If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Objectives were to determine the effect of supplementing 2 sources of vitamin D, cholecalciferol (CH) or calcidiol (CA), at 1 (1mg) or 3 mg/d (3mg) prepartum on concentrations of vitamin D metabolites in plasma, measures of innate immune function, and leukocyte mRNA expression. Parous Holstein cows (n = 99) were assigned to a daily treatment administered as top-dress containing either 1 or 3 mg of CH (CH1 or CH3) or of CA (CA1 or CA3) from 250 d of gestation until calving. Plasma concentrations of vitamin D, immune cell population in blood, cell adhesion markers, and granulocyte phagocytosis and oxidative burst were evaluated pre- and postpartum. The mRNA expression in leukocytes was determined at 270 d of gestation and 3 d postpartum for genes involved in cell migration, pathogen recognition receptors, cell signaling, cytokines, antimicrobial mechanisms, oxidative burst, and Ca and vitamin D metabolism. Concentrations of vitamin D3 increased in cows fed CH, whereas those of 25-hydroxyvitamin D3 increased in cows fed CA. Percentage of granulocytes from total leukocytes differed with amount of vitamin D pre- (1mg = 24.5 vs. 3mg = 37.9%) and postpartum (1mg = 22.0 vs. 3mg = 31.0%), thus shifting mononuclear cells in the opposite direction pre- (1mg = 75.5 vs. 3mg = 62.1%) and postpartum (1mg = 78.0 vs. 3mg = 69.0%). Granulocytes displaying phagocytosis (1mg = 69.0 vs. 3mg = 62.9%) and intensity of phagocytosis prepartum (1mg = 7.46 vs. 3mg = 7.28) tended to be less in cows fed 3mg compared with 1mg. During prepartum, CA increased mRNA expression of genes related to cell adhesion and migration (CD44, ICAM1, ITGAL, ITGB1, LGALS8, SELL), pathogen recognition receptor (NOD2, TLR2, TLR6), cell signaling (FOS, JUN, NFKB2), cytokine signaling (IL1B, IL1R1, IL1RN), antimicrobial mechanisms (CTSB, LYZ), and Ca metabolism (ATP2B1, STIM1, TRPV5) compared with CH. Similarly, postpartum, CA increased mRNA expression of genes related to cell adhesion and migration (CXCR2, SELL, TLN1), cell signaling (AKT2), cytokines (CCL2, IL1R1, ILRN), antimicrobial mechanisms (DEFB3), oxidative burst (RAC2), and calcium metabolism (CALM3) compared with CH. Feeding additional vitamin D in the last 3 wk of gestation changed the profile of blood leukocytes and attenuated granulocyte phagocytosis during the transition period, whereas supplementing CA prepartum increased mRNA expression of genes involved in immune cell function, including genes related to pathogen recognition and antimicrobial effects of leukocytes.
). Vitamin D3, also known as cholecalciferol (CH), can be endogenously synthesized in the skin when the latter is exposed to sunlight, or it can be acquired from the diet. In dairy cattle, supplementation with CH is a common practice (
) and, once absorbed, CH is hydroxylated by 25-hydroxylase in the liver to 25-hydroxyvitamin D3, also known as calcidiol (CA), the more stable metabolite of vitamin D. The most abundant circulating form of vitamin D is 25-hydroxyvitamin D3 and therefore is used as an indicator of the vitamin status of the animal (
). In the kidney, 25-hydroxyvitamin D3 is further hydroxylated by 1α-hydroxylase to form 1α,25-dihydroxyvitamin D3, also known as calcitriol, the steroid hormone that is the most active metabolite of vitamin D. This final conversion is a tightly regulated process (
). Inactivation of 25-hydroxyvitamin D3 and 1,25-dihydroxyvitamin D3 is carried out by 24-hydroxylase, producing 24,25-dihydroxyvitamin D3.
Plasma concentrations of vitamin D metabolites are responsive to the amount of dietary supplementation, although increasing serum 25-hydroxyvitamin D3 can be more easily achieved by supplementing CA than CH (
recommendations, supplementing CH did not increase concentrations of 25-hydroxyvitamin D3 in plasma above 100 ng/mL, suggesting a regulatory control on the conversion of vitamin D3 into 25-hydroxyvitamin D3 by the hepatic 25-hydroxylase (
). Therefore, it is possible that the effects of vitamin D supplementation may be enhanced by supplementing CA compared with CH. It is clear that adequate vitamin D status, based on concentrations of 25-hydroxyvitamin D3 in plasma (
demonstrated that pathogen-associated molecules such as LPS and peptidoglycan can stimulate local expression of 1α-hydroxylase in peripheral blood monocytes in dairy cows. In addition, intramammary endotoxin challenge in dairy cows resulted in activation of the vitamin D signaling pathway in mammary gland macrophages and neutrophils (
). Thus, it is possible that supplementation with CA could increase substrate availability for local synthesis of calcitriol in immune cells, which might benefit immune response. In support of that, calcitriol has been shown to increase β-defensins and inducible nitric oxide synthase mRNA in milk somatic cells (
). Interestingly, supplementation of dietary CA prepartum resulted in increased concentrations of CA in plasma in dairy cows pre- and postpartum and reduced incidence of retained placenta and metritis (
). It is possible that CA influences leukocyte mRNA expression for key regulatory proteins involved in vitamin D metabolism and immune cell functions important for pathogen clearance.
The hypothesis was that supplementing CA would increase concentrations of 25-hydroxyvitamin D3 in plasma and enhance measures of innate immune cell function by altering mRNA expression and intensity of expression of surface markers linked with innate immune response in peripheral blood leukocytes, compared with CH. It was anticipated that increasing the amount of vitamin D would result in increased effects when the source was CA rather than CH. Therefore, the objective of this experiment was to determine the effects of 2 sources of vitamin D3, CH or CA, supplemented at 2 amounts, 1 or 3 mg/d, during the last 3 wk of gestation on granulocyte function and peripheral blood leukocyte mRNA expression in dairy cows.
MATERIALS AND METHODS
The experiment was conducted at the University of Florida (Gainesville) Dairy Research Unit. Procedures for animal handling and care were approved by the University of Florida Institutional Animal Care and Use Committee under the protocol number 20171002. These data are part of an experiment evaluating the effects of vitamin D on performance and health that included 170 transition Holstein cows (
The experiment was a randomized complete block design with a 2 × 2 factorial arrangement of treatments and cow as the experimental unit. A 2-sided sample size was calculated for 4 treatments using α = 0.05 and β = 0.20 to allow detection of a 1-unit difference in delta cycle threshold (dCt) in mRNA expression, assuming that the standard deviation (SD) for dCt would be a maximum of 1.0 unit. The sample size calculated was 23 experimental units per treatment. Additional cows were enrolled to account for potential attrition at calving. Because of the arrangement of treatments as a 2 × 2 factorial, the interaction term would result in more than 95% power to detect a difference in mRNA expression, assuming a 1-unit difference in dCt and an SD for dCt of 1.0 unit.
Cows, Housing, Treatments, and Diets
A total of 99 parous Holstein cows were moved to the experimental facilities at 242 d of gestation, and dietary treatments started at exactly 250 d of gestation of each cow and ended at calving. Prepartum cows were housed in a freestall barn equipped with individual feeding gates (Calan Broadbent feeding system, American Calan Inc.) that were assigned randomly to each cow. Individual cow DMI was measured daily prepartum. Postpartum, cows were housed in a single group in a freestall barn for the first 42 d in lactation.
Every 2 wk, a cohort of 8 to 12 parous cows at 240 d of gestation were blocked based on parity group (1 vs. >1) and then by 305-d milk yield in the recently completed lactation and, within each block, assigned randomly to 1 of 4 treatments. Treatments included 2 sources of vitamin D, CH (Rovimix D3, 12.5 g of CH per kg; Division of Animal Nutrition and Health, DSM Nutritional Products LLC) or CA (Hy-D, 12.5 g of CA per kg; DSM Nutritional Products LLC), supplemented at 2 amounts, 1 or 3 mg/d arranged as factorial, resulting in 4 treatments: CH at 1 mg/d (CH1, n = 21), CH at 3 mg/d (CH3, n = 27), CA at 1 mg/d (CA1, n = 25), and CA at 3 mg/d (CA3, n = 26). Treatment CH1 represents what is typically used as vitamin D feeding to dairy cows in commercial farms (
for prepartum cows. The different numbers of experimental units per treatment was caused by some cows calving before the expected date, and, therefore, a prepartum sample was not obtained, resulting in incomplete blocks. Treatments were initiated at 250 d of gestation, and vitamin D supplements were diluted into 100 g of corn meal to result in the desired doses of CH (1 or 3 mg) and CA (1 or 3 mg). Cows were fed the same prepartum diet offered as TMR once daily, at 0700 h, which was formulated to meet the nutrient needs of late-gestation Holstein cows and formulated to have a negative value for DCAD (Table 1). The vitamin D treatments were administered as top-dress dispensed onto the TMR once daily until calving. Postpartum, cows were fed the same diet containing 1 mg of CH for every 25 kg of DM.
Table 1Ingredient composition and nutrient content of the prepartum diet
Contained (DM basis) 52.2% wheat middlings, 18.7% Reashure (Balchem), 16.0% magnesium sulfate heptahydrate, 4.8% magnesium oxide, 4.5% sodium chloride, 0.685% of a mixture containing vitamins A, D, and E and iodine, 0.45% Sel-Plex 2000 (Alltech Biotechnology), 0.40% Rumensin 90 (Elanco Animal Health), 0.27% IntelliBond Vital4 (Micronutrients), and 2.0% ClariFly Larvicide (Central Life Sciences). Each kilogram contained 8.8% CP, 0.20% Ca, 0.59% P, 4.4% Mg, 0.5% K, 1.8% Na, 2.7% Cl, 2.2% S, 740 mg of Zn, 168 mg of Cu, 562 mg of Mn, 9 mg of Se, 16.2 mg of Co, 12 mg of I, 103,000 IU of vitamin A, 18,160 IU of vitamin D, 1,475 IU of vitamin E, and 725 mg of monensin.
4 Contained (DM basis) 52.2% wheat middlings, 18.7% Reashure (Balchem), 16.0% magnesium sulfate heptahydrate, 4.8% magnesium oxide, 4.5% sodium chloride, 0.685% of a mixture containing vitamins A, D, and E and iodine, 0.45% Sel-Plex 2000 (Alltech Biotechnology), 0.40% Rumensin 90 (Elanco Animal Health), 0.27% IntelliBond Vital4 (Micronutrients), and 2.0% ClariFly Larvicide (Central Life Sciences). Each kilogram contained 8.8% CP, 0.20% Ca, 0.59% P, 4.4% Mg, 0.5% K, 1.8% Na, 2.7% Cl, 2.2% S, 740 mg of Zn, 168 mg of Cu, 562 mg of Mn, 9 mg of Se, 16.2 mg of Co, 12 mg of I, 103,000 IU of vitamin A, 18,160 IU of vitamin D, 1,475 IU of vitamin E, and 725 mg of monensin.
Blood was sampled by puncture of the coccygeal vessels into 10-mL evacuated tubes containing K2 EDTA (Vacutainer, Becton Dickinson). A sample was collected the day before cows initiated the dietary treatments and again on d 260, 267, 273, and 275 of gestation, which corresponded to the mean days relative to calving (range) −19 (−23 to −14), −10 (−13 to −4), −3 (−5 to −2) and −1 (−1). Postpartum, blood was sampled exactly on d 0, 4, and 8. Concentrations of vitamin D3, 25-hydroxyvitamin D3, and 24,25-dihydroxyvitamin D3 in plasma were analyzed using ultra high-performance liquid chromatography (1290 Infinity II LC System, Agilent) coupled with tandem MS detection (API 4000 LC-MS/MS System, AB Sciex LLC) by the DSM Nutritional Products Research and Development Solution Center (Kaiseraugst, Switzerland). In all assays, dedicated standard and quality-control samples were analyzed concurrent with unknown samples to ensure the accuracy and precision of the method. Data acquisition, integration, and quantification were performed using Analyst 1.7 software (AB Sciex LLC). Personnel running the assays were blind to treatments, and the lower limits of quantification for the assays were 0.5 ng/mL for vitamin D3 and 24,25-hydroxyvitamin D3, and 1.0 ng/mL for 25-hydroxyvitamin D3. The intra-assay coefficients of variation (CV) were 4.6, 3.6, and 6.2%, respectively, for vitamin D3, 25-hydroxyvitamin D3, and 24,25-dihydroxyvitamin D3.
Blood Minerals and Metabolites
Blood was sampled by puncture of the coccygeal vessels into 10-mL serum separator evacuated tubes (Vacutainer, Becton Dickinson) on gestation d 250, before treatments initiated, and again on d 260, 263, 266, and 269 of gestation, and then daily until calving. Postpartum blood was sampled on the day of calving and again on d 1 to 5 postpartum. Samples were collected before the morning feeding at 0600 h. Tubes were placed on ice, transported to the laboratory within 1 h, and centrifuged for 20 min at 2,500 × g for serum separation. Serum was frozen and stored at −20°C for further analyses. Prepartum serum samples collected from d −9 to −1 relative to calving and daily from d 0 to 5 postpartum were assayed for concentrations of total Ca, Mg, and P. Serum was analyzed for concentrations of Ca (Arsenazo III method, cat. no. CA2390), Mg (xylidyl-blue method, cat. no. MG531), and P (phosphomolybdate method, cat. no. PH8328) using a biochemical analyzer (RX Daytona, Randox Laboratories Ltd.) according to the manufacturer's instructions. Intra- and interassay CV were, respectively, 0.5% and 8.1% for Ca, 4.1% and 10.2% for Mg, and 1.1% and 9.8% for P.
Peripheral Blood Leukocyte Population Markers
Leukocytes were isolated from coccygeal blood sampled into 10-mL lithium-heparinized tubes (Vacutainer, Becton Dickinson) prepartum at 270 d of gestation, and on d 3 and 6 postpartum. Erythrocytes were lysed from an aliquot containing 100 μL of whole blood by adding 2 mL of a lysing solution (10.6 mM Na2HPO4, 2.7 mM NaH2PO4, pH = 7.2) and keeping tubes on ice for 2 min. Then 1 mL of a resuspension solution (10.6 mM Na2HPO4, 2.7 mM NaH2PO4, 462 mM NaCl, pH = 7.2) was added, and samples were homogenized and centrifuged at 500 × g for 5 min at 4°C. The supernatant was removed, and a small pellet of cells remained in the bottom of the tube. Cells were incubated for 20 min on ice with allophycocyanin (APC, Bio-Rad Laboratories)-conjugated anti-CD11b (CC126 clone, Bio-Rad Laboratories), fluorescein isothiocyanate (FITC)-conjugated anti-CD21 (CC21 clone, Bio-Rad Laboratories), phycoerythrin (PE, Bio-Rad Laboratories)-conjugated anti-CD62L (BAQ92A clone, Kingfisher Biotech Inc.), and phycoerythrin and cyanine dye (PE-Cy5.5) anti-CD14 (Tük4 clone, Life Technologies). An antibody cocktail solution was prepared containing 25 μL of each antibody for a total of 100 μL (1 μg/μL), which was sufficient for 100 samples, and stored at 4°C in the dark. On the day of the assay, 1 μL of antibody cocktail was diluted into 25 μL of flow staining buffer (PBS with 5% heat-inactivated fetal bovine serum and 0.1% NaN3) in a microtube kept in the dark and added to each pellet of cells isolated from a cow and incubated on ice for 15 min. Cells were washed with 2 mL of flow staining buffer to remove excess of antibodies not bound and then centrifuged at 500 × g for 5 min at 4°C. Cells were resuspended in 500 μL of cold PBS at 4°C and kept on ice in the dark until analysis.
Cell populations were then analyzed by flow cytometry (Accuri C6 digital analyzer flow cytometer, BD Biosciences) according to forward and side scatters to segregate cells based on size and granularity, and then identify those expressing CD14, CD21, CD11b, and CD62L. Data acquisitions from 10 μL of resuspended cells per sample (16,924 ± 13,881 cells) were analyzed using FlowJo software (version 10.6.2, FlowJo LLC). The parameters analyzed included the percentage of granulocytes and mononuclear cells from the total leukocyte population, and percentages of monocytes (CD14+), B lymphocytes (CD14− CD21+), and CD14− CD21− cells from the population of mononuclear cells (Supplemental Figure S1A; https://figshare.com/s/f19a05c4324597477f6e). We were limited to 4 parameters in the flow cytometry panel, and these markers were selected to most effectively capture the profile and status of circulating leukocytes. After identification of the distinct cell populations, the mean fluorescence intensity (MFI) for specific cell surface markers was quantified. In granulocytes, MFI was quantified for L-selectin (CD62L) and a β-integrin (CD11b); in monocytes, MFI was quantified for CD14, CD62L, and CD11b; in B lymphocytes, the MFI was quantified for CD62L and CD11b; and in CD14− CD21− cells, the MFI was quantified for CD62L and CD11b.
Blood was sampled by puncture of the coccygeal vessels into 10-mL lithium-heparinized tubes (Vacutainer, Becton Dickinson) at 270 d of gestation and on d 3 and 6 postpartum. Tubes were kept at room temperature, wrapped in foil to protect from light, and transported to the laboratory within 3 h of collection. Granulocyte function was assayed in vitro by dual-color flow cytometry according to
with a few modifications. Quantification of granulocyte phagocytosis and intracellular killing was assessed using an Escherichia coli 08:H19 strain KCJ852 previously isolated from a cow with metritis, labeled with propidium iodide. Briefly, 100 μL of blood was pipetted in triplicate aliquots, and 10 μL of 5 μM dihydrorhodamine 123 was added to each aliquot. One aliquot was used as a negative control, containing only blood and 5 μM dihydrorhodamine 123. The other 2 aliquots each received 40 μL of a solution containing 5.0 × 105 propidium iodide-labeled E. coli per μL. Samples were analyzed by flow cytometry (BD Biosciences). Data acquisitions from at least 10,000 and up to 100,000 cells per sample were analyzed using FlowJo software. Cells were then analyzed according to forward and side scatter gating, and responses analyzed included percentage of granulocytes that phagocytized bacteria, percentage of granulocytes with phagocytosis-induced oxidative burst, MFI for phagocytosis (an indication of the number of bacteria phagocytized per granulocyte), and MFI of oxidized dihydrorhodamine to estimate intensity of oxidative burst as an indication of reactive oxygen species produced per cell.
Leukocyte RNA Extraction and Gene Expression
Blood leukocytes were isolated from samples collected at 270 d of gestation and on d 3 postpartum. Blood was sampled by puncture of the coccygeal vessels into 10-mL lithium-heparinized tubes (Vacutainer, Becton Dickinson). Samples were centrifuged for 20 min at 2,500 × g for plasma separation, and the buffy coat fraction was collected by pipetting and transferring to a 12-mL conical tube. Red blood cell lysis buffer (Cell Lysis Solution, Promega Corp.) was added for a total volume of 12 mL. Tubes were homogenized and incubated at room temperature for 5 min. Samples were then centrifuged for 5 min at 500 × g and supernatant discarded. This step was repeated multiple times to remove all red blood cells. The leukocyte pellets were resuspended with 0.8 mL of Trizol (TRIzol LS Reagent, Invitrogen), transferred to 1.5-mL microtubes, and stored at −80°C until further analysis. For RNA extraction, chloroform was added to the microtube to bring up to a 20% chloroform solution. Samples were then centrifuged at 12,000 × g for 15 min at 4°C, and the supernatant containing the RNA was transferred to a new microtube. Purification of RNA was performed using the PureLink RNA Mini Kit (Invitrogen) according to the manufacturer's instruction. Purity and concentration were evaluated using a NanoDrop 2000 spectrophotometer (Thermo Scientific). Samples had a mean (±SD) 260:280 nm ratio of 2.00 ± 0.06, and 260:230 nm ratio of 1.75 ± 0.39.
The mRNA for a selected set of genes was quantified by the Fluidigm quantitative PCR microfluidic device Biomark HD system (Fluidigm Co.). The PCR primers were designed by Fluidigm Delta Gene assays and synthesized by Fluidigm (Fluidigm Co.). Details of genes and primers are in Supplemental Table S1 (https://figshare.com/s/f19a05c4324597477f6e). A pool sample containing mRNA from bovine leukocytes from 10 different samples was used for primer validation. Primers were validated using Fluidigm primer quality-control criteria (R2 ≥ 0.97; efficiency of 80 to 130%; slope = −3.92 to −2.76) were applied to cDNA serially diluted by 12× and evaluated in 8 replicates. Primers targeting 8 genes (CASR, CYP24A1, GC, IL12A, MMP2, NOS2, SELE, and TRPV6) failed to pass the quality control in the qualification run and were excluded from analyses. Primer efficiency in the genes passing quality control and used in this experiment ranged between 101.3 and 120% (Supplemental Table S1). Genes investigated included those involved in cell adhesion and migration, pathogen recognition receptor, cell signaling, synthesis of cytokines, antimicrobial mechanisms, oxidative burst, Ca metabolism, and vitamin D metabolism. Statistical analyses were performed on dCt values as described by
, whereby fold changes were calculated from least squares means (LSM) difference according to the formula 2−ddCt, where dCt = CtTarget gene − geometric mean of CtReference genes, and ddCt = dCttreatment A − dCttreatment B. Reference genes used were ACTB, GAPDH, RPL19, RPS9, and YWHAZ. Heatmaps were generated using Heatmapper online tool (
Distribution of residuals and homogeneity of variance were evaluated for all continuous dependent variables after fitting the statistical models. Responses that violated the assumptions of normality were subjected to power transformation according to the Box-Cox procedure (
. Concentrations in plasma of vitamin D3, 25-hydroxyvitamin D3, and 24,25-dihydroxyvitamin D3 had to be log-transformed before analysis either because of heteroscedasticity or because residuals were not normally distributed. The LSM and standard error of the mean (SEM) were back-transformed for data presentation according to
Data were analyzed by linear mixed-effects models using the MIXED procedure of SAS, and analyses were performed for the pre- and postpartum periods separately. The statistical models included the fixed effects of source of vitamin D (CH vs. CA), amount of vitamin D [1 mg/d (1mg) vs. 3 mg/d (3mg)], interaction between source and amount (CH1 + CA3 vs. CH3 + CA1), the linear covariate of the exact day relative to calving when the measurement was taken, and the random effect of block. For analysis of mRNA expression, the fixed effect of assay plate (1 or 2) was also included in the models. Data with repeated measures within cow were analyzed with the same mixed-effects model described previously but also included the fixed effects of day and the interactions between source of vitamin D and day, amount of vitamin D and day, and source and amount and day, as well as the random effect of cow nested within source and amount of vitamin D. For concentrations of vitamin D metabolites in plasma, the concentration obtained in the sample analyzed before initiation of treatments was used as covariate in the statistical models. The Repeated statement was included in all mixed models with repeated measurements, with day specified as the repeated effect. The covariance structure was selected based on spacing between measurements and model fit that resulted in the smallest corrected Akaike's information criterion. The Kenward-Roger method was used to approximate denominator degrees of freedom to compute the F tests. When an interaction between amount and source of supplemental vitamin D was significant, means were portioned using the SLICE command in SAS, and pairwise comparisons were conducted by the Fisher's least significant difference. Evidence of statistical significance against the null hypothesis was considered at P ≤ 0.05, and tendency was considered at 0.05 < P ≤ 0.10. For gene expression, only comparisons with P < 0.05 were considered.
Prepartum Vitamin Intake and Blood Concentrations of Vitamins and Minerals
The DMI prepartum did not differ with source or amount of vitamin D and averaged 11.3 kg/d. The diet offered contained some CH in the mineral-vitamin premix, which resulted in a mean prepartum intake of CH from the diet of 0.22 mg/d in all treatments. Therefore, cows consumed daily 1.22 mg of CH in CH1, 3.22 mg of CH in CH3, 0.22 mg of CH and 1 mg of CA in CA1, and 0.22 mg of CH and 3 mg of CA in CA3. The average duration of treatment prepartum did not differ among treatments, and means were 24.2 (9 to 32), 25.7 (13 to 33), 25.1 (14 to 32), and 25.5 (18 to 30) d, respectively, for CH1, CH3, CA1, and CA3. Six cows developed milk fever, 2 out of 21 in CH1 (9.5%), 1 out of 27 in CH3 (3.7%), 1 out 25 in CA1 (4.0%), and 2 out of 26 in CA3 (7.7%).
Of the 99 cows enrolled, 25 cows with 66 samples had concentration of vitamin D3 in plasma below the detection limit of the assay (<0.50 ng/mL). Of those 25 cows, 3 cows were fed CH, 2 in CH1 and in 1 CH3, and 22 cows were fed CA, 12 in CA1 and 10 in CA3. Details of the number of samples, day relative to calving, and number of cows with concentrations below the detection limit for vitamin D3 in plasma are presented in Supplemental Table S2 (https://figshare.com/s/f19a05c4324597477f6e). All other measurements of vitamin D3, 25-hydroxyvitamin D3, and 24,25-dihydroxyvitamin D3 resulted in concentrations above the detection limits of the assays.
All means and measures of variance reported in the text are LSM and respective SEM, unless otherwise stated. Concentrations of vitamin D3 in plasma prepartum decreased (P < 0.001) as parturition approached, in particular in cows fed CH3 (Figure 1A). Concentrations decreased (P < 0.001) over time postpartum, and the decline was observed only in cows fed CH based on the interaction (P < 0.001) between source and day relative to calving. However, feeding CA increased (P < 0.001) concentrations of 25-hydroxyvitamin D3 in plasma, and the increment over time was greater in cows fed CA3 than CA1 (Figure 1B). Concentrations of 25-hydroxyvitamin D3 in plasma declined (P < 0.001) in the first 8 d postpartum, and an interaction (P = 0.04) among source and amount and day was observed because the decrease in concentrations was greater in cows fed CA3 than in all other treatments. An interaction (P < 0.001) among source and amount and day was observed for prepartum concentrations of 24,25-hydroxyvitamin D3 in plasma because of the greater increase in concentrations in CA3 compared with all other treatments (Figure 1C). Concentrations of 24,25-hydroxyvitamin D3 in plasma remained elevated postpartum in cows fed CA3, followed by CA1, and then those fed CH.
Cows fed CA had greater (P = 0.03) concentrations of P in serum prepartum than those fed CH (CH = 1.84 vs. CA = 1.93 ± 0.03 mM; Table 2). Nevertheless, treatment or the interaction between treatment and day did not affect the concentrations of Ca or Mg in serum prepartum or any of the minerals postpartum.
Table 2Serum concentrations of minerals pre- and postpartum (LSM ± SEM)
Uterine diseases, defined as the combination of retained placenta and metritis, affected 22.2% (22/99) of the cows in this experiment; however, no difference between treatments was detectable (CH1 = 28.4; CH3 = 18.3; CA1 = 20.5; CA3 = 22.8 ± 8.6%). Mastitis affected 7.1% (7/99) of the cows and, similar to uterine diseases, was not affected by source or amount of vitamin D (CH1 = 3.8; CH3 = 13.3; CA1 = 3.5; CA3 = 3.3 ± 4.5%). During the first 60 d postpartum, 36.4% (36/99) had at least 1 disease event; however, no difference was detectable between treatments (CH1 = 47.5; CH3 = 36.9; CA1 = 29.1; CA3 = 34.5 ± 9.8%).
Profile of Immune Cells in Blood
Feeding the 3mg treatments, irrespective of source of vitamin D, increased (P = 0.008) the proportion of granulocytes (1mg = 24.5 vs. 3mg = 37.9 ± 3.5%) but consequently decreased that of mononuclear cells (1mg = 75.5 vs. 3mg = 62.1 ± 3.5%) in blood prepartum (Figures 2A and 2B). Source of vitamin D did not affect the proportions of different leukocytes in blood. Similar responses were observed during the early postpartum period. Feeding the 3mg treatments increased (P = 0.02) the proportion of granulocytes (1mg = 22.0 vs. 3mg = 31.0 ± 3.0%) and decreased that of mononuclear cells (1mg = 78.0 vs. 3mg = 69.0 ± 3.0%) in blood compared with feeding 1mg treatments (Figures 2A and 2B). Cows fed CA tended (P = 0.08) to have a smaller proportion of granulocytes (CH = 29.8 vs. CA = 23.3 ± 3.0%) but a greater proportion of mononuclear cells (CH = 70.2 vs. CA = 76.7 ± 3.0%) in blood than cows supplemented with CH. No interaction was observed between source and amount of vitamin D pre- or postpartum for the proportion of different leukocytes in blood.
Regarding mononuclear cells, feeding CH3 increased (P < 0.05) the proportion of monocytes pre- and postpartum (Figure 2C). Consequently, cows fed CH3 had the smallest proportions of B lymphocytes and CD14− CD21− cells (Figure 2D). Cows fed 3 mg of either source of vitamin D had the smallest proportions of CD14− CD21− cells pre- and postpartum (Figure 2E).
Feeding CA decreased (P = 0.05) the percentage of granulocytes expressing CD62L compared with CH prepartum (Table 3). In addition, MFI for CD62L was also less (P = 0.03) for granulocytes from cows fed CA compared with CH. Treatment did not affect the MFI for CD11b on granulocytes. On monocytes, MFI for CD14 was less (P = 0.02) in cows fed CA compared with CH, and those fed 3mg had greater (P = 0.02) MFI for CD14 compared with 1mg. Treatment did not affect the percentage of monocytes expressing CD62L and MFI for CD62L on those cells. Cows fed CA tended (P = 0.06) to have greater MFI for CD11b on monocytes than cows fed CH.
Table 3Effect of source and amount of vitamin D on blood cell markers prepartum (LSM ± SEM)
A tendency for interaction (P = 0.09) between source and amount of vitamin D was observed for the percentage of B lymphocytes expressing CD62L prepartum, because feeding CA1 tended to have greater percentage of cells expressing CD62L compared with those fed CA3. Also, MFI for CD62L in B lymphocytes increased (P < 0.05) with feeding CA1 compared with CA3. Treatment did not affect MFI for CD11b in B lymphocytes. For CD14− CD21− cells prepartum, feeding the 3mg treatments tended (P = 0.07) to increase the percentage of cells expressing CD62L compared with the 1mg treatments (1mg = 12.0 vs. 3mg = 18.4 ± 2.5%). Similarly, cows fed 3mg treatments tended (P = 0.09) to have increased MFI for CD62L in CD14− CD21− cells prepartum compared with those fed 1mg treatments (1mg = 2.35 vs. 3mg = 2.44 ± 0.04).
Treatment did not affect percentage of granulocytes expressing CD62L, or MFI for CD62L and CD11b in granulocytes postpartum (Table 4). In cows fed CH3, MFI for CD14 was increased (P < 0.05) compared with the other treatments. Percentage of monocytes expressing CD62L increased (P = 0.04) in cows fed CH compared with CA. Similarly, the MFI for CD62L increased (P = 0.02) in cows fed CH compared with CA. Within cows fed CA, 3mg reduced (P < 0.05) the MFI for CD11b in monocytes compared with 1mg, whereas no difference occurred among cows fed CH. In B lymphocytes, the percentage of cells expressing CD62L did not differ among treatments, whereas cows fed CH3 had greater (P < 0.05) MFI for CD62L compared with all other treatments. No difference was detectable between treatments for the MFI for CD11b in B lymphocytes. In CD14− CD21− cells, feeding 3mg increased (P < 0.05) the percentage of cells expressing CD62L (1mg = 11.0 vs. 3mg = 16.4 ± 2.2%) and MFI of CD62L (1mg = 2.35 vs. 3mg = 2.45 ± 0.04) but decreased (P < 0.05) the MFI of CD11b (1mg = 3.32 vs. 3mg = 3.19 ± 0.06) compared with those fed 1mg.
Table 4Effect of source and amount of vitamin D on blood cell markers postpartum (LSM ± SEM)
The proportion of granulocytes with phagocytic activity prepartum tended (P = 0.06) to be less in cows fed 3mg compared with 1mg (1mg = 69.0 vs. 3mg = 62.9 ± 3.0%), whereas treatment did not affect granulocyte phagocytosis postpartum (Figure 3A). Treatment did not affect the proportion of granulocytes with oxidative burst pre- or postpartum (Figure 3B). Intensity of phagocytosis prepartum tended to be smaller (P = 0.06) for cows fed 3mg compared with 1mg (1mg = 7.46 vs. 3mg = 7.28 ± 0.09).
No effect of treatment was observed for MFI of phagocytosis postpartum (Figure 3C). An interaction (P = 0.05) between source and amount of vitamin D was observed for MFI of oxidative burst prepartum, in part because of a tendency (P = 0.07) for CA1 to be greater than CH1 (Figure 2D). Treatment did not affect MFI for oxidative burst postpartum.
mRNA Expression in Blood Leukocytes Prepartum
A heatmap with the genes that were differentially expressed during the prepartum period was created using the fold-change relative to CH1 to illustrate the patterns according to treatment (Figure 4A). In addition, Supplemental Figure S2A depicts the pattern of mRNA expression according to the dCt for individual cows including only genes that were influenced (P < 0.05) by treatment (https://figshare.com/s/f19a05c4324597477f6e). In both Figure 4A and Supplemental Figure S2A, red represents reduced mRNA expression, whereas green represents increased mRNA expression. The genes differentially expressed (P < 0.05) by treatment during the prepartum period are reported in Table 5. Those included genes related to cell adhesion and migration, CD44, ICAM1, ITGAL, ITGB1, LGALS8, and SELL, which were upregulated (P < 0.05) by CA compared with CH. Expression of CXCR2, ITGAM, ITGB2, and TLN1 did not differ among treatments. Expression of genes related to receptors for pathogen recognition, including NOD2, TLR2, and TLR6, were upregulated (P < 0.05) in cows fed CA compared with CH. By contrast, expression of NOD1 was downregulated (P = 0.01) in cows fed CA compared with CH. Expression of TLR1, TLR4, and TLR9 did not differ among treatments. Expression of genes related to cell signaling cascade, including FOS, JUN, and NFKB2, were upregulated (P < 0.05) by CA compared with CH, whereas expression of AKT1, AKT2, IRAK4, MAPK1, MAPK3, MYD88, NFATC1, and NFKB1 did not differ among treatments. Calcidiol upregulated (P < 0.05) genes involved in cytokine signaling, including IL1B, IL1R1, and IL1RN, whereas expression of CCL2, CCL5, CXCL8, IFNG, IL1R2, IL6, IL10, IL23A, and TNF did not differ among treatments. Calcidiol upregulated (P < 0.05) genes involved in antimicrobial activity, including CTSB and LYZ, whereas expression of BPI, CATHL5, CATHL6, DEFB1, DEFB10, DEFB3, DEFB4A, DEFB5, DEFB6, DEFB7, ELANE, LAP, LTF, PRKCB, and TAP did not differ among treatments. In addition, treatment did not affect genes related to oxidative burst, including CYBA, CYBB, GPX1, GSR, MPO, NCF1, NCF2, NCF4, RAC2, and SOD1. Calcidiol upregulated (P < 0.05) genes involved in Ca binding and transport, including ATP2B1 and STIM1. Cows fed 3mg had less (P = 0.01) expression of TRPV5 than those fed 1mg. Expression of CALM1, CALM2, CALM3, ITPR1, ORAI1, PPP3CA, PPP3CB, and SLC8A1 did not differ among treatments. In addition, expression of genes related to vitamin D metabolism, including CYP27B1, RXRA, and VDR, did not differ among treatments.
Table 5Effect of source and amount of vitamin D on relative mRNA expression in leukocytes from prepartum cows for genes affected by treatment
Prepartum cows at 250 d of gestation were supplemented with 1 or 3 mg of cholecalciferol (CH) or calcidiol (CA). Blood was sampled and mRNA in leukocytes was analyzed on d 270 of gestation. Results are depicted as fold-change relative to cholecalciferol 1 mg.
Values within a row with different superscripts differ (P < 0.05).
a,b Values within a row with different superscripts differ (P < 0.05).
1 Prepartum cows at 250 d of gestation were supplemented with 1 or 3 mg of cholecalciferol (CH) or calcidiol (CA). Blood was sampled and mRNA in leukocytes was analyzed on d 270 of gestation. Results are depicted as fold-change relative to cholecalciferol 1 mg.
2 Source = effect of source of vitamin D (CH vs. CA); amount = effect of amount of vitamin D (1 vs. 3 mg); source × amount = interaction between source and amount (CH1 + CA3 vs. CH3 + CA1).
The pattern of expression of genes in leukocytes postpartum affected (P < 0.05) by treatment are depicted in the heatmap in Figure 4B. The same heatmap is depicted for individual cows postpartum in Supplemental Figure S2B (https://figshare.com/s/f19a05c4324597477f6e). Calcidiol upregulated (P < 0.05) genes involved in cell adhesion and migration, including CXCR2, SELL, and TLN1; cell signaling cascade, AKT2; synthesis of cytokines, including CCL2, IL1R1, and IL1RN; and antimicrobial mechanisms, DEFB3 and RAC2; but it downregulated (P < 0.05) a gene involved in Ca binding, CALM3, compared with cows fed CH (Table 6). Feeding 3mg upregulated (P < 0.05) the expression of DEFB5 but downregulated (P < 0.05) the expressions of MPO and PP3CA compared with cows fed 1mg. Interactions (P < 0.05) between source and amount of vitamin D were observed for LTF, MMP9, SOD1, and STIM1. The expression of LTF was greatest in CA1, MMP9 was greatest in CH3, SOD1 was greatest in CH1, and STIM1 was greatest in CA1.
Table 6Effect of source and amount of vitamin D on relative mRNA expression in leukocytes of postpartum cows for genes affected by treatment
Prepartum cows at 250 d of gestation were supplemented with 1 or 3 mg of cholecalciferol (CH) or calcidiol (CA). Blood was sampled and mRNA was quantified in leukocytes analyzed on d 3 postpartum. Results are depicted as fold-change relative to cholecalciferol 1 mg.
Values within a row with different superscripts differ (P < 0.05).
a,b Values within a row with different superscripts differ (P < 0.05).
1 Prepartum cows at 250 d of gestation were supplemented with 1 or 3 mg of cholecalciferol (CH) or calcidiol (CA). Blood was sampled and mRNA was quantified in leukocytes analyzed on d 3 postpartum. Results are depicted as fold-change relative to cholecalciferol 1 mg.
2 Source = effect of source of vitamin D (CH vs. CA); amount = effect of amount of vitamin D (1 vs. 3 mg); source × amount = interaction between source and amount (CH1 + CA3 vs. CH3 + CA1).
); thus, the treatments imposed evaluated the effects of manipulating source and amount of vitamin D in cows with adequate vitamin D status. As anticipated, source of vitamin D, CH or CA, had marked influence on concentrations of the respective vitamins in plasma, and the effect was more pronounced in those fed 3 versus 1 mg/d. Feeding a daily dose of 3mg of either source increased the proportions of granulocytes and monocytes pre- and postpartum and, consequently, reduced those of cells of the adaptive immune system, such as CD14− CD21− cells. However, cows fed 3mg tended to have reduced percentage of granulocytes with phagocytic activity and reduced MFI for phagocytosis prepartum. Feeding CA resulted in upregulation of several genes related to cell adhesion and migration, pathogen recognition receptor, cell signaling, synthesis of cytokines, and antimicrobial mechanisms in leukocytes prepartum, with some carryover to the postpartum. In contrast, measures of cell function investigated were not affect by treatment. Based on mRNA profiles, it is possible that immune cells from cows fed CA were better equipped to elicit a response and potentially to recognize and control pathogens. Cows fed 3mg CA had greater milk yield (
) with no reductions in disease observed; hence, it is possible that CA improved the immune cell response, mitigating detrimental effects of unnecessary inflammatory responses observed in severe cases of disease postpartum, by reducing either the intensity or the length of the inflammatory response, but benefits were insufficient to reduce the incidence of clinical disease.
Cell surface adhesion proteins are essential for leukocyte migration to peripheral lymph nodes and into target tissues. Regulation of CD62L controls the movement of lymphocytes, and once lymphocytes are activated by antigen-presenting cells in a lymph node, they shed CD62L from the membrane and move into blood (
). In the current experiment, we observed an increase in CD62L expression in CD14− CD21− cells in cows supplemented with 3mg compared with 1mg, suggesting greater presence of naïve lymphocytes in those cows. Manipulations of diet can affect leukocyte function, as supplemental fatty acids affect the abundance of granulocytes expressing CD62L, as well as altering the function of granulocytes from transition cows (
). Feeding CH in the present experiment increased prepartum expression of CD62L and the proportion of positive cells for CD62L in granulocytes of cows, although the effect was independent of amount of vitamin D fed. In contrast,
observed that supplementation with 3 mg/d of CA increased CD62L expression in milk macrophages and granulocytes, although the response observed was specific to CA compared with 1 mg/d of CH. The reduced percentage of monocytes CD62L+ and the reduced intensity of CD62L expression in monocytes in cows fed CA suggest more activation of those cells, perhaps linked with the observed changes in mRNA expression with indications of increased abundance of genes involved in recognizing and controlling pathogens.
Expression of leukocyte mRNA for several cell migration– and adhesion molecule–related genes were upregulated by supplementing CA. Of the 11 genes evaluated related to cell adhesion and migration, 7 were upregulated by CA prepartum, whereas only 3 were upregulated postpartum. Leukocyte migration is mediated by proteins transcribed by the selectin genes SELL and SELE, which reduce the rolling velocity of cells in the endothelium (
). This first adherence of leukocytes to endothelial cells uses CD44 for cell-to-cell interaction and, once cells are firmly attached, enables signaling and activation of integrins such as ITGAL and ITGB1. Intracellular integrin domains are interconnected with the cytoskeleton, and this binding of integrins to actin is mediated by TLN1 (
). Altogether, CA supplementation upregulated genes involved in the pathway for leukocyte recruitment, suggesting that those leukocytes would potentially be more readily able to migrate into the target tissue to mount a response to pathogens. This may be one of the reasons
observed increased mRNA expression of IL1B and inducible nitric oxide synthase in milk somatic cells after an intramammary bacterial challenge in cows supplemented with CA.
Defense against bacterial infection relies on recognition of antigens by pathogen recognition receptors. When activated, membrane receptors such as TLR2 and TLR6 upregulate the genes for COX-2 and PLA2 and the proinflammatory cytokines TNF and IL1b in a process mediated by MYD88 and NFKB1/2 (
). In addition, IL1B can further expand the response of the immune system by binding to its receptor, IL1R1, which can further stimulate the MYD88 and NFKB1/2 cascade, resulting in more proinflammatory elements and amplification of the immune response (
). Additionally, TLR2 can stimulate the MAPK pathway and activate FOS and JUN, consequently promoting transcription of target genes with AP-1 binding sites, including several cytokines, such as IL6, IL10, TNF, and GM-CSF. Calcidiol treatment upregulated several genes associated with these pathways during the prepartum period, including IL1RN, which is an antagonist of IL1R1 and prevents overstimulation of the IL1B response (
). Therefore, it appears that CA induced upregulated genes with potential for a proinflammatory response, yet at the same time CA also upregulated IL1RN as a control check point to prevent potential undesirable inflammatory responses.
Feeding 3mg tended to reduce granulocyte phagocytosis prepartum compared with feeding 1mg treatments, whereas no changes were observed postpartum. In contrast with our data,
found increased percentage of granulocytes positive for oxidative burst postpartum in cows supplemented with CA. Under bacterial exposure, reactive oxygen species are produced from superoxide anions and are used for microbial killing (
). A key enzyme in the synthesis of reactive oxygen species is NADPH oxidase, a complex of proteins formed by NCF1, NCF2, and NCF4, which were not affected by vitamin D supplementation in the present experiment. Thus, the potential benefits of CA supplementation on phagocytosis and oxidative burst remained unclear based on the results of this experiment. Nevertheless, enzymes that can aid during bacterial infection, such as those conferred by the gene LYZ, were upregulated by CA during the prepartum period, along with CTSB, the gene that encodes cathepsin B, an enzyme involved in supporting phagocytosis and metalloproteinases.
Calcium is an important intracellular second messenger, and activation of cells results in release of endogenous ionized Ca and increased uptake from the extracellular space in a process denominated store-operated Ca entry (
). An essential protein in this process is STIM1, which regulates and activates the store-operated Ca entry, together with ORAI1. Reduced concentrations of cytosolic ionized Ca have been associated with impaired granulocyte phagocytosis and oxidative burst in dairy cows (
). However, vitamin D metabolites are key regulators of Ca homeostasis, and treatments resulted in minor changes in expression of genes involved in intracellular Ca trafficking in leukocytes. Incidence of milk fever and serum concentrations of total Ca did not differ among treatments. Perhaps the lack of effects of treatment on serum concentrations of Ca and other minerals abolished any potential effect of source or amount of vitamin D on genes involved in Ca regulatory mechanisms in leukocytes.
Evidence exists for the role of vitamin D metabolites on immunity. Intramammary treatment with CA has been shown to reduce severity of bacterial-induced mastitis in dairy cows (
showed that intramammary treatment with CA increased abundance of the 1α-hydroxylase gene CYP27B1 in milk cells, suggesting that CA might stimulate local conversion of 25-dihydroxyvitamin D3 into 1α,25-dihydroxyvitamin D3. Their work also showed that mammary cell inflammation induced by LPS activates the vitamin D system with upregulation of CYP27B1 and CYP24A1 in cells isolated from milk. Increased machinery for local synthesis of 1α,25-dihydroxyvitamin D3 by immune cells has been observed in vitro in human macrophages and in vivo in bovine monocytes (
). In support of that, increasing CA resulted in greater β-defensin response (DEFB3, DEFB4, DEFB7, and DEFB10) in a dose-dependent manner in monocytes from lactating cows stimulated with LPS in vitro (
). Thus, vitamin D status might contribute to the ability of the immune response to confer resistance to bacterial infections. Although not conclusive, the present results suggest that immune cells from cows supplemented with CA experience upregulation of genes that might better equip them to harness an immune response, thereby providing an opportunity to improve the defense mechanisms in response to peripartum bacterial diseases. It is noteworthy that feeding CH1 resulted in a mean concentration of 25-hydroxyvitamin D3 of 57 ng/mL, sufficient to maintain adequate vitamin D status (
). The changes observed in this experiment suggest that providing vitamin D beyond the amounts needed to prevent perceived deficiency, in particular with supplemental CA, might play a role in supporting innate immune defenses in dairy cows.
Peripheral blood leukocyte mRNA expression is regulated by supplementation of vitamin D, and calcidiol enhanced expression of numerous genes involved in a multitude of pathways for cell adhesion and migration, intracellular signaling, and activation of antimicrobial mechanisms to eventually eliminate pathogens. Although a larger number of genes were affected by calcidiol prepartum than during the early postpartum period, genes in the same pathways were differentially expressed and followed a similar pattern pre- and postpartum, suggesting some carryover effect as lactation starts. Among the genes differently expressed, those related to cell adhesion and migration, pathogen recognition receptors, cell signaling, and synthesis of cytokines were more susceptible to changes. It is likely that changes in mRNA expression with calcidiol were induced by the increased substrate for local synthesis of 1α,25-dihydroxyvitamin D3 in immune cells. Nevertheless, it is important to note that, despite the observed changes in concentrations of vitamin D in plasma and of mRNA in blood leukocytes, treatments were unable to affect the risk of clinical diseases in the first 60 DIM. Supplementing calcidiol is a more effective means of manipulating the concentrations of 25-hydroxyvitamin D3 in plasma of dairy cows and provides an opportunity to alter leukocyte machinery to harness immune responses to fend off infections.
The authors thank J. Driver (University of Florida, Gainesville) for technical assistance and K. C. Jeong (University of Florida) for preparing the labeled Escherichia coli for granulocytes assays. The help of the staff of the University of Florida Dairy Research Unit is greatly appreciated. The authors thank Stephane Etheve (DSM Nutritional Products, Kaiseraugst, Switzerland) for assistance with vitamin D assays in plasma samples. Partial funding for this project was provided by the Southeast Milk Check-Off Dairy Research and Education Projects (Belleview, FL) and DSM Nutritional Products. The authors have not stated any conflicts of interest.
Biological role of interleukin 1 receptor antagonist isoforms.