MECHANISMS BY WHICH FEEDING SYNTHETIC ZEOLITE A AND DIETARY CATION ANION DIFFERENCE DIETS IMPACT MINERAL METABOLISM IN MULTIPAROUS HOLSTEIN COWS: PART I

The periparturient period is characterized by the increased demand for calcium ( Ca ) in dairy cows. This has resulted in the utilization of several different pre-partal nutritional strategies to prevent hypocalcemia postpartum. The objective of our study was to determine the effects of feeding synthetic zeolite A ( XZ ), a negative dietary cation-anion difference ( DCAD ) diet, or a positive DCAD diet ( CON ) during the close-up period on peripartal mineral dynamics and hormones involved in calcium metabolism. To this end, one hundred and 21 multiparous Holstein cows, blocked by lactation number and expected due date were enrolled at 254 d of gestation and randomly assigned to 1 of 3 prepartum diets: CON (+190 mEq/kg; n = 40), -DCAD (−65 mEq/kg; n = 41), or a diet supplemented with sodium aluminum silicate (XZ; +278 mEq/kg, fed at 3.3% DM, targeting 500 g/day; n = 40; Protekta Inc.). Blood, urine, and saliva samples were collected from enrollment until parturition, with data analyzed and presented beginning 14 d before parturition (D-14) until parturition (D0), and on D1, D2, D3, D6, D9, D12, D15, D18, D21, D35, and D49 postpartum to assess mineral and hormone dynamics. Total fecal collections were performed in a subset of 8 cows per treatment group to assess fecal mineral loss. Data was analyzed as a randomized complete block design in SAS. Cows fed XZ and -DCAD had higher blood Ca concentrations compared with CON fed cows, with XZ fed cows exhibiting the highest blood Ca concentrations pre and postpartum. Cows fed XZ had decreased blood and salivary phosphorous ( P ), increased fecal water extractable phosphate ( WEP ), and the highest blood calcium concentrations pre and postpartum. Parathy-roid hormone ( PTH ) was unaffected by diet but was increased at parturition in all treatments. Serotonin concentrations were increased in -DCAD and XZ fed cows compared with CON during the prepartum period. Our data indicate that XZ’s improvement in blood Ca concentrations pre and postpartum most likely is regulated by a dietary P restriction. Taken together, these data suggest that XZ and -DCAD diets improve postpartum calcium metabolism, however, they appear to work through different mechanisms.


INTRODUCTION
The periparturient period is considered to be the 3 weeks before and after parturition and is in part characterized by the dramatic increase in cows demand due to calcium (Ca) secretion into milk at the onset of parturition.Previous research indicates the presence of a high incidence of clinical (CH; 5%) and subclinical hypocalcemia (SCH; 25% in primiparous and 50% in multiparous cows) in the United States (Reinhardt et al., 2011).The occurrence of numerous other diseases during the peripartum period is associated with CH ([tCa] < 1.4 mmol/L) and SCH ([tCa] ≤ 2.0 mmol/L), such as ketosis, displaced abomasum, metritis, and mastitis, which can all lead to decreased milk production, reduced pregnancy rates, and increased culling risk (DeGaris and Lean, 2008).
The regulation of blood Ca concentrations is a tightly controlled process that is influenced by several hormones.As blood Ca concentrations decrease, parathyroid hormone (PTH) is released from the parathyroid gland into the bloodstream (DeGaris and Lean, 2008).Parathyroid hormone aids in mobilizing Ca from bone and regulates the hydroxylation of 25-hydroxycholecalciferol to 1, 25(OH) 2008).In addition to PTH and 1,25(OH) 2 D 3, a hormone called serotonin has recently been shown to regulate mammary gland Ca metabolism through stimulation of parathyroid hormone-related protein (PTHLH), thus influencing systemic calcium homeostasis.Research has established that PTHLH, a hormone similar in homology and biological action of PTH (Moseley et al., 1987), is secreted by the mammary gland during lactation and acts to support Ca homeostasis by increasing bone Ca mobilization (Slater et al., 2018;Connelly et al., 2021).Additionally, PTHLH and serotonin were positively correlated with Ca concentrations at calving (Laporta et al., 2013) and infusion of 5-hydroxy-L-tryptophan (5-HTP) prepartum, the immediate precursor to serotonin, has been demonstrated to increase Ca concentrations in blood and colostrum (Weaver, et al., 2016;Hernández-Castellano et al., 2017, Slater et al., 2018).
In addition to hormonal regulation of Ca metabolism, the dietary content of various macrominerals has been demonstrated to influence blood Ca concentrations and hypocalcemic risk.Previous research has also demonstrated that feeding prepartum diets high in phosphorus (P) concentrations increases the incidence of hypocalcemia (Lean et al., 2006).Recently, it was demonstrated that restricting dietary P (0.16% DM) during the last 4 weeks of gestation improved Ca metabolism in early lactation due to increased bone mobilization (Wächter et al., 2022).Another dietary approach to increasing peripartal blood Ca concentrations is the use of synthetic zeolite in prepartum diets.Zeolite A is a sodium aluminum silicate molecule that binds dietary ions.Previous research suggests that if the concentration of "free" Ca is low, it can bind to other ions, including Mg and P, decreasing the absorption of these minerals by the gut and leading to the activation of Ca homeostatic mechanisms (Thilsing et al., 2006).Recent research demonstrated that prepartum cows fed with synthetic zeolite A had higher blood Ca concentrations (approximately 0.4 mmol/L) compared with control-fed cows on the first day of lactation (Kerwin et al., 2019).Additionally, blood P concentrations were markedly decreased during the feeding period through d 2 of lactation in this study relative to control-fed cows.Interestingly, previous in vivo and in vitro experiments using synthetic zeolite A have demonstrated large reductions in P concentrations in blood and rumen fluid samples after the addition of an acid solution (Thilsing et al., 2006(Thilsing et al., , 2007;;Kerwin et al., 2019).However, previous work has yet to clearly delineate the exact mechanism of how and where P in the cow is being altered.Circulating magnesium (Mg) concentrations are maintained between 0.75 to 1 mmol/L, and when concentrations decrease below normal, the ability of PTH to bind to its receptor in the target tissues is negatively impacted (Goff, 2000).Consequently, bones and kidneys are less responsive to PTH during hypomagnesaemic states (Goff, 2008).Thus collectively, increasing dietary Mg (0.40 -0.45% DM) has been shown to reduce the risk of developing hypocalcemia (Lean et al., 2006;Goff 2008;Lean et al., 2019).Interestingly, feeding XZ has previously been demonstrated to reduce blood Mg concentrations (Grabherr et al., 2009;Kerwin et al., 2019).
Numerous studies have demonstrated that feeding diets with a negative dietary cation-anion difference (DCAD) 21 d before the expected due date improved postpartum Ca homeostasis compared with feeding neutral or positive DCAD diets prepartum (Block, 1984).Compared with neutral or positive DCAD diets, negative DCAD diets induce a metabolic acidosis, allowing bone to accept hydrogen ions, liberating Ca from the bone into circulation (Block, 1984;Charbonneau et al., 2006).Additionally, prepartum negative DCAD diets increase bone sensitivity to PTH (Goff et al., 1991), increase production of 1,25(OH) 2 D 3 by the kidneys (Goff et al., 1991), improves ruminal absorption of Ca independent of 1,25(OH) 2 D 3 (Wilkens et al., 2016), and increases gut absorption and bone resorption of Ca (Block, 1984) which aid in maintaining plasma Ca concentrations at parturition, which occur independent of decreases in Mg or P concentrations.
Numerous studies demonstrate the benefits of feeding negative DCAD diets prepartum on Ca homeostasis postpartum relative to positive DCAD diets (Goff et al., 1991;Lean et al., 2019;Santos et al., 2019).In the recent 2 decades, studies have started to investigate the benefits of feeding synthetic zeolite A prepartum on Ca homeostasis (Thilsing-Hansen et al., 2002;Grabherr et al., 2009;Kerwin et al., 2019).However, there are no studies that directly compare these 2 prepartum strategies that appear to improve blood calcium concentrations through 2 different mechanisms.Therefore, the objective of this study was to determine the effects of feeding a negative DCAD diet, a positive DCAD diet, and a diet supplemented with synthetic zeolite A to multiparous Holstein cows during the close-up dry period and investigate their effects on mineral metabolism and hormones regulating Ca metabolism.We hypothesized that feeding cows a negative DCAD diet or synthetic zeolite starting at 254 d of gestation would result in improved blood Ca concentrations after parturition and that the mechanisms regulating Ca homeostasis by negative DCAD and synthetic zeolite A would differ.

MATERIALS AND METHODS
All experimental procedures were approved by the College of Agriculture and Life Sciences Animal Care

Experimental Design, Animals, and Dietary Treatments
We enrolled a hundred and 28 multiparous Holstein cows in a randomized complete block design, of which 121 remained in the final data set.Random numbers were generated in Microsoft Excel (Microsoft Corporation, 2016), cows were assigned to treatment, and balanced by parity.Cows that received Ca boluses, received their dietary treatment for less than 14 d, or longer than 35 d were not eligible for inclusion in the data set.At 254 d of gestation, cows were moved from the far-off dry cow barn at Blaine Dairy to the close-up dry cow pen which was bedded with straw.No more than 30 animals were allocated at the same time to the closeup pen.Cows were then blocked by expected due date and lactation number and randomly assigned to 1 of 3 different prepartum diets: control (CON; +190 mEq/ kg; n = 40), negative DCAD (-DCAD, −65 mEq/kg; n = 41; Ultra Chlor; Vita Plus), or a diet containing sodium aluminum silicate zeolite (XZ; +278 mEq/kg, targeting 500 g/day AF; n = 40; X-Zelit, Protekta Inc./ Vilofoss).The sample size determination was based on biologically important differences in serum total Ca concentrations between parturition and 2 DIM from previous work reported in Kerwin et al., 2019.The detectable difference in Ca concentration between the 3 treatments were at least 0.20 mmol/L, with α = 0.05 and β = 0.20.Blinding was attempted when possible.The first 3 authors were not blinded to treatment since they were responsible for collection of the samples from the animals; however, all other individuals working in the laboratory with assays were unaware of treatment groups.Further, statistical analysis was additionally assessed by the University of Wisconsin-Madison, College of Agriculture and Life Sciences Statistical Consulting group and they were blinded to treatment groups during analyses.
Diets were formulated based on the assumption of 13.5 kg of DMI, and 5% refusals using NRC, 2001 models (Table 1).Feed intakes were recorded daily prepartum.Cows were housed in a straw-bedded packpen with 15 Insentec Roughage Intake Control (RIC) gates (RIC System, Holofarm group) and fed once daily at approximately 1200h.The 15 RIC gates in the pen were randomly assigned to receive 1 of the 3 prepartum diets, resulting in 5 RIC gates for each treatment so that a group of 10 cows had access to 5 RIC gates that contained the same diet (Weld and Armentano, 2018).Each cow had an electronic ear tag (FDX EID TAG, Allflex) that allowed them to have access to feeders that contained the same diet to which they were assigned.Twice a week, TMR samples were collected and composited, and sent to Dairyland Laboratories Inc. for wet chemistry analysis.
Postpartum cows were placed on a common diet balanced to NRC, 2001 requirements for a 650 kg Holstein cow milking 50 kg/d of ECM.Postpartum cows were housed in a freestall facility, bedded with sand bedding, received a common TMR 1x per day, and were milked 2x daily.Collection of individual cow feed intakes was not possible due to facility limitations for housing for postpartum cows.
At 254 d of gestation, all cows received an intramuscular shot of Endovac-Dairy (Endovac Animal Health, Columbia, MO, USA) for E. coli mastitis and multiple gram-negative bacterial diseases, for the prevention of mastitis, a subcutaneous shot of injectable trace mineral containing zinc, manganese, selenium, and copper (MultiMin 90, Multimin North America, Fort Collins, CO, USA), and a pour-on deworming treatment with moxidectin (Cydectin, Bayer HealthCare LLC, Animal Health Division, Shawnee Mission, KS, USA) based on manufacturer's recommendation.

Blood, Saliva, and Urine Sampling
Blood, saliva, and urine samples were collected daily beginning at enrollment until parturition (D0) approximately one hour before feeding and subsequently on D1, D2, D3, D6, D9, D12, D15, D18, D21, D35, and D49 postpartum.Blood samples were collected from the coccygeal vessel using 20-gauge Vacutainer needles (Greiner Bio-One GmbH; Exelint International, Co.).Whole blood was collected into 10-mL BD Vacutainer K2 EDTA Plus (368589; Becton, Dickinson, and Company), 10-mL BD Vacutainer Serum (367820; Becton, Dickinson, and Company), and 10-mL Lithium Heparin 158 USP Units (367880; Becton, Dickinson, and Company) blood collection tubes and inverted gently.Immediately after inversion, 3 to 4 mL of whole blood was transferred from the 10-mL Vacutainer K2 EDTA Plus tube to a 5 mL Eppendorf preloaded with approximately 35 mg of ascorbic acid (10 mg/mL) to prevent oxidation of serotonin, and frozen at −20°C until further analysis (Connelly et al., 2021).Samples from whole blood collected into lithium heparin tubes were analyzed for blood pH, concentrations of HCO - 3 , base excess, partial pressure of CO 2 (pCO 2 ), total dissolved CO 2 (tCO 2 ), partial pressure of O 2 , oxygen saturation, concentrations of sodium (Na), potassium (K), and ionized calcium (iCa), hemoglobin, and hematocrit using a hand-held biochemical analyzer and analyzed using the CG8+ cartridge (VetScan i-Stat, Abaxis), measured and calculated results were not temperature corrected.Blood samples collected into serum tubes were kept at room temperature and allowed to clot for 1 h before centrifugation.Serum and plasma were harvested after centrifugation at 3,000 x g for 20 min at 4°C.Samples were allocated into triplicate aliquots and stored at −80°C until further analysis.
Saliva samples were collected using Salivette tubes (51-1534.500;Sarstedt, Aktiengesellschaft & Co).Each tube contained a cotton swab clipped with a surgical clamp and cows were allowed to chew it (Riek et al., 2019).Then, the cotton swab was placed back into the tube and centrifuged at 1,000 x g for 10 min at 4°C.Saliva samples (approximately 1 to 2 mL per cotton swab) were stored at −20°C until analysis.Before analyzing saliva, samples were scored from 1 to 4 ac-cording to clarity.This allowed for the classification of samples based on saliva ruminal fluid differences due to potential contamination of sampling (Figure 1) as it can influence concentrations of various biological outcomes (Storm et al., 2013;Contreras-Aguilar et al., 2021;Contreras-Aguilar et al., 2022).Acetic acid (0.5 M) was then added to samples using a 1:2 dilution before mineral analysis and centrifuged at 13,000 x g for 12 min at 4°C and the supernatant was separated and stored at −20°C for further analysis (Connelly et al., 2023, accepted).
Urine samples were collected into 5 mL Eppendorf by gently stimulating the area between the udder and the vulva.Urine pH was measured using an electronic pH meter (Horiba LaquaTwin Compact PH Meter) and calibrated daily before measurement to ensure -DCAD efficacy.After measuring pH, urine samples were frozen at −20°C until further analysis.Acetic acid (0.5M) was added to samples in a 1:2 dilution and centrifuged at 13,000 x g for 12 min at 4°C and the supernatant was separated and used for mineral analysis.

Mineral Digestibility
Fecal samples were collected approximately 10 d before parturition over 3 d at 8-h intervals covering a 4-h clock period (Cook et al., 2016) to encompass a composite sample per cow in a random subset of 24 cows enrolled in the study (n = 8/treatment).Samples were obtained rectally with clean palpation sleeves and placed in a plastic container, one per cow.After collection, samples were frozen at −20°C until further analysis.The TMR and orts samples were collected the same week as the feces collection.Samples were thawed and dried in a forced-air oven at 60°C for 72 h, ground to pass through a 1-mm screen of a Wiley mill (A.H. Thomas) and submitted for undigested NDF (uNDF) after 240 h of incubation and minerals analysis (Dairyland Laboratories Inc., Arcadia, WI).Calculations were made for fecal output as described in (Fustini et al., 2017).Dried and ground fecal samples were also analyzed for water-extractable phosphate (WEP; PO 4 ) concentrations (UW Soil and Forage Laboratory).For this analysis, a representative sample with 1 to 3 g of solids is extracted with water in a 100:1 (solution: solids) dilution ratio, centrifuged, and filtered.The filtration product is analyzed for P concentration using an inductively coupled argon plasma spectrophotometer (Kleinman et al., 2007).
Serotonin concentrations were analyzed in whole blood samples (IM1749, Immunotech, Beckman Coulter) according to the manufacturer's instructions and as previously described (Connelly et al., 2020).The intra-and interassay CV's for blood 5-HT were 3.73% and 15%.Parathyroid hormone concentrations were analyzed in plasma samples, using Bovine iPTH (Intact PTH) ELISA Kit (DEIA 1826B, Creative Diagnostics) per the manufacturer's instructions.Briefly, plasma samples were diluted 1:2 using a dilution buffer.The bovine iPTH standards were prepared and used according to the manufacturer's instructions (Connelly et al., 2021).The intra-and inter-assay CV were 5.73% and 14.40%.

Statistical Analysis
Prepartum and postpartum data were analyzed separately, except for Se due to sampling time points and data are presented as LSMeans ± SEM.Prepartum data was restricted from D-14 until D0 for analysis.A baseline measurement was determined using the first sample collected from each cow before the start of treatment and was used as a covariate for the respective analyses.Prepartum cows at 254 d of gestation were blocked by lactation number [lactation 2 (Lact2) vs lactation >3 (Lact3+)] and expected due date.All statistical analyses were conducted using the MIXED procedure of SAS (version 9.4 SAS Institute Inc.).Fixed terms in the model for blood and urine variables were treatment, day, lactation number, covariate, the interaction of treatment x day, and interaction of treatment x lactation number.Fixed terms in the model for blood gases analyzed using a hand-held biochemical analyzer were treatment, day, lactation number, weeks on treatment, the interaction of treatment x day, and interaction of treatment x lactation number.For saliva variables, the same variables were used as in the blood biochemical analysis as fixed terms with the addition of salivary grade.The random statement in all models included cow.Two separate analyses were used to fit the proper covariance structure per sampling day due to differences in sampling schemes pre and postpartum.Day was considered the repeated measure in both analyses.The first analysis included the prepartum period with daily sampling and to account for autocorrelated errors, the ar(1) structure was utilized.Due to different sampling timeframes the analysis in the postpartum period the spatial power structure was utilized as a covariance structure.Data were analyzed for normality, using Shapiro-Wilk, AIC values, and Q-Q plots.When normality assumptions failed, data was transformed.Transformations were based on diagnostics plots and overall model fit.If transformations were necessary values are depicted as median and interquartile.To evaluate trace mineral status, serum Se was measured at 3 time points across the peripartal period (D-6, D0, D6) in a subset of cows (n = 20 per treatment).Therefore, for analysis of Se concentrations, a single analysis was utilized and the fixed effects were treatment, lactation number, day, and treatment x day.The random effect was cow and the covariance structure was ar(1).Statistical significance was declared if P ≤ 0.05.

RESULTS
Seven cows were excluded from the analysis (3 cows from CON diet, 1 cow from -DCAD diet, and 3 cows from XZ diet).Four animals were removed from the study due to receiving supplemental Ca treatment postpartum, 2 cows received their respective dietary treatment for less than 14 d, and 1 cow remained on their respective dietary treatment for more than 35 d.

Blood, Saliva, and Urinary Mineral Concentrations
Serum iCa concentrations for cows fed 3 prepartum diets are presented in Figure 2A and serum tCa concentrations are presented in Figure 2B.Cows fed the XZ diet had the highest serum iCa and tCa concentrations during the prepartum period (P < 0.01).On D0, D1, and D2 cows fed the XZ diet continued to maintain the highest iCa concentrations (P < 0.01) compared with -DCAD and CON diets (Figure 2A).Prepartum concentrations of iCa (P < 0.01) were the highest in cows fed XZ, followed by -DCAD fed cows, and were lowest in cows fed the CON diet (Figure 2A).Cows supplemented with synthetic zeolite A prepartum had the highest iCa concentrations (1.26 ± 0.01 mM) compared with CON (1.17 ± 0.01 mM) and -DCAD (1.22 ± 0.01 mM) cows postpartum.Concentrations of iCa were dynamic across the prepartum period with D0 being the lowest concentrations.Concentrations of iCa on D0 were the lowest (P < 0.01) during the prepartum period.Postpartum, iCa concentrations (P < 0.01) on D1 were the lowest.Consistent with the iCa concentrations on D0, D1, and D2, cows fed the XZ diet continued to maintain higher tCa concentrations (P < 0.01) compared with -DCAD and CON diets (Figure 2B).After D3 there were no differences in iCa concentrations between treatments, days, or lactation number (P > 0.05).After D3 there were no differences in serum tCa concentrations between treatments, except on D9 when cows fed the CON diet had the highest serum tCa concentrations, cows fed the XZ diet had the lowest serum tCa concentrations, and cows fed the -DCAD diet did not differ from either treatment.Cows in Lact2 had higher serum tCa concentrations compared with cows in their Lact3+ in both the prepartum (2.44 ± 0.01 and 2.40 ± 0.01 mmol/L, respectively; P < 0.01) and postpartum (2.40 ± 0.01 and 2.35 ± 0.01 mmol/L, respectively; P < 0.01) periods.
Salivary Ca concentrations are presented in Figure 2C.Cows fed the XZ diet had the highest salivary Ca concentrations compared with -DCAD and CON cows during the prepartum period (1.00 [0.78, 1.51], 0.87 [0.68, 1.29], and 0.85 [0.67, 1.21] mmol/L, respectively; P < 0.01).Because of the substantial variation in debris present in the saliva samples during collection, an analysis that considered the sample coloration grade was conducted.This approach was implemented to account for the potential impact of debris on mineral concentrations.A prepartum (P < 0.01) and postpartum (P < 0.01) effect of the salivary grade was observed on salivary Ca concentration, with the highest concentrations being detected in grade 4 (1.82 [1.46, 2.41] and 3.79 [2.31, 5.28] mmol/L, respectively), followed by grade 3 (1.60 [1.14, 2.13] and 2.33 [1.68, 3.58] mmol/L,  There was a treatment effect (P < 0.01) and day effect (P < 0.01) during the prepartum and a day effect (P < 0.01) and treatment by day effect (P < 0.01) during the postpartum period for iCa.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), and treatment by day effect (P < 0.01) during the prepartum period and a day effect (P < 0.01), lactation effect (P < 0.01), and treatment by day effect (P < 0.01) during the postpartum period for serum Ca.There was a treatment effect (P < 0.01), day effect (P < 0.01), grade effect (P < 0.01), and treatment by lactation effect (P = 0.01) during the prepartum period, and grade effect (P < 0.01) during the postpartum period.There was a treatment effect (P < 0.01) and day effect (P < 0.01) during the prepartum period and a day effect (P < 0.01) and lactation effect (P < 0.01) during the postpartum period for urinary Ca. a, b, c Treatment by day effects when treatment means differed significantly (P ≤ 0.05).
No differences were detected due to treatment (P > 0.05) for prepartum mineral intake, fecal excretion, absorption, or digestibility (Table 3).It is important to notice that the observed numerical differences may not achieve statistical significance, likely due to the low sample size within the subset (n = 8 per treatment).There was an effect of number of days on diet before fecal sampling on fecal excretion of P (P = 0.05).Additionally, Mg digestibility was affected due to number of days on diet before fecal sampling (P < 0.01).No other effects of days on diet before fecal sampling was observed (P > 0.05).While we did not observe treatment effects on mineral intake, fecal excretion, absorption, or digestibility, we did observe that cows fed XZ had increased excretion of WEP (P = 0.01) compared with cows fed either CON or -DCAD diets (Figure 5).There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), treatment by day effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the prepartum period, and a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), treatment by day effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the postpartum period for serum P.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), grade effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the prepartum, and treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), and grade effect (P < 0.01) during the postpartum period for salivary P.There was a treatment effect (P < 0.01), and day effect (P < 0.01) during the prepartum period, and a day effect (P < 0.01) and, treatment by day effect (P < 0.01), during the postpartum period for urinary P. a, b, c Treatment by day effects when treatment means differed significantly (P ≤ 0.05).
Prepartum concentrations of Na (P < 0.01) and K (P < 0.01) were the highest in cows fed the -DCAD diet, followed by the CON diet, and were lowest in cows    fed the XZ diet (Table 4).No differences in Na (P > 0.05) and K (P > 0.05) were observed during the prepartum period.Cows in Lact2 had lower Na (P < 0.01) and higher K (P = 0.04) concentrations compared with cows in Lact3+ during the postpartum period.On D1 the concentration of Na (P < 0.01) was highest.Cows in their Lact2 had lower Na concentration (P < 0.01) than cows in their Lact3+ during the prepartum period.The concentration of Na (P < 0.01) on D0 was the highest during the prepartum period, while concentrations of K (P < 0.01) were lowest.

Blood Hormone Concentrations
Whole blood serotonin concentrations are presented in Table 4.We did not detect an effect of treatment during the prepartum (P > 0.05) or postpartum (P > 0.05) periods on serotonin concentrations.However, there was a day effect during the prepartum period (P < 0.01), with the highest blood serotonin concentrations on D-14 (1936.05 ± 36.68 ng/mL).Additionally, there was a tendency for interaction between day and treatment during the prepartum (P = 0.09) and postpartum (P = 0.05) periods.On D-3, -DCAD and XZ cows tended to have higher serotonin concentrations compared with CON cows (1828CON cows ( .44 ± 62.82, 1821CON cows ( .89 ± 63.67, and 1676.94 ± 64.42 .94 ± 64.42 ng/mL, respectively; P = 0.09).On D3, XZ and CON cows tended to have increased serotonin concentrations compared with -DCAD fed cows (1999.38 ± 81.26, 1828.18 ± 82.82, and 1805.86 ± 79.42 ng/mL, respectively; P = 0.05).Plasma PTH concentrations according to pre and postpartum periods are presented in Table 4.There was no treatment effect on plasma PTH during the prepartum (P > 0.05) or postpartum (P > 0.05) periods.However, there was an effect of the day during the prepartum period (P < 0.01), with the highest PTH concentrations occurring on D0 (183.15 ± 6.45 pg/mL).

DISCUSSION
The purpose of this study was to investigate the effects of 2 different prepartum dietary hypocalcemic mitigation strategies, -DCAD and synthetic zeolite A, that are utilized to alter the pre-fresh cow's physiology to improve calcium homeostasis and metabolism.Previous research has demonstrated that both synthetic zeolite A supplementation and negative DCAD diets during the close-up period can lead to similar increases in blood Ca concentrations in cows postpartum (Thilsing et al., 2007;Kerwin et al., 2019;Santos et al., 2019).However, there is a gap in knowledge comparing these 2 approaches directly.Studies feeding synthetic zeolite A supplementation have demonstrated a marked reduc-tion in blood P concentrations over the feeding period.However, it is not clear how this collectively affects cow metabolism and production responses (Thilsing-Hansen et al., 2002;Grabherr et al., 2009;Kerwin et al., 2019), but Wächter and colleagues did not observe health-related outcomes related to feeding restricted dietary P for 4 weeks antepartum.However, it is crucial to note that their study lacked sufficient statistical power to address this specific question.In a study conducted examining prepartum P restriction researchers demonstrated that serum Ca concentrations were improved postpartum and this was due to increased bone mobilization (Wächter et al., 2022).Further, they observed a reduced incidence of subclinical, persistent, and prolonged hypocalcemia indicated by improvements in blood calcium across the corresponding days postpartum.Interestingly, this reduction in subclinical, persistent, and prolonged hypocalcemic cows was repeated in the present study, with blood calcium concentrations being higher in XZ fed cows on D0, D1 and D2 relative to -DCAD and CON fed cows.Additionally, as previously described (Lean et al., 2006;Lean et al., 2019), cows fed negative DCAD diets in our study also demonstrated increased postpartum Ca concentrations compared with the CON group.
On average, serum P concentrations were 52% lower in the XZ group compared with the -DCAD and CON groups during the prepartum period.Similar results were observed when cows were supplemented with synthetic zeolite A prepartum in previous experiments (Pallesen et al., 2008;Grabherr et al., 2009;Kerwin et al., 2019) with the thought that supplementing synthetic zeolite A prepartum reduces P availability in the gastrointestinal tract for the cow (Thilsing et al., 2007;Kerwin et al., 2019).This was supported in the present study through an increase in fecal WEP and a reduction in serum and salivary P concentrations.Our study is the first in vivo experiment to demonstrate that it appears that XZ is working primarily through binding P, rather than Ca, as P is increased in fecal excretions.While we did not observe significant changes in fecal excretion of P or Ca using an undigestible marker, we did observe an increase in fecal WEP.However, these findings should be repeated in a larger sample size, given that we collected fecal samples from only 8 cows per treatment and there is a significant amount of error associated with this type of analysis, particularly during a period when DMI is decreased.Additionally, given the use of the Insentec feeders, we were unable to collect orts from individual animals, which is an additional limitation of the uNDF240 analysis in our experiment.
Zeolites are commonly used in ion exchange reactions and are well known for their ability to bind ions.For example, in an in vitro study, zeolite was demonstrated to extensively bind P, in addition to the binding of Ca (Inglezakis, 2004, Thilsing et al., 2006).This study, in addition to Kerwin et al., suggests that feeding synthetic zeolite A induces a mild to moderate hypophosphatemia which can significantly improve Ca homeostasis (Kerwin et al., 2019) and improve blood Ca concentrations.This improvement in blood Ca may be occurring through a variety of mechanisms, with one being increased synthesis of vitamin D. Elevated blood P concentrations (>2 mmol/L) have been shown to inhibit the conversion of 25-hydroxyvitamin D to 1,25(OH) 2 D 3 , reducing vitamin D action and thus decreasing Ca absorption at the gut (Goff, 2006).However, when cows were submitted to a dietary P restriction (0.16% of P in DM) and had decreased blood P concentrations during the last 4 weeks of gestation, blood Ca concentrations were increased in the peripartal period without an increase in PTH or 1,25(OH) 2 D 3 concentrations (Wächter et al., 2022).It is important to state that this result should be interpreted with caution, as the authors acknowledged limitations in statistical power for addressing this specific question.This is contrary to Thilsing-Hansen and colleagues (2002) who indicate that decreased plasma P concentrations after zeolite supplementation are due to increased circulating PTH and increased salivary and renal P excretion.Despite clear understanding to how PTH and 1,25(OH) 2 D 3 contribute to this collective improvement in blood calcium, cows P-restricted prepartum did have higher carboxy-terminal collagen crosslinks (CTX) concentrations, a marker of increased bone resorption, compared with cows fed adequate dietary P (0.30% of P in DM) (Wächter et al., 2022).The exact mechanism of the hormonal relationship between hypophosphatemia and bone resorption remains poorly understood but has been suggested to be driven through a phosphorus regulating hormone called fibroblast growth factor (FGF23).
Fibroblast growth factor 23 is a hormone synthesized by the osteocytes and regulates P concentrations during hyperphosphatemia.This occurs by increasing renal P excretion and decreasing the synthesis of 1α-hydroxylase, the enzyme responsible for converting vitamin D into 1,25(OH) 2 D 3 (Kuro-o, 2010).Previous research has demonstrated that FGF23 is stored in osteocytes and is released under conditions of rapid bone mobilization, leading to increased serum Ca and further increased serum FGF23 in non-ruminant species (Yamazaki et al., 2015).Additionally, administration of 1,25(OH) 2 D 3 has been demonstrated to increase FGF23 concentrations in the circulation by stimulation of transcriptional activity in the osteoblasts (Liu et al., 2006).It is hypothesized that this effect is to aid in maintenance of blood P concentrations during sup-pression of PTH concentrations.In a recent study in sheep, restricted dietary P decreased RNA expression of FGF23, increased 1,25(OH) 2 D 3 concentration, decreased renal P excretion, and increased bone mobilization (Köhler et al., 2021).The decrease in P availability from the diet and corresponding decrease in blood P concentrations in XZ cows may have decreased FGF23 expression in the bone.This possibly leads to increased bone resorption and supporting Ca demands during the peripartal period.Unfortunately, we are unable to measure FGF23 due to the lack of a bovine assay for FGF23 and function has not been explored in the bovine as of yet.Future studies should aim to explore FGF23, 1,25(OH) 2 D 3 , and PTH during P restriction to determine the relationship and impacts of these hormones on mineral metabolism in the transition dairy cow.
Feeding cows a negative DCAD diet during the close-up period has been extensively studied and clearly shown to aid in maintaining and improving Ca homeostasis through the transition period (Block, 1983;Charbonneau et al., 2006).Feeding a negative DCAD diet prepartum helps improve Ca homeostasis through inducing a mild metabolic acidosis, which was clearly instigated in the present experiment.Negative DCAD diets instigate a mild compensated metabolic acidosis.This leads to liberation of Ca from the bone into circulation (Charbonneau et al., 2006), increased bone sensitivity to PTH (Goff et al., 1991), increased production of 1,25(OH) 2 D 3 by the kidneys (Goff et al., 1991), improved ruminal absorption of Ca independent of 1,25(OH) 2 D 3 (Wilkens et al., 2016), and increases in gut absorption and bone resorption of Ca (Block, 1984).These physiological changes collectively aid to maintain blood Ca concentrations at parturition.The present findings align with previous studies, which also observed a decrease in urinary and blood pH, pCO 2 , base excess, HCO 3 -, and tCO 2 in cows fed a negative DCAD diet compared with those fed a positive DCAD diet (Vagnoni and Oetzel, 1998;Charbonneau et al., 2006;Zimpel et al., 2018).However, within the current study when synthetic zeolite A was added to the closeup diet, we observed an increase in blood pH and base excess, as well as a decrease in Na and K concentrations, relative to the negative DCAD and CON diets.These results are consistent with a study by Li et al. (2013), where a different type of zeolite (zeolite 5A -Na 0.4 Ca 5.8 (SiO 2 ) 12 (AlO 2 ) 12 • 27 H 2 O) was tested as an agent to prevent blood loss.The authors found an increase in Ca concentrations and a decrease in Na and K concentrations in the blood after the addition of zeolite (Li et al., 2013).These data along with our observations indicate that while negative DCAD and XZ both mitigate postpartum hypocalcemia, they do so through what appears to be due to 2 distinct mechanisms, one which acts via binding of P and the other via induction of metabolic acidosis.Future research should be aimed at understanding the physiological implications of these findings in the dairy cow due to 2 distinctly separate prepartum hypocalcemic mitigation strategies.
While zeolite supplementation has been previously demonstrated to reduce blood P concentrations, there is a lack of knowledge surrounding how this modulates other mineral homeostasis during the peripartal period (Thilsing et al., 2006(Thilsing et al., , 2007;;Kerwin et al., 2019).The aluminum present in the zeolite molecule is released in a low pH environment and is then free to readily bind P, which then forms a non-absorbable aluminum phosphate complex in the small intestine (Cook et al., 1982;Thilsing et al., 2007).Decreased dietary P absorption is reflected by the decreased plasma P concentrations, likely because the main site of P absorption is in the proximal small intestine (Thilsing et al., 2007).Further, Thilsing-Hansen et al., (2002) suggested that decreased plasma P concentrations after zeolite supplementation were due to an increase in circulating PTH and an increase in salivary and renal P excretion.However, we observed that the XZ group had the lowest salivary P concentrations during the prepartum period, as well as the lowest plasma P concentrations.Interestingly, no differences in circulating PTH were observed between treatments.Further, after cessation of XZ supplementation, postpartum salivary P concentrations increased in a similar pattern to serum P concentrations, further suggesting that XZ binds P and is responsible for the decreased blood and salivary P concentrations observed.Upon analysis of fecal samples using a WEP method, we determined that P concentrations were increased in the feces of cows fed XZ compared with cows fed CON or -DCAD.This suggests that P is being bound by XZ and secreted into the feces, bypassing absorption in the intestine and resulting in a phosphorus restriction during the prepartum period.
The regulation of plasma Mg concentrations is not as precise as that of Ca, and it involves a balance between Mg absorption in the rumen and intestine and Mg excretion by the kidneys (Deetz et al., 1982;Greene et al., 1983;Martín-Tereso and Martens, 2014).Consistent with previous studies, our results showed that plasma Mg concentrations were decreased in the XZ fed cows compared with those fed -DCAD and CON diets during the prepartum through D1 (Thilsing-Hansen et al., 2002;Grabherr et al., 2009;Kerwin et al., 2019).However, the Mg concentrations in all treatment groups were within the normal range for dairy cows (0.8 to 1.0 mmol/L) (Goff, 2014).In an in vitro study, it was demonstrated that zeolite has a high affinity to bind to Ca, but when free Ca concentrations in solution decrease, zeolite can bind to other divalent cations, such as Mg, resulting in numerical reductions in Mg intake, as well as absorption and digestibility of Mg (Thilsing et al., 2006).In our experiment, there were higher serum Mg concentrations in cows fed the -DCAD and CON diets, accompanied by increased excretion of Mg into the urine in these 2 groups.This is contrary to cows fed XZ, where urine and blood Mg concentrations were decreased.Given the decrease in blood Mg concentrations during the prepartum period and the decrease in urine Mg concentrations, it is plausible that some Mg is bound by XZ.This finding should be further explored as serum Mg is elevated in cows that experience more hypocalcemia (Goff, 2014), of which XZ fed cows experienced the least hypocalcemia and had the highest blood Ca concentrations.
The primary hormone involved in restoring Ca concentrations is PTH.Parathyroid hormone regulates the reabsorption of Ca in the renal tubules and promotes the mobilization of Ca from bone (Goff, 2006).As PTH concentrations increase, the secretion of 1,25(OH) 2 D 3 from the kidneys also increases to prevent Ca release into the urine, and 1,25(OH) 2 D 3 is responsible for the absorption of dietary Ca in the intestine (Goff, 2006).Therefore, based on the increase in Ca concentrations during early lactation, we expected that PTH concentrations would be elevated in XZ and -DCAD groups.However, no statistical differences in PTH were detected among the dietary treatment groups (Table 4).In a study by Thilsing-Hansen et al. (2002), an increase in 1,25(OH) 2 D 3 concentration was detected in cows supplemented with zeolite A around calving.Although we did not measure 1,25(OH) 2 D 3 in our study, our findings related to circulating PTH concentrations do not align with previous experiments in which they were increased.This suggests that an alternative pathway is being stimulated, potentially FGF23, which may be contributing to the observed results.However, we think it is relevant to note that in our study there was a numerical increase in PTH concentrations for the CON and -DCAD groups during the postpartum period.This could imply that cows in these dietary groups may have faced more challenges in maintaining their Ca concentrations, as they had lower blood Ca concentrations relative to XZ fed cows, thus necessitating a higher production of PTH.Nevertheless, the analysis lacked statistical power due to the limited sampling frequency which is influenced by the short half-life (approximately 4 min) and pulsatility of PTH (Bieglmayer et al., 2002).
Additionally, we assessed Se concentrations in serum as this mineral is critical to immune function and reproductive capacity during pregnancy and the peripartal period (Mehdi and Dufrasne, 2016).Given that XZ is capable of binding ions in their 2 + form, we analyzed the effect of both -DCAD and XZ on Se concentrations in serum.We observed the -DCAD supplementation decreased Se concentrations at calving, while XZ supplemented cows had lower serum Se at D6 compared with the -DCAD and CON.Overall, Se was decreased in XZ compared with the CON, but not in -DCAD fed cows.Despite differences observed, all Se ranges fell within previously reported ranges (0.05-0.08 mg/L; Villar et al., 2002;Wang et al., 2021).Previous work examining the effects of anionic salt supplementation demonstrated that there was no significant impact on blood Se concentrations (Gant et al., 1998).
The onset of lactation demands large amounts of Ca to support milk synthesis.The canonical PTH -1,25(OH) 2 D 3 pathway is not entirely sufficient to provide the Ca that is needed due to the onset of lactation and the adaptation to a new physiological state.Recent studies have focused on exploring the presence of these alternative pathways that play supportive roles during lactation to help facilitate the demands that are being placed upon the animal.Two such supportive mechanisms are serotonin and PTHLH, which act as additional mechanisms to maintain Ca concentrations around parturition and during lactation to support maternal homeostasis (VanHouten et al., 2003;Laporta et al., 2013).Infusion of 5-HTP, the immediate precursor to serotonin, during the prepartum period resulted in higher circulating serotonin concentrations and improved Ca homeostasis in early lactation (Hernández-Castellano et al., 2017;Slater et al., 2018).Rodney and colleagues (2018) demonstrated that cows fed a negative DCAD diet prepartum exhibited increased serotonin and Ca concentrations, and it has also been observed that feeding a negative DCAD diet in combination with 5-HTP infusions for 7-10 d prepartum exhibited increased iCa concentrations in the peripartum period compared with a negative DCAD diet or 5-HTP infusion alone (Slater et al., 2018).Our data demonstrate a tendency for increased serotonin concentrations in both the DCAD and XZ fed cows prepartum, suggesting that serotonin could be aiding the cow's physiology to improve Ca homeostasis in negative DCAD and XZ fed cows in our experiment.While this could be possible, we are unclear how feeding DCAD or XZ could each induce this response and further research should be conducted to understand the underlying mechanisms.
In conclusion, multiparous Holstein cows fed a negative DCAD diet or supplemented with synthetic zeolite A during the close-up period had improved Ca concentrations in early lactation compared with cows fed the CON diet.Cows fed XZ had the highest blood Ca concentrations and had decreased blood P concentrations and salivary P concentrations.This study suggests that feeding synthetic zeolite A appears to work via restricting dietary P and decreasing blood and salivary P concentrations which improves Ca metabolism.Collectively, the data herein demonstrate that cows fed XZ have improved Ca metabolism in the peripartal period being driven by a restriction in P, as evidenced by increased fecal WEP and decreased blood and salivary P. It also appears that XZ and -DCAD increase serotonin concentrations prepartum, and serotonin may be an additional pathway that supports peripartal Ca metabolism.Future studies should be focused on delineating and validating the hormonal pathways regulating P metabolism, like FGF23, with the use of XZ in the prepartal dairy cow.
2 D 3 ), leading to increased intestinal Ca absorption and increased renal tubular reabsorption of Ca (DeGaris andLean,

Frizzarini
et al.: MECHANISMS BY WHICH FEEDING…

Frizzarini
Figure 1.Representative saliva samples by salivary grade: grade 1 = clear sample, without coloration or debris; grade 2 = clear sample, with some coloration, and without debris; grade 3 = sample not clear and with some debris; grade 4 = sample completely cloudy and with debris.

FrizzariniFigure 2 .
Figure 2. Least squares means and SE for prepartum and postpartum (A) iCa (mmol/L) (B) serum Ca (mmol/L); median and interquartile for prepartum and postpartum (C) salivary Ca (mmol/L), and (D) urinary Ca (mg Ca/mg of creatinine) for multiparous cows fed with control (CON), negative DCAD (DCAD), or with supplementation of synthetic zeolite A (XZ) diet during the close-up period.Prepartum and postpartum data were analyzed separately.There was a treatment effect (P < 0.01) and day effect (P < 0.01) during the prepartum and a day effect (P < 0.01) and treatment by day effect (P < 0.01) during the postpartum period for iCa.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), and treatment by day effect (P < 0.01) during the prepartum period and a day effect (P < 0.01), lactation effect (P < 0.01), and treatment by day effect (P < 0.01) during the postpartum period for serum Ca.There was a treatment effect (P < 0.01), day effect (P < 0.01), grade effect (P < 0.01), and treatment by lactation effect (P = 0.01) during the prepartum period, and grade effect (P < 0.01) during the postpartum period.There was a treatment effect (P < 0.01) and day effect (P < 0.01) during the prepartum period and a day effect (P < 0.01) and lactation effect (P < 0.01) during the postpartum period for urinary Ca. a, b, c Treatment by day effects when treatment means differed significantly (P ≤ 0.05).

FrizzariniFigure 3 .
Figure3.Least squares means and SE for prepartum and postpartum (A) serum Mg (mmol/L); median and interquartile for prepartum and postpartum (B) salivary Mg (mmol/L), and (C) urinary Mg (mg Mg/mg of creatinine) for multiparous cows fed with control (CON), negative DCAD (DCAD), or with supplementation of synthetic zeolite A (XZ) diet during the close-up period.Prepartum and postpartum data were analyzed separately.There was a treatment effect (P < 0.01), day effect (P < 0.01), and treatment by day effect (P < 0.01) during the prepartum period, and a treatment effect (P < 0.01), day effect (P < 0.01), treatment by day effect (P < 0.01), and treatment by lactation effect (P = 0.01) during the postpartum period for serum Mg.There was a treatment effect (P < 0.01), and grade effect (P < 0.01) during the prepartum period and a grade effect (P < 0.01) for saliva during the postpartum period for salivary Mg.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the prepartum period, and a treatment effect (P = 0.05), day effect (P < 0.01), lactation (P < 0.01), and treatment by day effect (P < 0.01) during the postpartum period for urinary Mg. a, b, c Treatment by day effects when treatment means differed significantly (P ≤ 0.05).

FrizzariniFigure 4 .
Figure 4. Least squares means and SE for prepartum and postpartum (A) serum P (mmol/L); median and interquartile for prepartum and postpartum (B) salivary P (mmol/L), and (C) urinary P (mg P/ mg of creatinine) for multiparous cows fed with control (CON), negative DCAD (DCAD), or with supplementation of synthetic zeolite A (XZ) diet during the close-up period.Prepartum and postpartum data were analyzed separately.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), treatment by day effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the prepartum period, and a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), treatment by day effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the postpartum period for serum P.There was a treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), grade effect (P < 0.01), and treatment by lactation effect (P < 0.01) during the prepartum, and treatment effect (P < 0.01), day effect (P < 0.01), lactation effect (P < 0.01), and grade effect (P < 0.01) during the postpartum period for salivary P.There was a treatment effect (P < 0.01), and day effect (P < 0.01) during the prepartum period, and a day effect (P < 0.01) and, treatment by day effect (P < 0.01), during the postpartum period for urinary P. a, b, c Treatment by day effects when treatment means differed significantly (P ≤ 0.05).

Frizzarini
et al.: MECHANISMS BY WHICH FEEDING…

2
CON = control diet; -DCAD = negative DCAD diet; XZ = control diet plus synthetic zeolite A; prepartum cows beginning at 254 d of gestation were fed diets.3 T = treatment; D = days of dietary treatment on the first day of sampling.

Figure 5 .
Figure 5. Least squares means and SE for water extractable phosphorus (WEP) (mg/kg) in manure for multiparous cows fed with control (CON), negative DCAD (DCAD), or with supplementation of synthetic zeolite A (XZ) diet during the close-up period.There was a treatment effect (P = 0.01).a, b Treatment effect when treatment means were significantly different (P ≤ 0.05).

Frizzarini
et al.: MECHANISMS BY WHICH FEEDING…

Frizzarini
et al.: MECHANISMS BY WHICH FEEDING…

Table 1 .
Frizzarini et al.: MECHANISMS BY WHICH FEEDING… Dietary ingredient composition of experimental prepartum diets fed to multiparous Holstein cows beginning at 254 d of gestation

Table 2 .
Analyzed nutrient composition (mean ± SE; %DM unless otherwise noted) for experimental prepartum diets fed to multiparous Holstein cows beginning 254 d of gestation 1Samples from each diet were collected weekly for chemical analyses.2 CON = control diet; -DCAD = negative DCAD diet; XZ = control diet plus synthetic zeolite A; prepartum cows beginning at 254 d of gestation were fed diets.

Table 3 .
Least squares means and SE for prepartum minerals intake, fecal excretion, absorption, and digestibility for the multiparous Holstein cows fed experimental prepartum diets beginning 254 d of gestation from which fecal samples were collected 1 Samples collected on 3 consecutive days prepartum, from a subset of 8 cows per treatment.