1. Introduction
The transition from pregnancy to lactation is one of the most susceptible periods of a dairy cow’s life. Approximately 75% of health disorders in adult cows, such as retained placenta, mastitis, and metritis, typically occur within the first month after calving [
1]. The high susceptibility to these diseases is mainly correlated with an impaired immune response, resulting in drastic metabolic and endocrine changes during the transition period. From a metabolic standpoint, dairy cows commonly experience an increase in energy requirements during the transition period, primarily for fetal growth prepartum, followed by milk synthesis postpartum [
2,
3]. Then, cows show a clear inflammatory response [
4,
5] after calving, although changes in immune response often proceeded parturition [
6,
7,
8]. Additionally, during the prepartum, cows commonly experience a decrease in dry matter intake (DMI) [
9]. Although immediately after parturition, there is a progressive increase in feed intake, this is generally insufficient in parallel with the nutrient requirements in early lactation. This results in a negative energy balance (NEB) condition, and consequently, non-esterified fatty acids (NEFA) are mobilized from the adipose tissue [
10,
11].
During NEB, triacylglycerides (TAG) are hydrolyzed into glycerol and NEFA and released into the bloodstream [
12]. NEFA can be used as an energy source by different tissues, such as the mammary gland for milk fat synthesis, or taken up by the liver [
13]. In the liver, NEFA are either completely oxidized to generate energy, partially oxidized to produce ketone bodies, or re-esterified into TAG. Then, TAG are either exported as very low-density lipoproteins (VLDL) or stored in the hepatocytes [
13,
14]. At periods of intense lipid mobilization, as the transition period, NEFA entering the liver exceeds its capacity of β-oxidation and VLDL secretion, which follows to a greater production of ketone bodies, such as b-hydroxybutyrate (BHB), and the accumulation of TAG in the liver [
14,
15].
The increased plasma NEFA and BHB, aligned with augmented metabolic demand characteristic of the transition period, is commonly accompanied by a surge in reactive oxygen species (ROS) leading to oxidative stress [
11,
16]. Parallel to this, the pro-inflammatory cytokines released by the immune system induce ROS production by both phagocytic and non-phagocytic cells [
11], which can further alter redox balance and lead to an oxidative stress condition during early postpartum. Oxidative stress can be defined as the damage occurring to cellular macromolecules as a consequence of serious and prolonged imbalance between oxidants and antioxidants [
17,
18]. ROS are particularly reactive towards DNA, proteins, and lipids, resulting in cellular alterations, including lipid peroxidation, cell membrane damage, DNA mutation, impaired protein activity, and cell death [
16,
18,
19]. In immune cells, those alterations can result in diminished functional activities, such as reduction in neutrophils phagocytosis and killing capacity [
2]. In addition, oxidative stress has been correlated with increased inflammation through the activation of pro-inflammatory signaling pathways [
16] and with insulin resistance, which can further enhance the lipolytic state of transition dairy cows [
19]. Overall, oxidative stress as an underlying factor for dysfunctional immune response and enhanced inflammation underscores the importance of this condition in the increased susceptibility of dairy cattle to health disorders, particularly during the transition period [
2,
11,
20].
Antioxidant compounds can be synthesized by the body and are also naturally present in feedstuffs; however, for indoor animals fed low forage diets or especially during periods of high antioxidant demands such as the peripartal period, dietary supplementation is necessary to meet the increased requirements [
21,
22,
23]. Vitamin E and selenium are the most widely used antioxidants included in dairy cattle diets, especially in the form of premixes added to the total mixed ration [
11]. However, in the last few decades, there has been an increased interest in studying alternative natural compounds with antioxidant and anti-inflammatory properties as supplements for food-producing animals [
24,
25,
26]. Besides being a source of antioxidants in the final products for the human diet, those natural compounds have the added benefit of being positively perceived by consumers [
22,
27].
Rosemary (
Salvia rosmarinus), an herb from the Labiatae family, is known to have a particularly high concentration of phenolic diterpenes with antioxidant and antimicrobial properties [
28]. Among those phenolic diterpenes, carnosic acid is the most abundant in rosemary leaves [
29], and it has one of the highest antioxidant activity [
30,
31]. Carnosic acid presents a high reactivity toward ROS, and it acts as a scavenger that can eliminate ROS [
28]. Moreover, carnosic acid has been shown to play a role in the activation of the PI3K/Akt/Nrf2 signaling pathway in human cells [
32,
33]. Nuclear factor erythroid-2 related factor 2 (Nrf2) is a transcription factor involved in the cellular response to oxidative stress, inducing the expression of several protective enzymes, e.g., glutathione peroxidase and superoxide dismutase [
32]. Furthermore, carnosic acid has been shown to reduce the expression levels of pro-inflammatory cytokines in human and mice cells [
34,
35] and regulate fatty acid metabolism [
36].
However, in ruminants, studies evaluating carnosic acid have been mainly correlated with an increase in meat quality through reduction of lipid oxidation [
37,
38,
39]. Surprisingly, its potential use as a strategy to reduce the oxidative and metabolic stress of transition dairy cows remains
terra incognita. Based on the above, we hypothesized that providing carnosic acid during early lactation might minimize the typical exposure to oxidative stress and attenuate the typical inflammatory response during peripartum while allowing cows to reach peak performance. The objective of the present study was to evaluate the effects of a compound found in rosemary plants (
Salvia rosmarinus) named carnosic acid during the transition period of dairy cows.
2. Materials and Methods
2.1. Experimental Design and Treatments
The Institutional Animal Care and Use Committee (IACUC) of the South Dakota State University approved all the procedures for this study (protocol no. 2003-017A). The experiment was conducted from September to December 2020 at the South Dakota State University Dairy Research and Training Facility (Brookings, SD). Twenty late pregnant multiparous Holstein dairy cows were used in a randomized complete block design from −21 days prior to expected calving until 21 days in milk (DMI). Cows were blocked according to expected calving day, parity, and previous lactation milk yield, then assigned into one of two treatments. A total of 4 cows were removed from the experiment due to calving outside the range of −7 to 7 d relative to the expected calving date (
n = 2), displaced abomasum (
n = 1), and euthanasia due to low calcium/potassium at calving coupled with a lack of response to treatment (
n = 1) (
Table 1).
From day 1 to 3 after calving, cows received either a daily intravenous infusion of 500 mL of sterile saline solution (NaCl 0.9%; Saline; n = 8) or an infusion with carnosic acid at a rate of 0.3 mg/kg of BW (CA; n = 8). The complete dose of carnosic acid (Combi-Blocks, San Diego, USA, cat. number QC-4383) based on BW was supplied on a total volume of 500 mL of sterile saline solution. Treatments were infused into the external jugular vein before evening milking (17:00).
In rodents, pharmacokinetic studies have been performed using intravenous infusions with CA at a rate of 10 or 20 mg/kg [
40,
41]. To the authors’ knowledge, a carnosic acid pharmacokinetic study in ruminants has never been performed. Milk is the only body fluid that has been used to detect carnosic acid in ruminants supplemented with distilled rosemary leaves [
42]. In this study Jordan, et al. [
42], fed goats a basal diet supplemented with 0, 10, and 20% of the diet with a pellet containing barley and distilled rosemary leaves, and observed a carnosic acid concentration in milk ranging from 0.31 to 0.77 mg/kg. Therefore, given the limitations on pharmacokinetic data on carnosic acid in ruminants and the cost of procuring the CA couple with the large amount of CA needed to supply the corresponding dose for an adult dairy cow, a 0.3 mg/kg BW dose was adopted as the most feasible daily dosage from 1 to 3 d postpartum.
2.2. Animal Management
Cows were enrolled in the experiment from early September 2020 to early January 2021. Weather data from Mesonet at South Dakota State University (
https://climate.sdstate.edu/) was used to evaluate the daily ambient temperature during the experimental period. Cows were fed using an individual gate system (American Calan, Northwood, NH, USA), and intakes were recorded daily. Diets were formulated using the CNCPS model contained within the Agricultural Modeling and Training Systems (AMTS) CattlePro diet-balancing software (version 4.16.1, AMTS LLC, Lansing, NY, USA) to meet the requirements of the average cow in the group (
Table 2). Dry matter content of feed ingredients was determined once a week throughout the experiment, and diets were adjusted accordingly to maintain formulated DM ratios.
During the dry period, cows were housed in bedded pack pens. Immediately after calving, cows were reallocated in individual pens bedded with straw. On day 3, after calving, cows were moved to a lactation free-stall barn. Cows were fed once daily (6:00) and milked twice daily (6:30 and 18:00). Body weight was measured weekly for each cow in the morning (9:00). Body condition score (BCS) (scale 1 = thin to 5 = obese) was assigned by two individuals, and the average score was used for statistical analysis. All cows received the same close-up diet (1.46 Mcal of NEL/kg and 15.1% CP;
Table 2) and lactation basal diet (1.75 Mcal of NEL/kg and 19.6% CP;
Table 2), as a total mixed ration.
2.3. Blood Collection and Analyses
Blood was sampled from the coccygeal vein before morning feeding using a 20-gauge vacutainer needle (Becton Dickinson, Franklin Lakes, NJ) at –7, 2, 5, 7, 14, and 21 d relative to parturition. Blood was collected into evacuated tubes (BD Vacutainer, Becton Dickinson, Franklin Lakes, NJ) containing either serum clot activator or lithium heparin. After collection, tubes that contained lithium heparin were placed on ice, and tubes with serum clot activator were kept at 21 °C until centrifugation. Serum and plasma were obtained by centrifugation at 1300× g for 15 min at 21 °C and 4 °C, respectively. The aliquots were frozen at −80 °C until further analysis.
Blood samples were analyzed for biomarkers related to energy metabolism [i.e., glucose, β-hydroxybutyric acid (BHB), non-esterified fatty acids (NEFA)], muscle mass catabolism (i.e., urea and creatinine), inflammation (i.e., ceruloplasmin and haptoglobin), liver function [i.e., albumin, bilirubin, glutamic-oxaloacetic transaminase (GOT), γ-glutamyltransferase (GGT), cholesterol, and paraoxonase (PON)], and oxidative stress [i.e., myeloperoxidase (MPO), reactive oxygen metabolites (ROM), ferric reducing antioxidant power (FRAP), nitrates (NO
3−), nitrites (NO
2−), nitric oxide metabolites (NO
x), and oxygen radical absorbance capacity (ORAC)]. Furthermore, we calculated the ratios between oxidants and antioxidant defenses, e.g., ROM/FRAP and ROM/ORAC. Those ratios provide an integrated oxidant status index, which seems to better assess changes in oxidative status during transition period [
43,
44].
Albumin, cholesterol, bilirubin, urea, creatinine, GOT, GGT, and glucose were analyzed using the IL Test purchased from Instrumentation Laboratory Spa (Werfen Co., Milan, Italy) in the ILAB 600 clinical auto-analyzer (Instrumentation Laboratory, Lexington, MA, USA), following the procedures described previously [
45,
46,
47]. Haptoglobin was analyzed using the method described by Skinner, et al. [
48], while ceruloplasmin was determined based on Sunderman and Nomoto [
49], with modifications described by Jacometo, et al. [
50]. Antioxidant potential was assessed as ferric reducing antioxidant power (FRAP) using a colorimetric method [
51]. Paraoxonase, NO
x, NO
2−, and NO
3− were analyzed according to methods described by Trevisi, et al. [
52]. Myeloperoxidase was determined via colorimetry based on the reaction of MPO contained in the plasma sample with hydrogen peroxide, which forms H
2O and O
−; the O
− dianisidine dihydrochloride, and electron donor, reacts with the O
−, releasing H
2O and a colored compound [
50,
53]. Non-esterified fatty acids and BHB were measured using kits from Wako (Chemicals GmbH, Neuss, Germany) and Randox (Randox Laboratories Ltd., Crumlin, UK), respectively, following the procedures described previously [
45,
53,
54]. Finally, total antioxidants were assessed through the oxygen radical absorbance capacity (ORAC) assay. This method estimates the overtime antioxidant capacity to inhibit phycoerythrin hydroxyl radical damage [
55].
2.4. Milk and Feed Samples
Total mixed ration samples were collected weekly and frozen at −20 °C after DM analysis until further nutrient profile analysis. Monthly composites were analyzed for contents of DM, CP, NDF, and ADF and NEL was calculated using wet chemistry methods at a commercial laboratory (Dairy One; Ithaca, NY, USA).
Consecutive morning and evening milk samples were collected once weekly until 21 DMI. Composite milk samples were performed in proportion to milk yield at each milking, preserved (Broad Spectrum Microtabs II, Advanced Instruments, Norwood, MA, USA), and analyzed for fat, protein, lactose, solids, milk urea nitrogen (MUN), and somatic cell count (SCC) (Dairy One; Ithaca, NY, USA). Energy corrected milk (ECM) was calculated based on milk yield and milk sample analysis as follows: ECM = (12.82 × fat yield (kg)) + (7.13 × protein yield (kg)) + (0.323 × milk yield (kg)) [
56].
Energy balance (EB) for each cow was calculated based on equations described previously [
21]. The net energy intake (NE
I) was determined based on daily DMI multiplied by NE
L density of the diet, and net energy of maintenance (NE
M) was calculated as BW
0.75 × 0.080. Requirements of net energy of lactation (NE
L) were calculated as NE
L = (0.0929 × fat % + 0.0547 × protein % + 0.0395 × lactose %) × milk yield. The net energy requirement for pregnancy (NE
P) was calculated as NE
P = ((0.00318 × day of gestation − 0.0352) × (calf birth weight/45))/0.218. The equation used to calculate prepartal EB (EB
PRE; Mcal/d) was EB
PRE = NE
I − (NE
M + NE
P) and EB
PRE (as % of requirements) = (NE
I/(NE
M + NE
P)) × 100. Finally, to calculate postpartal EB (EB
POST), the equation used was EB
POST (Mcal/d) = NE
I − (NE
M + NE
L) and EB
POST (as % of requirements) = (NE
I/(NE
M + NE
L)) × 100.
2.5. Statistical Analysis
The effects of carnosic acid blood biomarkers were evaluated separately at 2 d relative to parturition (during infusions) and from 5 to 21 d postpartum as residual effects. Performance data and residual effects were evaluated repeated measures using the MIXED procedure of SAS 9.4 (SAS Institute Cary NC, USA). The statistical model contained the effects of treatment, time (day or week), and their interactions as fixed effects, while the cow within treatment was considered as a random effect. Single time-point data were analyzed following the same model, without the time statement. Blood biomarkers were log-scale transformed if needed to comply with normal distribution of residuals.
Residual data on blood biomarkers from 5 to 21 d postpartum was unequally spaced; therefore, the SP(POW) covariance structure was used for this analysis. For the equally spaced measures, the covariance structure was chosen between first-order autoregressive and heterogeneous first-order autoregressive based on goodness of fit (smaller Akaike information criteria). Covariates, including previous 305 d milk yield, prepartum DMI, ambient temperature, and blood metabolites at –7 d relative to calving, were maintained in the model when
p ≤ 0.20. Observations were considered outliers when Cook’s distance >0.50 and consequently excluded from the analysis. The CORR procedure of SAS was used to test the Pearson correlation coefficient (
r) between milk performance and prepartum DMI, BW prepartum, and change in energy balance. The occurrence of health problems was analyzed using the FREQ procedure of SAS and interpreted based on Fisher’s exact test probabilities. However, none of the health issues observed in this experiment were affected (
p ≥ 0.26) by treatment (
Table 1). Statistical significance was declared at
p ≤ 0.05 and tendencies at
p ≤ 0.10.