Effects of feeding 25-hydroxyvitamin D 3 with an acidogenic diet during the prepartum period in dairy cows: Mineral metabolism, energy balance, and lactation performance of Holstein dairy cows

Our objective was to determine the effects of feeding 25-hydroxyvitamin D 3 [25(OH)D 3 ], or vitamin D 3 (cholecalciferol) on plasma, mineral, and metabolite concentrations, mineral balance, mineral excretion, ru-mination, energy balance, and milk production of dairy cows. We hypothesized that supplementing 3 mg/d of 25(OH)D 3 during the prepartum period would be more effective than supplementing vitamin D 3 at the National Research Council (2001) levels to minimize calcium imbalance during the transition period and improve milk production of dairy cows. Forty multiparous, pregnant nonlactating-Holstein cows were enrolled in this study. Body weight, body condition score, parity, and milk yield in the previous lactation (mean ± standard deviation) were 661 ± 59.2, 3.46 ± 0.35, 1.79 ± 0.87, and 33.2 ± 6.43 kg/d, respectively. Cows were enrolled into the blocks (n = 20 for each treatment) at 30 d of the expected day of calving to receive an acidogenic diet (373 g/kg of neutral detergent fiber and 136 g/ kg of crude protein, dry matter basis; −110 mEq/kg) associated with the treatments: (1) control (CTRL), vitamin D 3 at 0.625 mg/d (equivalent to 25,000 IU of vitamin D 3 /d) or (2) 25(OH)D 3 at 3 mg/d (equivalent to 120,000 IU of vitamin D 3 /d). All cows were fed with the base ration for 49 d after calving. Blood samples were taken on d 7, 0, 1, 2, 21, and 42, relative to calving. No effect of treatment was observed for prepartum dry matter intake or body condition score. A trend for increase of ionized Ca was observed for the cows fed 25(OH)D 3 , compared with the CTRL, but no effect of treatment was detected for total Ca or total P. Feeding 25(OH)D 3 increased colostrum yield. The plasmatic concentration of 25-hydroxyvitamin D 3 was increased with 25(OH)D 3 supplementation. 25-Hydroxyvitamin D 3 supplementation increased plasma glucose concentration at parturition. The postpartum dry matter intake was not influenced by treatments. Feeding 25(OH) D 3 increases milk yield, 3.5% fat-corrected milk, and energy-corrected milk and improves milk yield components in early lactation. Overall, these findings suggest that 25(OH)D 3 at 3 mg/d can improve the energy metabolism and lactation performance, compared with the current-feeding practice of supplementing vitamin D 3 at 0.625 mg/d.


INTRODUCTION
The transition period is the most challenging phase for dairy cows due to all the metabolic and physiological changes that occur over late pregnancy and early lactation (Bell, 1995;Huzzey et al., 2005;Glosson et al., 2020).The negative energy balance (NEB) faced by cows plays a critical role in many changes that occur during the transition period, such as an increase in plasma concentrations of ketone bodies (Cardoso et al., 2020) and nonesterified fatty acids (NEFA; Wankhade et al., 2017), impaired neutrophil function (Hammon et al., 2006), fatty liver (McArt et al., 2012), and hypocalcemia (Horst et al., 1994).
Calcium metabolism plays an important role during the transition period and into the subsequent lactation period (Golder et al., 2021).Impaired health and reproduction were observed in cows that developed clinical and subclinical hypocalcemia (Martinez et al., 2012a; Effects of feeding 25-hydroxyvitamin D 3 with an acidogenic diet during the prepartum period in dairy cows: Mineral metabolism, energy balance, and lactation performance of Holstein dairy cows Rodríguez et al., 2017).Feeding acidogenic prepartum diets with a negative DCAD results in stimulation of Ca mobilization before calving and decreases clinical and subclinical hypocalcemia postpartum (Diehl et al., 2018;Rodney et al., 2018;Glosson et al., 2020).
The NRC (2001) recommends 25,000 IU/d of vitamin D for a 680-kg mature dairy cow during the transition period; however, some studies have shown that vitamin D 3 supplementation was not enough to prevent subclinical hypocalcemia (Wilkens et al., 2012;Martinez et al., 2018b).Additionally, there are not sufficient data to indicate whether 25,000 IU/d of vitamin D 3 is adequate for optimal Ca homeostasis.According to Martinez et al. (2018b), replacing vitamin D 3 with 25(OH)D 3 (at 3 mg/d) during the prepartum period can improve postpartum lactation performance of Holstein cows.Poindexter et al. (2020) also concluded that feeding 25(OH)D 3 (at 1 or 3 mg/d) increased serum 25(OH)D 3 more effectively than supplemental vitamin D 3 , resulting in increased serum mineral concentrations.Additionally, supplementing vitamin D 3 at 3 mg/d increased mammary immunity when compared with supplementing 1 mg/d (Poindexter et al., 2020).The use of an acidogenic diet associated with an active form of vitamin D would be more effective in improving Ca homeostasis reflecting a more integrated bone and mineral metabolism (Rodney et al., 2018a).However, these studies were performed under temperate settings, with greater milk production (+2.5 kg/d) and lower acidogenic salt addition to achieve a greater negative DCAD (−130 mEq/kg of DM; Martinez et al., 2018a) needed to promote a reduction in the blood pH.In addition, the greater K content of tropical forages (McNeill et al., 2002) presents a greater challenge in formulating diets that induce a compensated metabolic acidosis (Wildman et al., 2007).
To our knowledge, no studies have evaluated 25(OH) D 3 , in association with an acidogenic diet, fed to Holstein dairy cows under tropical conditions in the transition period.We hypothesized that supplementing 3 mg/d of 25(OH)D 3 to Holstein cows during the transition period would increase the serum Ca concentration and, thus, increase the urine mineral excretion compared with the traditional feeding practice of supplementing 0.625 mg/d of vitamin D 3 (NRC, 2001).We also hypothesized that supplementation with 25(OH)D 3 could improve the homeostatic mechanisms of Ca, decreasing the negative energy status and increasing the productive performance of dairy cows (Rodney et al., 2018b).Our objective was to evaluate the effects of feeding vitamin D 3 at 25,000 IU/d (0.625 mg/d) or 25-hydroxyvitamin D 3 at 120,000 IU/d (3 mg/d) during the last 30-d prepartum period associated with an acidogenic diet on vitamin D and mineral metabolism, colostrum yield, and composition and lactation performance of Holstein dairy cows.

MATERIALS AND METHODS
The experiment was conducted at the "José Henrique Bruschi" Experimental Field, Embrapa Dairy Cattle, Coronel Pacheco, Minas Gerais, Brazil (21°33′22″ S, 43°06′15″ W).According to Köppen's classification, the climate of the region is a humid subtropical-climate, mesothermic, with hot and rainy summer and cold and dry winter seasons.All the procedures were approved by the Embrapa Dairy Cattle Animal Care and Use Committee (protocol #7946051218).

Animals, Treatment Diets, and Housing
Forty multiparous-Holstein cows were enrolled in the study.Body weight, BCS, lactation number, and milk yield in the previous lactation (mean ± SD) were 661 ± 59.2 kg, 3.46 ± 0.361, 1.79 ± 0.865, and 33.2 ± 6.43 kg/d, respectively.Sample size was calculated with the POWER procedure of SAS (version 9.3, SAS Institute Inc.) to detect differences in blood concentrations of total Ca, ionized Ca, and vitamin D metabolite as described by Rodney et al. (2018a).The experiment was a randomized complete-block design with cows as the experimental unit.Cows were blocked according to the expected calving date (n = 40 cows; 5 blocks, 8 cows/ block, and 20 cows/treatment) to receive the acidogenic diet ( 373 Cows were housed in a freestall barn equipped with electronic devices (feeders, water drinkers, and weighing devices; Intergado) and a cooling system (DeLaval) consisting of automatic fans and sprinklers located in the feed line.
During the 30 d prepartum, cows were fed the acidogenic ration containing the treatments (Table 1) once daily at 0800 h.Upon calving, all cows were fed the base lactation ration twice daily at 0800 and 1530 h during 49 DIM.All diets were fed as TMR (Table 1), formulated to meet the nutrient requirements specified by the NRC (2001).Refusals were weighed daily before the morning feeding, and feed allowances were calculated daily with a goal of 5% refusals.

Chemical Analysis and Measurement of Dry Matter Intake
Feed samples (ingredients, diet offered, and orts) were collected daily, dried in a forced ventilation oven at 55°C for 72 h, ground (Wiley mill; A. H. Thomas) through a 1-mm screen sieve, and composited (DM basis) by week for each cow and then for each treatment.Samples were analyzed for DM, OM, total N, and ether extract according to methods 930.15, 942.05, 984.13, and 920.39, respectively (AOAC International, 2005).To estimate the CP, a conversion factor of 6.25 was used to convert N values into CP.Neutral detergent fiber and ADF were determined sequentially using Ankom 220 Fiber Analyzer (Ankom Technology).For NDF analysis, a thermostable α-amylase was used without the addition The energetic density of the diets was estimated using chemical analysis of dietary ingredients, calculated for 11.0 and 20.0 kg of DMI for the prepartum and postpartum periods, respectively, using the NRC (2001).

Body Weight and Body Condition Score
Cows were weighed daily throughout the trial using the electronic weighing device (Intergado).The BCS was evaluated on the day of enrollment in the experiment and then weekly throughout the trial by the same trained evaluator, using a 1 to 5 scale (Ferguson et al., 1994) with increments of 0.25 units as described in the Elanco BCS chart (Elanco Animal Health, 2009).

Yield and Mineral Content of Colostrum
Cows were milked within the first 6 h after calving and colostrum yield was measured.Duplicate samples were collected and frozen at −20°C for later analysis of total Ca and Mg according to Bird et al. (1961).

Net Energy Balance
Energy balance was measured using the caloric intake obtained from the DMI and the energy content of the diets estimated according to NRC (2001), using the NE L system.The energy requirement for maintenance was calculated according to NRC (2001), using the metabolic BW (0.08 × BW 0.75 ).Calories required for gestation during the prepartum period were estimated at 3.7 Mcal of NE L /d for a calf that would eventually be born with a BW of 43 kg (NRC, 2001).The calories secreted as milk were calculated according to yields of fat, protein, and lactose according to NRC (2001): calories secreted = milk yield × [(0.0929 × fat %) + (0.0563 × protein %) + (0.0395 × lactose %)].

Blood Sample Collection
Blood samples were collected 2 h after the morning feeding (between 0800 and 1000 h) by puncture of the coccygeal blood vessels into 10-mL evacuated tubes (Vacutainer, Becton Dickinson) containing no anticoagulant agents for serum separation or into tubes containing K 3 -EDTA or sodium fluoride for plasma separation.After collection, samples were immediately placed on ice until processing.The tubes were centrifuged for 15 min at 2,500 × g (LS-3 Plus, CELM) at environmental temperature (25°C), for plasma and serum separation.Plasma and serum samples were transferred into multiple aliquots of 2.5 mL and stored in a freezer at −20°C for later analyses.Serum ionized Ca (iCa), total Ca (tCa), total Mg (tMg), and total P (tP) were analyzed immediately after collection.Blood samples used for plasma vitamin D analysis were centrifuged for 10 min at 1,800 × g at environmental temperature (25°C) for plasma separation and stored at −80°C until analysis.

Measurements of Serum Minerals
Serum iCa, tCa, tMg, and tP were determined on d 7 relative to calving and on d 0, 1, 2, and 21 after calving, using the ion-selective technique with an automatic ion analyzer (MHlab-ise; MH Equipamentos e Materiais para Laboratório Ltda.).

Vitamin D and Parathyroid Hormone in the Plasma
Plasma samples were collected on d −7 and 0, relative to calving, and on d 21, after calving, and analyzed for 25(OH)D 3 , using HPLC coupled with mass spectrometry (LC-MS/MS) detection by the LabFor Análises Laboratoriais Ltda.Parathyroid hormone was analyzed by chemiluminescence (VetLab).

Urine Collection and Analyses
Urine samples were collected twice per week before calving and on d 21 and 42 after calving, by massaging  TO DAIRY COWS the perineal area, until a copious stream of urine was obtained.Samples were stored in plastic tubes on ice and the pH was measured within 10 min after collection using a portable pH meter (MPA 210, MS Tecnopon).Creatinine concentration was analyzed by the calorimetric method (ref. 35,Labtest).Calcium and Mg concentrations in the urine samples were performed by spectrophotometry using an automatic analyzer (ref. 35,Labmax Plenno,Labtest).
Creatinine was used as a marker to estimate daily urinary volume using the constant value of 24.04 mg of creatinine per kg of BW per day (Chizzotti et al., 2008).The estimate of daily urine volume was calculated as follows: BW × 24.04 urinary concentrations of creatinine (mg/L) .
Daily urinary excretion of Ca and Mg were calculated as a product of urinary volume, and the concentrations of those minerals in the urine sample.

Estimated Mineral Balance
Calcium and Mg balances were calculated based on the differences in Ca and Mg intake and urinary losses of Ca and Mg during the last week of gestation, as well as the Ca and Mg concentrations in the colostrum on the day of calving.Estimated absorptions of Ca and Mg and fetal tissues accretion were computed according to diet and calf BW at birth (NRC, 2001).The mineral digestion was performed using 1.0 g of sample, burned at 525°C by 8 h, then, the residue was dissolved in 6.0 mL of nitric acid (HNO 3 ) with 25 g/L of perchloric acid (HClO 4 ; nitroperchloric digestion; Sarruge and Haag, 1974).The determination of mineral concentration (Ca, Mg, and K) in the diets was made by atomic absorption using a spectrophotometer (model AA-6800, Shimadzu Scientific Instruments).The P concentration analysis was performed using a spectrophotometer (UV Genesys 10s Vis, Thermo Scientific).

Physiological Responses
Seven days before calving and on d 0, 7, and 21 after calving, blood samples were collected for determination of BHB, NEFA, insulin, and glucose concentration.The glucose concentrations in the plasma were analyzed by glucose oxidation with glucose oxidase kits (Glucose PAP Liquiform, Labtest), using a microplate spectrophotometer and the software of interphase Gen5 (Eon Biotek).Nonesterified fatty acids and BHB were analyzed using colorimetric enzymatic assays (Randox Laboratories Ltd.).Insulin concentration was analyzed by enzyme immunoassays (Bovine Insulin ELISA, Mercodia Inc.).The threshold (iCa ≤1.0 mM) was selected to define subclinical hypocalcemia using whole blood (Oetzel et al., 1988).Hyperketonemia was defined as serum BHB concentrations >1.20 mM (Martinez et al., 2012).

Rumination Monitoring
Total time spent ruminating was measured using an Hr-Tag rumination monitoring system (SCR Engineers Ltd.).The system consisted of neck collars with rumination loggers built in, stationary readers, and software for processing the electronic data (Data Flow software, SCR Engineers Ltd.).A neck collar positioned the logger on the left side of the neck.The logger contained a microphone to record the distinctive sounds of regurgitation and rumination.Data were calculated and summarized at 2-h intervals and stored in the memory of the logger.For data analysis, the rumination time was taken during wk 1, 2, 3, 4, 5, and 6 relative to calving.

Statistical Procedures
The experiment was a randomized complete block design with cows used as the experimental unit in each block.The prepartum and postpartum periods were analyzed separately.The normality of the residues and the homogeneity of variance were examined for each continuous dependent variable analyzed after model fitting.Variables that violated the assumptions of normality were subjected to transformation, according to the Box-Cox procedure (Box and Cox, 1964), using the TRANSREG procedure of SAS version 9.3 (SAS/ STAT, SAS Institute Inc.).All data were analyzed using the MIXED procedure of SAS.The Kenward-Roger method was used to calculate the approximate denominator degrees of freedom.The statistical model included the fixed-effect treatment of vitamin D 3 (CTRL), or 25-hydroxyvitamin D 3 25(OH)D 3 , experimental day (or week for milk yield and composition), and its interaction.Block was included as a random effect, as well as cows nested within the interaction of treatment and day.All results were reported as least squares means.Significance was set at P ≤ 0.05 and tendencies were determined if P > 0.05 and ≤0.10.

Colostrum Yield and Mineral Content
Feeding 25(OH)D 3 increased (P = 0.02) colostrum yield compared with the CTRL (5.44 vs. 7.50 kg, respectively).Treatments did not affect (P ≥ 0.13) the concentration (g/L) and secretion (g/d) of Ca and Mg in colostrum (Table 3).
No treatment × day interactions were observed for serum concentration of tCa, tP, tMg, and PTH (P ≥ 0.26; Table 4).However, a treatment × day interaction was detected for plasma 25(OH)D 3 (P < 0.01; Table 4).Plasma 25-hydroxyvitamin D 3 was greater for cows fed 25(OH)D 3 compared with the CTRL (P < 0.001) from d −7 to 21 relative to calving (Figure 1D).An effect of day (P < 0.01) was observed for serum concentration of tCa and tP (Figure 1B), and PTH (Figure 1C).Serum concentration of tCa was less on d 2 and 21 compared with d −7 (P < 0.05).Total serum concentration of P was greater on d 2 compared with d −7, 0, and 21 (P < 0.05).Serum concentration of PTH was greater on d 21 compared with d −7 (P < 0.05; Figure 1C).

Blood Metabolites
No treatment or treatment × day interactions were observed for plasma concentration of insulin (P ≥ 0.44; Table 5), but an effect of day was observed (P = 0.004; Table 5).The concentration of insulin was less on d 7 compared with d −7 and 21 (P < 0.05), but not compared with d 0 (P = 0.27; Figure 2B).
A treatment × day interaction was observed for glucose and NEFA (P ≥ 0.004; Table 5).The concentration of glucose was greater for cows fed 25(OH)D 3 on d 0 compared with the CTRL (Figure 2A: P ≤ 0.05), and NEFA was less for cows fed 25(OH)D 3 on d 7 postpartum compared with the CTRL (Figure 2C; P ≤ 0.05).
No treatment, day, or treatment × day interactions were observed for plasma concentrations of BHB (P ≥ 0.10, Table 5).
Treatments did not affect the percentage of protein, fat, and lactose, or urea (P ≥ 0.78; Table 6).

Postpartum Dry Matter Intake, Energy Balance, and Rumination Time
A tendency of treatment × week interaction was observed for NE intake and NEB (P = 0.09; Table 7; Figure 4C).Cows fed 25(OH)D 3 had lower energy intake only during the first week of lactation, but CTRL had lower NE intake during the 4 wk after parturition (P < 0.05; Figure 4A).
An effect of week was observed for DMI, NE secreted as milk, BW, BW change, and BCS (P < 0.003; Table 7).Overall all these variables increased with the progress of lactation (data not shown).
No effects of treatment, day, or treatment × day interaction were observed for postpartum urinary volume, creatinine, and Mg excretion (g/d; P ≥ 0.23; Table 8).A treatment × day interaction was detected for urinary concentration of Mg (P = 0.02; Table 8).The urinary concentration of Mg (mg/L) was higher on d 42 for cows fed CTRL than 25(OH)D 3 (P ≤ 0.05), and for cows fed 25(OH)D 3 the opposite occurred (P = 0.019; data not shown).

DISCUSSION
The Panel A: effect of treatment (P = 0.06), effect of the day (P = 0.02), and interaction between treatment and day (P = 0.28).Panel B: effect of treatment (P = 0.38), effect of the day (P < 0.0001), and interaction between treatment and day (P = 0.34).Panel C: effect of treatment (P ≥ 0.16), effect of the day (P < 0.0001), and interaction between treatment and day (P ≥ 0.41).Means with different letters (a, b) are different (P < 0.05).Panel D: effect of treatment (P < 0.0001), effect of the day (P < 0.0001), and interaction between treatment and day (P < 0.0001).Error bars denote SEM.conditions.Under tropical conditions, the incidence of sunlight is higher than in temperate conditions, so the cows in the tropics may synthetize vitamin D naturally, which can influence the effects of vitamin D supplementation for dairy cows, although hypocalce-mia is still persistent in tropical settings (Paiano et al., 2020).Additionally, the greater K content of tropical forages presents a greater challenge in formulating diets that induce a compensated metabolic acidosis.Therefore, the use of a more effective vitamin D metabolite effect of treatment (P = 0.0017), effect of the day (P < 0.001), and interaction between treatment and day (P = 0.0036).Panel B: effect of treatment (P = 0.43), effect of day (P < 0.004), and interaction between treatment and the day (P = 0.52).Panel C: treatment effect (P = 0.0001), effect of day (P < 0.0001), and interaction between treatment and day of collection (P = 0.0001).Panel D: effect of treatment (P = 0.36), effect of day (P < 0.10), and interaction between treatment and day (P = 0.76).Means with different letters (a,b) are different (P < 0.05).Error bars denote SEM; asterisks denote interaction between vitamin D source and day of collection.[25(OH)D 3 ] can be important in farms in the tropics to decrease the incidence of hypocalcemia and to improve the performance of Holstein dairy cows.To our knowledge, this is the first study reporting a decrease in plasma NEFA concentration and an increase in plasma glucose concentration, suggesting an improvement in energy metabolism of Holstein dairy cows in the transition period that were fed 25(OH)D 3 .
One of the most common effects of acidogenic diets is the reduced DMI when compared with alkalogenic diets in the prepartum period (Lean et al., 2019;Santos et al., 2019).The exact mechanism involved in this reduction in the number of cows fed acidogenic diets is not completely understood, but the metabolic acidosis caused by these diets can be one of the reasons for the reduction of DMI in the prepartum period (Zimpel et al., 2018).Average DMI in the current study was similar to the report by Santos et al. (2019), which observed 11.2 kg/d of DMI for multiparous Holstein cows with a weight, BCS, and milk yield of 715 kg, 3.57, and 36.7 kg/d, respectively, after being fed an acidogenic diet.However, in contrast to Martinez et al. (2018a), in the present study cows fed 25(OH)D 3 tended to increase the prepartum DMI (+1.0 kg/d).
The lack of treatment effect on prepartum energy balance can be explained by the similar NE of the diets, although the caloric intake was higher for cows fed 25(OH)D 3 supported by the higher DMI.Martinez et al. (2018a) did not observe differences in prepartum energy balance or BW change for cows fed vitamin D 3 or 25(OH)D 3 21 d before calving.Cows fed 25(OH)D 3 had a higher BW change in the present study, possibly related to higher weight of the calves at birth (data not shown).
Around calving days, energy requirements increase due to colostrum production and increased mammary gland metabolism.The increase in colostrum yield (+2.06 kg/d) observed in the current study for cows fed 25(OH)D 3 is probably due to the higher glucose concentration in the plasma for 25(OH)D 3 group at calving day.Glucose is an important substrate to produce lactose (Cant et al., 2002) and, consequently, one of the main substrates for milk production (Danes et al., 2020).Cows fed 25(OH)D 3 tended to increase the DMI (+1.0 kg/d) and this could explain the higher glucose concentration on calving day that resulted in higher colostrum yield; however, the extra amount of colostrum did not result in more Ca secreted.In contrast to the current study, Rodney et al. (2018a) observed a trend for higher colostrum yield and Ca secretion in colostrum from cows fed 25(OH)D 3 .
In the present study, the iCa and tCa concentrations were 1.3 and 2.5 mM, respectively.This normocalcemic state is likely due to an effective acidogenic diet, given that the urine pH was below 6.0 for both groups in the prepartum period.The value to diagnose hypocalcemia is iCa <1.0 mM (Oetzel et al., 1988), tCa <2.0 mM (Reinhardt et al., 2011), or tCa <2.15 mM (Martinez et al., 2012) and even though the cows remained with a mean value above the threshold, a decrease in the serum Ca at calving was present when compared with prepartum and 21 d postpartum.The lack of effect of treatment on plasmatic PTH is probably due to very minimal differences in serum Ca at calving.Additionally, PTH is mainly a calcicotrophic hormone when Ca in the blood is under normal levels (Hernández-Castellano et al., 2020).Calcium is so tightly regulated that very low Ca concentration in the serum could promote changes in plasma PTH (DeGaris and Lean, 2008).Even the PTH challenge model by Vieira-Neto et al. ( 2021) could barely detect changes in 1,25(OH) 2 D 3 ; therefore, in the present study, it is unlikely a change in PTH would be observed from dietary vitamin D.
When tCa is <2.5 mM, the parathyroid gland produces PTH (Goff et al., 2014).As mentioned above, the cows in the present study had normal levels of blood Ca.Another reason all cows remained normocalcemic was that the serum Mg concentration was above 2.2 mg/dL.The Mg concentration above 1.9 mg/dL suggests no suppression of PTH responsiveness (Rude and Gruber, 2004).
In the current study, cows fed 25(OH)D 3 tended to increase (+0.07 mM) the serum iCa concentration in the postpartum period; however, Rodney et al. (2018a) observed an increase in serum iCa concentration in the prepartum period, but not during the postpartum period for cows fed 25(OH)D 3 at 3 mg/d associated with a negative DCAD diet (−126.6 mEq/kg of DM).The increase in serum iCa observed in the present study (+0.07 mM) is consistent with Poindexter et al. ( 2020) and Rodney et al. (2018a), which also evaluated the effects of feeding an acidogenic diet supplemented with 25(OH)D 3 during the transition period.
The intestinal absorption of Ca is mediated by the transcellular pathway through the Ca entrance in the specific sites present in the brush border and, finally, an intracellular transport mediated by Ca-binding protein and calbindin-D 9k (Lieben et al., 2011).The activation of intestinal absorption of Ca is mediated mainly by the active form of vitamin D. In the current study, the cows fed 25(OH)D 3 had 268% higher 25(OH)D 3 in the plasma.This increase of plasma 25(OH)D 3 suggests an increased Ca intestinal absorption; however, this is not conclusive in our study because we do not use markers for the determination of Ca absorption, although previous data showed increased Ca absorption in the intestine of Holstein dairy cows fed 25(OH)D 3 (Oehlschlaeger et al., 2014).In addition, the involvement of vitamin D on intestinal absorption of Ca is largely acceptable in the literature (McGrath et al., 2013).
Moreover, cows fed vitamin D 3 had a mean 25(OH) D 3 of 66.6 ng/mL during the experiment in the present study.Dairy cows fed 25,000 IU/d of vitamin D 3 normally present this range of plasma 25(OH)D 3 , although the data of the present study suggest that it is not enough to improve Ca metabolism.In a previous study, Rodney et al. (2018a) found a possible limitation of the conversion of vitamin D 3 to 25(OH)D 3 on the vitamin D 3 supplemented group with Holstein dairy cows fed 21 d before calving with 120,000 IU/d vitamin D 3 .This lower blood plasma 25(OH)D 3 for the vitamin D 3 treatment possibly decreased the intestinal absorption of Ca compared with cows fed 25(OH)D 3 in the present study.Additionally, at the onset of lactation, the Ca supplied from the diet is the main source of Ca for the transition dairy cows because the supply is mainly mediated by the 1,25(OH) 2 D 3 (Horst et al., 1994) that increases the Ca efflux and decreases the Ca influx in bone (Felsenfeld et al., 2013).
When the blood concentration of Ca and Mg increases above normal levels, the urinary excretion increases to maintain normal physiological levels.Therefore, the lack of effect of 25(OH)D 3 supplementation in urinary Mg excretion in the pre-and postpartum periods re-  Rodney et al., 2018a).Treatment did not affect tP in the current study (mean = 2.22 mM).This is important to note because P concentration in blood is directly related to P intake (Horst, 1986).In the current study, the P concentration of the prepartum diets was 0.27%; therefore, high P diets can decrease the conversion of 25(OH)D 3 to 1,25(OH) 2 D 3 by impairment of the hydroxylation mediated by 1-α hydroxylase (Anderson et al., 2017).
A large amount of glucose is needed for a variety of functions and physiological mechanisms such as the immune system, energy metabolism, milk production, and reproduction.The glucose requirement is usually 2.7 times greater for cows in early lactation (Sundrum, 2015).In the present study, supplementation with 25(OH)D 3 increased plasma glucose concentration by 0.45 mM, which means more glucose was available to be used for a variety of functions.Cows in early lactation can experience some degree of insulin resistance in response to physiological changes (De Koster and Opsomer, 2013;Mair et al., 2016) and low insulin concentration (Gärtner et al., 2019).The present study, in agreement with Martinez et al. (2018a), found insulin concentration was not affected by treatments.
Lipolysis also occurs frequently in early lactation due to great energetic demand during this period (Contreras and Sordillo, 2011).Lipid mobilization is related to decreased postpartum DMI, lower percentage of neutrophils with oxidative burst activity, and lower production of reactive oxygen species per neutrophil (Martinez et al., 2014).Plasma concentration of NEFA  is exacerbated by the increased demand for energy during the postpartum period followed by low DMI.In the present study, cows fed 25(OH)D 3 had lower plasma NEFA concentrations compared with the CTRL probably due to the higher glucose concentration on the day of calving.In the present study, the average plasma NEFA concentration on the cows fed 25(OH)D 3 during the experimental period was 0.32 mM, with the higher values on calving day (0.68 mM) and 7 d postpartum (0.58 mM).Similar to diabetes type 2, insulin secretion is not compromised by β cells in the pancreas, but the glucose uptake on target tissues is affected by low insulin sensitivity.As mentioned earlier, no effect of the treatment on insulin concentration was observed in the present study; however, insulin resistance can occur in specific target tissues (De Koster and Opsomer, 2013), and the higher lipid mobilization of the cows fed vitamin D 3 supports this hypothesis in the present study.Furthermore, Weber et al. (2013) showed that cows classified with low, medium, and high fat liver concentrations had higher glucose and lower plasma NEFA concentrations.
The plasma BHB concentration was not affected by treatments in the present study.Martinez et al. (2018a) did not observe an effect of the source of vitamin D on plasma BHB concentration.One possible explanation is that the liver's oxidative capacity was not exceeded in the present study.Plasma BHB levels are usually lower than 1.2 mM and when they increase to levels above 1.2 to 1.5 mM, cows can exhibit subclinical or clinical ketosis, respectively (Iwersen et al., 2013).In the present study, the cows fed 25(OH)D 3 or the CTRL had 0.72 and 0.70 mM plasma BHB concentration, respectively.
Cows fed 25(OH)D 3 produced 2.8 kg/d more milk than those fed CTRL.During lactation, most of the glucose available (50 to 80%) is directed to the mam-  (Duhlmeier et al., 2005;Zhao and Keating, 2007).In this context, Danes et al. (2020) observed an increase of 6.4 kg/d of milk yield for cows receiving abomasal infusion of 3.1 Mcal/d of glucose.In the current study, cows fed 25(OH)D 3 increased plasma glucose concentration and milk yield.Glucose in ruminants is mostly from the hepatic gluconeogenesis, and its transport is mediated by glucose transporter (GLUT4) stimulated by insulin secretion and GLUT1 to the mammary gland that does not depend on insulin signals (De Koster and Opsomer, 2013).Although not expected, the plasma insulin concentration was not influenced by treatments in the present study.Rodney et al. (2018b) observed a positive correlation between 25(OH)D 3 and the IGF-1, and also between the IGF-1 and glucose; therefore, this could explain, in part, the higher milk yield for the cows fed 25(OH)D 3 in the current study.Unfortunately, the IGF-1 was not measured in the present study.Additionally, Lee et al. (2007) demonstrated in mice models an endocrine regulation of energy metabolism.In mouse models, IGF-1 regulates the levels of nutrients in the blood and drives cell growth and division (Feng and Levine, 2010).
The supplementation with 25(OH)D 3 increased the energy secreted as milk (+1.8 Mcal/d), although the milk composition was not affected by treatments.Due to higher energy secreted as milk, the cows fed 25(OH) D 3 tended to have a higher NEB in the present study.Likewise, Martinez et al. (2018a) reported an increase of energy secreted in milk produced by cows fed an acidogenic diet with 25(OH)D 3 21 d before calving; therefore, such authors also reported a lower NEB (−7.5 Mcal/d), but higher than the results of the present study (−2.3 Mcal/d).The lower energy balance in the present study compared with Martinez et al. (2018a) was probably due to the higher DMI (19.4 vs. 18.4 kg/d) and lower milk fat concentration (3.59 vs. 4.77%).
The BW change and BCS were not affected by treatments.Due to the higher milk yield for the cows fed 25(OH)D 3 , a greater fat mobilization was expected to compensate for the higher energy demand for production; however, 25(OH)D 3 supplementation decreased fat mobilization and showed an improvement of homeorhetic mechanisms, and cows fed 25(OH)D 3 had a decreased NEB when compared with cows fed the CTRL.The β-cell proliferation, normal insulin secretion, and decrease of insulin resistance through the uncarboxylated osteocalcin promote the suppression mechanisms of reserved mobilization (Wolf, 2008).Furthermore, the active metabolite of vitamin D (calcitriol) induces the osteocalcin by the osteoblasts (Wolf and Phil, 1996).
Several studies have shown the influence of the near parturition on rumination time (Calamari et al., 2014;Goff et al., 2014;Kaufman et al., 2016) and the reaction to physiological and metabolic changes in relation to the low blood Ca concentration (Goff et al., 2020) and strong reduction of the DMI (Liboreiro et al., 2015).In the present study, both treatments had an abrupt decrease in DMI at parturition.Nevertheless, rumination time of cows fed 25(OH)D 3 decreased on calving day, but the cows fed vitamin D 3 had lower rumination time from parturition until the third week postpartum.Rumination time is closely linked with ruminal motility, which is also controlled by the Ca availability and can affect DMI (Jørgensen et al., 1998).The ruminal contraction usually ceases when the iCa is lower than 0.6 mM (Jørgensen et al., 1998).Additionally, a direct relationship exists between plasma Ca concentration and rumination rate (r = 0.75; P = 0.01) on the first day of lactation (Goff et al., 2020).Our data suggest that the trend for higher serum iCa postpartum on cows feeding 25(OH)D 3 was enough for increased rumination time after parturition.
The increased prepartum Ca excretion (+4.82 g/d) in the current study suggests a compensatory mechanism (Kovács et al., 2015) on cows fed 25(OH)D 3 .The normal levels of blood Ca are tightly regulated, and the higher concentration of blood Ca results in a hypercalciuria status (Constable et al., 2019) that can lead to a higher urinary Ca excretion; however, when blood Ca decreases, the renal reabsorption increases to maintain the normal levels of blood Ca (Blaine et al., 2015).The cows fed 25(OH)D 3 in the current study possibly have increased Ca transport by the increase of the expression of luminal Ca channels mainly by the transient receptor potential vanilloid type 5 and 6 (Hoenderop et al., 2005).This increase in the Ca flux exchange promotes a decrease in renal Ca reabsorption (Fredeen et al., 1988) and, therefore, more Ca is excreted, as observed in the current study.In the same way, the urinary Ca excretion increased (+4.10 g/d) for cows fed an acidogenic diet (−68 mEq/kg of DM) supplemented with 25(OH)D 3 (Oehlschlaeger et al., 2014).
Due to the higher urinary excretion of Ca in cows fed 25(OH)D 3 (8.01 vs. 12.83 g/d for the cows fed the CTRL treatment and cows fed 25(OH)D 3 , respectively), the Ca balance (even positive) was lower during the prepartum period and negative during the postpartum period.Nevertheless, the estimation of Ca balance did not account for a likely effect of treatment on Ca absorption; therefore, Ca balance may be underestimated in the current study.In addition, cows fed 25(OH)D 3 produced 2.06 kg/d more colostrum, which resulted in 7.0 g/d more Ca secreted in the colostrum.Although  TO DAIRY COWS this increase was not statistically significant, it is possible the Ca secreted in colostrum affected the Ca balance.In agreement with the present study, Rodney et al. (2018a) observed similar values for the Ca balance, but the negative balance in that study was smaller than the observed in the current trial, probably due to the lack of the effect of treatment on colostrum yield observed by Rodney et al. (2018a).In the present study, the Mg balance was not influenced by treatments, probably due to the lack of effect of treatment on urinary Mg excretion and Mg concentration in the colostrum.In contrast, according to Rodney et al., (2018a), feeding cows 25(OH)D 3 during the transition period increased the negative balance of Mg.

CONCLUSIONS
Cows fed 25(OH)D 3 at 3 mg/d tended to have an increased serum iCa concentration and increase urinary excretion of Ca.Also, feeding 25(OH)D 3 at 3 mg/d increases colostrum yield and plasma glucose concentration on the day of calving.Feeding 25(OH)D 3 increases milk yield, 3.5% FCM, and ECM, and improves milk yield components in early lactation.Overall, these findings suggest that 25(OH)D 3 at 3 mg/d can improve the energy metabolism and lactation performance compared with the current feeding practice of supplementing vitamin D 3 .

Figure 2 .
Figure 2. Plasma concentration of glucose (A), insulin (B), nonesterified fatty acids (NEFA; C), and BHB (D) of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3 .Acidogenic prepartum diets containing the treatments were fed 30 d before the expected calving date.CTRL: vitamin D 3 at 0.625 mg/d, equivalent to 25,000 IU of vitamin D 3 /d (Rovimix D 3 500, DSM Nutritional Products) or 25-hydroxyvitamin D 3 [25(OH)D 3 ] at 3 mg/d (Rovimix Hy-D 1.25%, DSM Nutritional Products).Panel A:effect of treatment (P = 0.0017), effect of the day (P < 0.001), and interaction between treatment and day (P = 0.0036).Panel B: effect of treatment (P = 0.43), effect of day (P < 0.004), and interaction between treatment and the day (P = 0.52).Panel C: treatment effect (P = 0.0001), effect of day (P < 0.0001), and interaction between treatment and day of collection (P = 0.0001).Panel D: effect of treatment (P = 0.36), effect of day (P < 0.10), and interaction between treatment and day (P = 0.76).Means with different letters (a,b) are different (P < 0.05).Error bars denote SEM; asterisks denote interaction between vitamin D source and day of collection.

Figure 5 .
Figure 5. Rumination time of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with control (CTRL) or 25(OH)D 3 .Acidogenic prepartum diets containing the treatments were fed 30 d before the expected calving date.CTRL: vitamin D 3 at 0.625 mg/d, equivalent to 25,000 IU of vitamin D 3 /d (Rovimix D 3 500, DSM Nutritional Products) or 25-hydroxyvitamin D 3 [25(OH)D 3 ] at 3 mg/d (Rovimix Hy-D 1.25%, DSM Nutritional Products).Error bars denote SEM; asterisks denote interaction between vitamin D source and week of lactation.

Table 1 .
TO DAIRY COWS Ingredient and chemical composition of the experimental diets

Table 2 .
Dry matter intake, energy balance, BW, and BCS of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3

Table 3 .
Colostrum yield and Ca and Mg secretion in colostrum of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3 1 Acidogenic prepartum diets containing the treatments were fed 30 d before the expected calving date.CTRL: vitamin D 3 at 0.625 mg/d, equivalent to 25,000 IU of vitamin D 3 /d (Rovimix D 3 500, DSM Nutritional Products) or 25-hydroxyvitamin D 3 [25(OH)D 3 ] at 3 mg/d (Rovimix Hy-D 1.25%, DSM Nutritional Products).

Table 7 .
TO DAIRY COWS Postpartum DMI, net energy (NE) secreted as milk, caloric intake, net energy balance, BW, BW change, BCS, and rumination of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3

Table 8 .
Silva et al.: FEEDING 25-HYDROXYVITAMIN D 3 TO DAIRY COWS Urinary volume, creatinine, Ca, and Mg of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3 Acidogenic prepartum diets containing the treatments were fed 30 d before the expected calving date.CTRL: vitamin D 3 at 0.625 mg/d, equivalent to 25,000 IU of vitamin D 3 /d (Rovimix D 3 500, DSM Nutritional Products) or 25-hydroxyvitamin D 3 [25(OH)D 3 ] at 3 mg/d (Rovimix Hy-D 1.25%, DSM Nutritional Products). 2Urinary production and mineral excretion were performed at d 21 and 42 relative to calving.

Table 9 .
Mineral balance during prepartum and at the day of calving of multiparous pregnant Holstein dairy cows fed a prepartum acidogenic diet supplemented with CTRL or 25(OH)D 3