesterol, albumin, and paraoxonase). Nevertheless, the magnitude of this response varied among the explored breeds, even though all cows were managed identically and fed the same diet. This unusual setting allowed us to compare different breeds, mitigating the environmental variability. Nevertheless, the facilities used to perform the present study did not enable us to record individual feed intake. Therefore, we were unable to fully distinguish between the effects of a different adaptation to the acute phase response from those that could have been driven by a different feeding behaviour (and likely by a different nutrient intake) among the three explored breeds. Another limitation is represented by the limited sample size. Thus, results of this observational study should be interpreted accordingly.

During the transition period, SIM cows had the lowest ceruloplasmin concentration, suggesting that this breed had the least severe acute phase response at the onset of lactation. Ceruloplasmin is a positive acute-phase protein, whose recovery is slower than that of haptoglobin (Bertoni et al. 2008). This was likely driven by the lower metabolic demand faced by those cows, as reflected by their higher BCS and plasma fructosamine concentration compared with the other breeds. Fructosamine is a stable glycated protein formed by an irreversible non-enzymatic reaction between glucose and proteins (mainly albumin) (Armbruster 1987). Fasting plasma glucose concentration is finely reflected by fructosamine within one to three weeks (Caré et al. 2018), and higher plasma fructosamine in early lactation suggested there was a lower amount of energy diverted to galactopoietic processes in SIM cows compared with BRO and, especially, HOL cows. This effect could have been driven by the least milk yield characterising SIM cows while they were fed the same diet as the other breeds. In fact, SIM cows have a greater attitude towards gaining muscle mass rather than prioritising milk synthesis during energy partitioning, as reflected by the higher creatinine concentration detected in their plasma throughout the study, which can indicate greater utilisation of phosphocreatine by exercising muscles (Hayden et al. 1992; Finco et al. 1997) compared with the other breeds. The lower milk yield in SIM compared with other breeds can also account for the difference in P concentration, as mammary gland uptake contributes to P losses, particularly at the onset of lactation (Goff 2000; Grünberg 2014). Previous research highlighted how the shift of metabolic functions towards homeorhetic regulation at calving is smoother in SIM compared with HOL cows (Lopreiato et al. 2019) and could also result in different leukocyte function, with SIM cows seeming to have a heightened immune response (Lopreiato et al. 2020).

As milk volume is mainly driven by the osmotic effect exerted by lactose (Sadovnikova et al. 2021), a higher glucose diversion to lactose synthesis accounted for the higher milk yield that BRO and HOL cows had during the first week of lactation compared to SIM cows. Furthermore, HOL cows had the highest milk yield from the second to the fourth weeks of lactation, suggesting the mammary demand for milk synthesis at the onset of lactation was the highest among the explored breeds. This is consistent with their lowest BCS at 28 DFC, suggesting that milk yield in HOL cows was supported a greater extent by mobilising body reserves compared with the BRO cows, as previously reported (Gruber et al. 2014). Genetic selection in HOL has prioritised milk yield, promoting a metabolic asset that favours more nutrient partitioning towards the mammary gland and body reserve mobilisation (De Koster and Opsomer 2013) compared with other less-selected breeds (Yan et al. 2006; Lucy et al. 2009; Friggens et al. 2013).

The highest concentrations of cholesterol and paraoxonase during the week after calving measured in HOL compared with other breeds suggest that HOL cows benefitted from the fastest recovery from the inflammatory condition. Cholesterol decrease can be considered an indicator of reduced lipoprotein synthesis by the liver, suggesting an acute phase response occurring (Bertoni and Trevisi 2013). Paraoxonase is a liver enzyme whose concentration is decreased by diseases and inflammatory states (Bionaz et al. 2007). In contrast, BRO cows had the highest concentration of ceruloplasmin throughout the experimental period and the lowest concentration of paraoxonase at 7 DFC. These results suggest that, in this study, BRO cows faced the most severe acute phase response. A possible explanation for the different adaptations to lactation could be due to the varying functionality of the leukocytes between these two breeds. This was supported by finding the greatest IL-1β production in leukocytes from HOL cows following ex-vivo stimulation, consistent with the results obtained by Gibson et al. (2016) in HOL-stimulated macrophages compared with those from BRO cows. These results suggested that, in HOL, immune cells rely on greater cytokine production in response to harmful stimuli, likely accounting for a faster resolution of systemic inflammatory conditions at the onset of lactation. Conversely, the intense acute phase response observed in BRO cows could have been driven by a stronger activation of leukocytes following parturition, as reflected by the highest concentration of ROMt detected in the plasma from these animals in the first week of lactation (Celi and Gabai 2015). The latter could be consequential to a stronger leukocyte killing capacity, as already documented by Gibson et al. (2016), who found macrophages from BRO cows had a greater ROMt production compared to those from HOL following a bacterial ex-vivo stimulation with LPS.

Nevertheless, trends of acute phase biomarkers in BRO cows were also accompanied by the highest concentration of bilirubin, AST-GOT, and GGT at the onset of lactation. Elevated bilirubin is mainly attributable to the lower synthesis of enzymes responsible for its clearance, while AST-GOT and GGT are markers of liver damage (Bertoni et al. 2008). Thus, liver function was somewhat impaired in BRO cows. Furthermore, although the loss in BCS was lower in BRO compared with HOL cows, they had higher plasma BHB concentration, suggesting a lower oxidising efficiency of the liver against mobilised NEFA (Herdt 2000). Both liver damage and impaired liver oxidation have been reported in dairy cows facing a remarkable acute phase response (Bertoni et al. 2008), and trends of the aforementioned plasma analytes detected in BRO cows during the present study were beyond the physiological ranges reported for HOL dairy cows at the onset of lactation (Premi et al. 2021). Of interest, BRO cows did not show any clinical metabolic disorder and maintained their milk yield throughout the experimental period. Thus, these data could suggest that, despite severe inflammatory conditions, BRO might be more resilient to the intense metabolic changes typical of the transition period. However, the metabolic stress caused by nutrient mammary uptake is likely lower than that occurring in HOL. Analysing the impact of heat stress on BRO cows, Maggiolino et al. (2020) found that, as temperature and humidity rise, BRO does not reduce milk yield. In our study, we observed that, in another stressful condition (as pointed by the blood markers investigated), BRO cows had a similar response. Nevertheless, it may be possible that reference ranges based on HOL are not suitable for BRO, and specific values should be calculated.

Conclusions

This study suggests that the three explored breeds have different adaptations to the systemic inflammatory state occurring at the onset of lactation. The acute phase response observed in SIM cows was mitigated by their lower metabolic load (a consequence of their lower milk yield). Moreover, HOL cows produced more milk and relied on the most efficient activation of the immune system, resulting in a faster resolution of the acute phase response. BRO cows maintained systemic inflammatory conditions longer than the other breeds, but negligible effects were detected regarding their general performances, suggesting this breed might have the greatest ability to cope with the metabolic shifts driven by the acute phase.

We speculate that the immune system function of the three explored breeds could have been affected by the selection procedures aimed at maximising the different productive traits, possibly accounting for the different metabolic adaptations to the transition period detected here. Nevertheless, these should be considered preliminary results due to the limitations represented by the facilities where this study was carried out and the limited sample size. Further research is required to fully elucidate the adaptive strategies adopted by different dairy breeds to cope with the dramatic changes occurring at the onset of lactation.

Acknowledgments

This work was conducted in the framework of projects supported by CREI (Romeo and Enrica Invernizzi Research Center of the Università Cattolica del S. Cuore) funded by the ‘Fondazione Romeo ed Enrica Invernizzi’, Milan, Italy. The authors wish to convey sincere thanks and appreciation to Professor Luigi Calamari (Università Cattolica del Sacro Cuore, Piacenza, Italy), who substantially contributed to the conceptualization of this experiment, but prematurely passed away before completing the data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethical Approval

This study complied with Italian laws on animal experimentation (DL n. 26, 04/03/2014) and ethics (authorization of the Italian Ministry of Health n. 851/2018).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.