Effect of feeding rumen-protected lysine through the transition period on postpartum uterine health of dairy cows

Feeding rumen-protected methionine as an indispensable amino acid source has been shown to improve reproductive performance in dairy cows, but the effect of feeding rumen-protected lysine (RPL) during the peri-partum period on reproductive performance is not well explored. Therefore, we aimed to determine the effects of feeding RPL (AjiPro-L Generation 3, Ajinomoto Heartland Inc.) prepartum, postpartum, or both on follicular dynamics, uterine health, and mRNA gene expression of the endometrium. Seventy-five multiparous Holstein cows were assigned to 1 of 2 dietary treatments with or without RPL in a randomized, complete block design. A 2 × 2 factorial arrangement of treatments was used. Prepartum (–28 d to calving), animals were fed a diet (68% of dietary DM from forage) with RPL [PRE-L; 0.54% RPL of dietary dry matter intake] or without RPL (PRE-C). After calving, half of the cows from each prepartum treatment group were assigned to a diet (56% forage) with RPL (PRE-L POST-L; PRE-C POST-L; 0.40% RPL of dietary dry matter intake) or without RPL (PRE-C POST-C; PRE-L POST-C) until 28 d in milk (DIM). Vaginal discharge was detected with a Metricheck device (Simcro) to detect metritis, and at 28 DIM polymorphonuclear leukocytes were evaluated as a percentage of the epithelial cells using a cytology brush (Andwin Scientific) and an endometrial tissue biopsy was collected for mRNA expression and histology. The first postpartum follicular growth cycle was monitored at 7, 10, 11, 13, 15, 17, 19, 21, 23, 25, 27, and 28 DIM via transrectal ultrasonography. Time to first ovulation did not differ between treatments and averaged 18 ± 1.6 DIM. Follicular diameter at first ovulation was not affected by the treatments, but the growth rate of dominant follicle before first ovulation tended to be lower for cows in POST-L in comparison with cows in POST-C. Prevalence of fetid vaginal discharge and metritis did not differ between treatments. Cows in PRE-L POST-L had lower polymorphonuclear leukocytes percentage at 15 and 28 DIM than cows in PRE-L POST-C, PRE-C POST-L, and PRE-C POST-C. Feeding RPL prepartum downregulates the expression of TLR4 , SLC7A6 , EHMT2 , and tends to down-regulate the expression of PTGES3 in uterine tissues at 28 DIM. Additionally, it upregulates the expression of APOL3 and NFKB1 , and tends to upregulate the expression of AHCY and MAT2A . In conclusion, feeding RPL pre-and postpartum improved indicators of uterine immune status, but did not change days to first ovulation postpartum.


INTRODUCTION
Nearly all dairy cows experience at least some degree of negative energy (Drackley, 1999) and protein (Larsen et al., 2014) balance during early postpartum.Coupled with an immunosuppression state (Pascottini and LeBlanc, 2020), this metabolic status is associated with increased risk of uterine diseases among other metabolic disorders (Velázquez et al., 2019).This is partly a result of impaired endometrial function, as a decrease in the energy supply can alter the inflammatory response and increase the risk of uterine diseases (Sheldon and Owens, 2017).Additionally, the mechanisms for adaptation to negative energy balance such as mobilization of fatty acids from adipose tissue may contribute to reduced innate immune function, which increases the risk of reproductive diseases (LeBlanc, 2020).Thus, in this critical period for the dairy cows' productive life, there might be competing demands for nutrients for lactation and for immune response, including AA (Iseri and Klasing, 2014).Lysine (Lys) and methionine (Met) are considered the most limiting AA in dairy cows' diets, particularly in the United States, and research indicates that it should be fed at a ratio of 3:1 to optimize milk production (NRC, 2001).Although focusing on the ratio of Lys to Met could be of practical use when formulating diets, it could lead to deficiencies of these AA when actual DMI does not meet the predicted, such as during the transition period (Vyas and Erdman, 2009).Therefore, quantifying the indispensable AA (IAA) is a more accurate approach, and providing these IAA as a ruminal-protected source improves the duodenal flow of AA (Patton, 2010;Robinson, 2010).For instance, reports indicate increased milk yield, milk protein, and DMI upon supplementation of rumen-protected methionine and rumen-protected lysine (RPL) on Holstein cows' diets (Xu et al., 1998;Socha et al., 2005;Zhou et al., 2016;Batistel et al., 2017).Additionally, greater MP and Lys intake during the precalving period increased DMI postpartum (Girma et al., 2019;Fehlberg et al., 2020).
The reproductive success of dairy cows is associated with multiple factors, such as uterine health, involution and regeneration, and ovarian resumption (Galvão et al., 2004;Chebel et al., 2006;McCoy et al., 2006;Santos et al., 2009;LeBlanc, 2014).Innate immunity is crucial for the health of the reproductive tract of dairy cows following parturition and is affected by AA supply (Zhou et al., 2016;Batistel et al., 2017).Uterine infection is common in the postpartum period and can have a detrimental effect on ovarian and uterine function (Bromfield and Sheldon, 2013).Therefore, improving immune function and reducing the risk of reproductive tract inflammatory diseases could lead to better reproductive outcomes.Uterine infections can also be detrimental to ovarian resumption, because inflammation can affect the first dominant follicle (DF) growth and function through neuroendocrine mechanisms of inhibition of hypothalamic GnRH release and pituitary LH secretion (Williams et al., 2001).Moreover, there is also evidence of direct localized inflammatory mediators, resulting from uterine bacterial contamination after calving, affecting the ovary by suppressing estradiol secretion and decreasing the growth rate of follicles (Sheldon et al., 2002).Additionally, chronic inflammation can result in the disruption of uterine regeneration processes in the early postpartum period (Lucy, 2003;LeBlanc, 2014), which can potentially alter the functional capacity of the uterus (Gray et al., 2001a) and future reproductive efficiency (Gray et al., 2001b).Therefore, ovarian resumption could benefit from modulation of the uterine immune response through nutritional strategies.However, the effects of feeding RPL on the reproductive tract physiology and immune response are still lacking.
Research conducted mainly in monogastric animals provided evidence of the immune system requirements for Lys; for example, Lys consumption by the immune system increased 10-fold in an LPS challenge in poultry (Klasing and Calvert, 1999).Lysine can also play a role in biosynthesis processes, such as the synthesis of acute-phase proteins in response to an increase in circulating cytokines (Iseri and Klasing, 2014) or the synthesis of nonessential AA (Lapierre et al., 2009).These processes are pertinent to and activated during the periparturient period when the immune response of the high-producing dairy cow is activated and the animal is under a state of systemic inflammation (Bradford et al., 2015;Pascottini and LeBlanc, 2020).Though there is limited research in dairy cows relating Lys supply to immune response and inflammatory status, there is evidence of decreased inflammatory response upon supplementation of RPL through the transition period (Fehlberg et al., 2021).The decreased inflammatory response is demonstrated by an increase in negative acute-phase proteins, a decrease in positive acute-phase proteins, and downregulation of IL-1β prepartum and IL-8 and serum amyloid A3 (Fehlberg et al., 2021).
Therefore, our objective was to determine the effects of feeding RPL prepartum, postpartum, or both on uterine health, endometrial morphology, and transcriptional expression of genes related to endometrial metabolism and immunity of multiparous Holstein cows.Additionally, we aimed to evaluate whether supplementation with RPL could affect follicular dynamics of the first follicular wave postpartum.Our experimental design is unique in allowing for the evaluation of the effects of prepartum and postpartum supplementation of RPL separately or the effect of prepartum supply of RPL on postpartum outcomes.We hypothesized that supplementing RPL would improve markers of uterine health due to modulation of uterine metabolism and immune defense system, leading to earlier ovarian resumption.

Animal Care and Housing, and Experimental Design
All experimental procedures were approved by the University of Illinois (Urbana-Champaign) Institutional Animal Care and Use Committee (#18157).Animal handling, experimental design, and diets have been previously described in depth by Fehlberg et al. (2020).Based on the tertiles for expected milk production for each cow through a 305-d lactation (ME305), cows were categorized into low, intermediate, or high ME305, and a similar concept was used for BCS.Eighty-nine multiparous Holstein cows were blocked by lactation number (3.3 ± 1.1), previous 305-d mature-equivalent The number of cows per treatment was calculated for the study reported by Fehlberg et al. (2020) to detect a minimum of 1.1 ± 0.75 kg/d difference in postpartum DMI between groups, assuming a power of 0.9 and a 2-tailed α of 0.05.Additional calculations were made to ensure that the experimental design had the power to detect a minimum of 4 ± 0.5% in endometrial cytology from different independent groups, assuming a power of 0.8 and a 2-tailed α of 0.05.This additional power analysis determined that, to detect such a difference among groups for PMN % evaluations, it would be required a minimum of 16 cows per treatment group, which was met by the experimental design.The exclusion criteria included calving with twins or not having consumed the treatment for at least 16 d during the prepartum period.Four cows were excluded due to twins (PRE-C POST-C, n = 1; PRE-C POST-L, n = 1, PRE-L POST-C, n = 1; PRE-L POST-L, n = 1) and 1 cow was excluded for calving too early (PRE-L POST-C, n = 1).Nine cows were excluded postpartum due to health problems (PRE-C POST-C, n = 1; PRE-C POST-L n = 1; PRE-L POST-C n = 2; and PRE-L POST-L, n = 5).A total of 75 cows concluded the study.
According to the manufacturer, there is 80% rumen bypass and 80% intestinal digestibility to result in 64% bioavailability of this encapsulated RPL product.This would provide 1.4 g of intestinally available Lys prepartum and 1.0 g of intestinally available Lys postpartum per kilogram of DMI (Miura et al., 2017).Top-dress was applied to the TMR once daily following morning feeding.The amount of RPL top-dressed was adjusted daily and for each cow based on their individual DMI of the previous day.Cows that were not receiving RPL were top-dressed with 300g of dried sugarcane molas-ses.Daily DMI was determined for each cow by weighing refusals and total amount fed and determining the difference.Cows were fed for 10% refusals to allow for ad libitum feed intake.All cows had free access to water.Diets (TMR) were formulated using AMTS.Cattle.Pro version 4.7 (2017, AMTS, LLC)

Ultrasonography of Ovarian Structures
The first postpartum follicular growth cycle was monitored at 7,10,11,13,15,17,19,21,23,25,27,and 28 DIM via transrectal ultrasonography (IBEX Pro,7.5-MHzlinear array probe, E.I. Medical Imaging).Patterns of development of the follicles from the first follicular wave postpartum were mapped and evaluated following descriptions adapted from Sirois and Fortune (1988) and reported in Ryan et al. (2020).For every cow, each ovary was scanned several times and in at least more than one plane or direction (lateral to medial direction, medial to lateral direction, dorsoventral, intermediate oblique, and cranio-caudal).All measurements were performed by the same trained person.Ultrasound videos of ovarian structures were recorded to allow measurement of the follicles.Follicular structures from both ovaries were and patterns of development were characterized by follicular mapping.Diameter of follicles were measured through ImageJ software (version 1.47, National Institutes of Health), using a software tool calibrated against a scale provided by the ultrasound unit.Measurements of all follicles ≥5 mm in diameter were recorded and a DF was defined as a follicle >10 mm in diameter in the absence of other follicles (Savio et al., 1990;Beam and Butler, 1997).From there, only the DF was mapped until its disappearance or until 28 DIM, the last day for ultra-sound measurements.Ovulation was classified as the disappearance of the previously identified DF and the appearance of a corpus luteum (CL) in the subsequent examinations.Growth rate was calculated as the difference in diameter from the last measurement by the first measurement, divided by the difference in days from those 2 measurements.A cow was considered to have a follicular cyst if the ovarian structure had a thin walled (≤3 mm) round, anechoic density consistent with the presence of an antrum >25 mm in diameter, in the absence of a CL (Hamilton et al., 1995;Garverick, 1997).By the end of the experiment, a CL was not detected in 25 cows.From these, 14 cows developed follicular cysts

Vaginal Discharge Evaluation
Evaluations of vaginal discharge were performed at 4, 7, 10, 13, 15, and 17 DIM.These days were chosen because although technically described as occurring at any time within 21 DIM (Sheldon et al., 2006;LeBlanc et al., 2011), the majority of metritis cases occur in the first 14 DIM, peaking at around 5 to 7 DIM (Galvão et al., 2011).The evaluation was performed using the Metricheck device (MC, Simcro) following guidelines presented in LeBlanc and Bicalho (2017).The device was composed of a 50-cm-long stainless-steel rod with a 4-cm rubber hemisphere to collect vaginal contents.The MC was disinfected with chlorhexidine diacetate (Nolvasan Solution, Zoetis Animal Health) before and after each use in a single cow.The evaluation began with cleaning the perineal region of the cow with a paper towel and disinfectant solution.The tail was moved to the side and the MC was inserted into the vaginal canal until the cervix was reached.The device was then retracted and removed from the reproductive tract with the vaginal contents remaining in the rubber hemisphere.With the vaginal content in the rubber hemisphere, evaluation of smell was scored (smell 0 = no odor or smell 3 = fetid odor).The vaginal content was then poured onto a paper towel for examination and scored on a scale of 0 to 3: score 0 = clear or translucent mucus; score 1 = mucus containing flecks of white or off-white pus; score 2 = discharge containing ≤50% white or off-white mucopurulent material; and score 3 = discharge containing ≥50% purulent material, which may be white, yellow, or sanguineous (Sheldon et al., 2006).Cows were classified as having metritis if the MC score plus smell was equal or exceeded 3, meaning a fetid odor was detected along mucus-or mucopurulent characteristic of the discharge.

Cytology of the Uterine Endometrium
Cytology of the endometrium was performed using a cytology brush (Andwin Scientific) at 28 DIM.The sterile cytology brush was mounted to a sterile stainless-steel rod and inserted into a larger sterile stainless-steel rod (SSR) covered with a plastic sleeve for passage through the cervix and into the uterine body without contamination.Before the procedure, the cow was restrained, and the vulva was cleaned with water and 70% ethanol.After passage of the cytology rod through the first ring of the cervix, the SSR was exposed through the plastic sleeve and was advanced into the uterine body.Once inside the uterine body, the outer SSR was pulled back to expose the cytology brush.The SSR that was mounted to the cytology brush was then rotated 3 times while the cytology brush remained in contact with the endometrium.Finally, the cytology brush was retracted back into the outer SSR and removed from the reproductive tract.The SSR were washed and autoclaved between each day of use.If multiple samples were being taken within a single day, the SSR were sanitized in a chlorhexidine diacetate disinfectant solution between each animal.Cytology slides were prepared immediately by rolling the cytology brush onto a clean glass microscope slide and fixed using a cytology fixative (Cytoprep, Fisher Scientific).Once the fixative was dry, the samples were transported to the laboratory where they were stained (Camco Quik Stain 2 -Self Buffered Differential Wright-Giemsa Stain).After being allowed to dry for 24 h, the slides were covered using mounting medium (Permount, Fisher Scientific) and dried for at least 48 h before being scanned.Following guidelines described in Stella et al. (2018), all slides were scanned at the Institute for Genomic Biology at the University of Illinois with 20× magnification using whole slide imaging (Nanozoomer Digital Pathology System, Hamamatsu Photonics).Five areas were captured at 20× magnification from 5 separate locations, one image from each corner of the sample area of the slide and one image from the center, to represent the entire the slide (NDP.view2 software, ver.2.9.29,Hamamatsu Photonics).A minimum of 100 cells were manually counted using the software ImageJ (National Institutes of Health) to determine the percentage of PMN [PMN/(PMN + epithelial cells)].Cell counting was performed by the same technician for all samples.Seven samples were excluded due to staining issues (PRE-L POST-L n = 1; PRE-C POST-L n = 2; PRE-C POST-C n = 4).

Endometrial Biopsy
Endometrial tissue samples were collected transcervically from the body of the uterus at 28 ± 2 DIM.Biopsy samples were harvested for histological evaluation of uterine glands and expression of transcripts in the endometrial tissue.Uterine biopsy was not performed on 3 cows (PRE-L POST-C = 1, PRE-C POST-L = 1, and PRE-C POST-C = 1) due to MC score plus smell >3.The biopsy instrument (48 cm in length; 2 cm diameter; Aries Surgical) was covered with a sanitary disposable sleeve and inserted into the vagina.The biopsy forceps were positioned at the cervical opening, and the sleeve was retracted over the instrument.The exposed biopsy forceps were then threaded through the cervix and into the uterine body.Endometrial tissue was collected at a location approximately 1 cm beyond the end of the cervix.A subset of the sample was flashfrozen in liquid nitrogen for gene expression analysis.The other subset of the sample was placed into PBS containing 4% paraformaldehyde for 24 h.The samples were then set in a block of paraffin wax at the University of Illinois Veterinary Diagnostic Laboratory for hematoxylin and eosin staining.

Hematoxylin and Eosin Stain for Endometrial Gland Analysis.
After whole slide scanning, individual gland structures were labeled and images were captured (NDP.view software, Hamamatsu Photonics).Total glandular area and perimeter, glandular epithelial height, number of glandular epithelial cells, and number of glands per tissue sample were manually measured through ImageJ software (version 1.47, National Institutes of Health).The same trained technician obtained all glandular measurements.Twenty-two samples were excluded due to incorrect processing (PRE-L POST-

RNA Extraction and Real Time Quantitative PCR
The extraction and quantitative PCR analysis were performed using previously established protocols from the Mammalian NutriPhysioGenomics Laboratory at the University of Illinois (Vailati-Riboni et al., 2015).All evaluated transcripts and primer information are reported in the supplemental materials (Supplemental Tables S1, S2, and S3; https: / / uofi .box.com/s/ rhs3dxrlk0niswq3fjrkqz5key8fz2nu).
Primers were chosen based on previous research demonstrating the expression of such targets involved in inflammatory and metabolic processes in the bovine reproductive tract (Brewer et al., 2020;Guadagnin et al., 2021).To control analytical and tissue sampling variation, the final data were normalized to expression of the geometric mean of GAPHD, ACTB, and H2AFZ, which were validated as suitable internal control transcripts in bovine uterine tissue (Cerri et al., 2012;Gómez et al., 2017).The change in cycle threshold (CT) for each gene was calculated following guidelines reported by Schmittgen and Livak (2008), by subtracting the geometric mean of the CT of GAPDH, ACTB, and H2AFZ from the CT of the gene of interest.Fold change was calculated using the 2 −ΔΔCT method, as described by Schmittgen and Livak (2008).
For RNA extraction from endometrial cells, the Direct-zol RNA Miniprep system (Zymo Research) was used following the manufacturer's protocols.To start, 600 μL of TRI-Reagent was added to uterine tissue and homogenized completely; 600 μL of 100% ethanol was directly added to the solution and homogenized.The sample was added to a Zymo-Spin IIC column (Zymo Research) with the collection tube and centrifuged.The column was moved to a new collection tube and the collection tube containing the filtrate was discarded.The RNA samples were treated with DNase.Next, 400 μL of prewash Direct-zol RNA (Zymo Research) was added to the column and centrifuged.Subsequently, the filtrate was discarded, and this step was repeated; 700 μL RNA wash buffer was added to the column, and it was centrifuged.The column was carefully transferred from the collection tube into the RNase-free tube, 20 μL DEPC-treated nuclease-free water was added to it (Life Technologies), and it was centrifuged.One and a half microliters of solution was used to measure the concentration of RNA in Nano-Drop ND-1000 (NanoDrop Technologies).Complementary DNA was synthesized using 100 ng of RNA.First, random primers (10 mM; Invitrogen Corp.) and DNase-and RNase-free water were mixed and incubated at 65°C for 5 min and kept on ice for 3 min.Then a second mix containing DNase-and RNase-free water, first strand buffer (5%), oligo dT18 (Operon Biotechnologies), dNTP mix (10 mM; Invitrogen Corp.), Re-vertAid Reverse Transcriptase (200 U/mL; Fermentas Inc.), and RNase Inhibitor (20 U/mL; Promega) was added.The reaction was performed in an Eppendorf Mastercycler (Eppendorf North America) using the following temperature program: 25°C for 5 min, 42°C for 60 min, and 70°C for 5 min.The cDNA was then diluted 1:3 with DNase-and RNase-free water.Sample concentration was measured using the NanoDrop ND-1000 (NanoDrop Technologies), and RNA quality was measured using an Agilent 2100 Bioanalyzer (Agilent).Samples used in the analysis had a mean RNA integrity number of 7.1 ± 0.4.
Messenger RNA expression was analyzed on Quant-Studio 7 Flex PCR system (Applied Biosystem) using primers designed using Primer Express 2.0 with minimum amplicon size of 80 bp (when possible, amplicons of 100 to 120 bp were chosen) and limited 30 G þ C (Applied Biosystems).When possible, primer sets were designed to fall across exon-exon junctions.Primers were aligned against publicly available databases using BLASTN (Nucleotide BLAST, 2008;

Statistical Analyses
Statistical analysis was performed using SAS 9.4 (SAS Institute Inc.).The MIXED procedure of SAS was used to conduct statistical analysis.The model included treatment with RPL prepartum (PRE) or not, RPL postpartum (POST) or not, and PRE × POST interaction.The following mixed linear regression model was used: where Y jklm = the observations for dependent variables; μ = the overall mean; B j = the fixed effect of block; C k = the fixed effect of PRE; D l = the fixed effect of POST; (CD) kl = the interaction of PRE and POST; A m = the random effect of cow; and ε jklm = the random residual error.A log-transformation for the variables glandular area, number of cells per gland, and diameter of DF at first measurement was performed because residuals were not normally distributed and with homogeneous variance.Data presented for these variables were back transformed.The Kenward-Roger degrees of freedom approximation was used to determine the denominator degrees of freedom for tests of fixed effects (Littell, 2002).Statistical analysis of transcript expression was performed after the 2 −ΔCT transformation was calculated to obtain the mean ± standard deviation, following guidelines reported by Schmittgen and Livak (2008).
Endometritis was classified if the proportion of PMN at 15 DIM was greater than 40% (Ryan et al., 2020) and at 28 DIM if the proportion was greater than 18% (Sheldon et al., 2006).Metritis was classified as MC score plus the MC smell ≥3 (Sheldon et al., 2006;Skenandore et al., 2017).Fetid vaginal discharge was classified as MC smell = 3, independently of the MC score.
Endometritis at 28 DIM, metritis, and fetid vaginal discharge were analyzed as binary traits for the determination of their prevalence.A model for development of CL by 28 DIM was evaluated by logistic regression using a binomial distribution in the GLIMMIX procedure of SAS.Odds ratio was used to compare the likelihood of 2 treatments to incur in the event (development of a CL by 28 DIM).Association between treatments and time to first ovulation was assessed using Kaplan-Meier curves and Cox's proportional hazard regression.Treatments and block were forced into models, with cow was considered as a random effect.Statistical difference was considered at P ≤ 0.05 and trends at 0.05 > P ≤ 0.10.

Follicular Dynamics
Follicular dynamics data are in Table 1.Cows that received RPL postpartum tended to have smaller growth rate of the first DF than cows that did not receive RPL postpartum.
Treatments did not differ in the likelihood of observing a follicular cyst at 28 DIM (P = 0.23).Cows in PRE-C POST-L were more likely (P = 0.04, OR =  11/22).Days to first ovulation were 18.5 ± 1.64 for PRE-L POST-L, 16.8 ± 1.09 for PRE-L POST-C, 17.9 ± 1.51 for PRE-C POST-L, and 17.7 ± 1.26 for PRE-C POST-C (P = 0.39).A total of 27.8% (5/18) of PRE-L POST-L, 55.0% (11/20) of the PRE-L POST-C, 15.8% (3/19) of the PRE-C POST-L, and 33.3% (6/18) of the PRE-C POST-C were treated as right censored because they had not ovulated by or had a follicular cyst at 28 DIM.Kaplan-Meier survival curves showed no difference (P = 0.28) between treatments in time to first ovulation (Figure 1).

Vaginal Discharge, Cytology of the Uterine Endometrium, and Morphology of Uterine Glands
The prevalence of fetid vaginal discharge (MC smell = 3) did not differ and were 11.For the main effects of RPL prepartum (PRE), RPL postpartum (POST), and their interaction.lesser PMN % at 28 DIM (P < 0.01) than cows in PRE-L POST-C (15.8 ± 3.68%) and PRE-C POST-C (17.2 ± 3.68%), whereas cows in PRE-C POST-L did not statistically differ from the rest (12.9 ± 3.68%; Table 2).
The morphology of the uterine glands was assessed through histology and is described in Table 2. Treatments did not affect the number of uterine glands, the glandular area or perimeter (P = 0.20).However, there was a tendency for a treatment effect of POST (P = 0.09) on the number of cells per gland, with cows that were fed RPL postpartum having more cells per gland (68.4 ± 7.68 cells) than cows that did not receive RPL postpartum (55.1 ± 7.68 cells).The same tendency for an effect of POST (P = 0.06) was observed when the glandular epithelial height was evaluated, but in this case cows that did not receive RPL postpartum tended to have greater epithelial cell height (8.44 ± 0.19 μm) than cows that received RPL postpartum (7.90 ± 0.18 μm).

Expression of mRNA Transcripts from Uterine Samples
The mRNA expression of measured transcripts is reported in Figure 2. Feeding RPL prepartum downregulated the expression of TLR4, SLC7A6, EHMT2 (P ≤ 0.05), and tended to downregulate the expression of PTGES3 (P = 0.06) in uterine tissues at 28 DIM; additionally, it upregulated the expression of APOL3 and NFKB1 (P = 0.04), and tended to upregulate the expression of AHCY (P = 0.08) and MAT2A (P = 0.07).When fed postpartum, RPL upregulated the mRNA expression of MUC1 (P = 0.04), tended to upregulate MUC4 (P = 0.06), and tended to downregulate SOD1 (P = 0.07) mRNA transcript.There was a PRE × POST interaction effect for the FGF10 mRNA transcript, with a downregulation for cows in PRE-C POST-L, PRE-L POST-C; and POST-L PRE-L.Additionally, there was a tendency for a PRE × POST interaction effect for the expression of HGF, where cows in PRE-C POST-L, PRE-L POST-C, and PRE-L POST-L had a downregulation of HGF mRNA transcript in comparison with cows in PRE-C POST-C.

Follicular Dynamics
The first ovulation following parturition is associated with fertility, as multiple estrus cycles before the first artificial insemination result in a greater probability of pregnancy (Butler, 2003).Although the importance of the DF size at ovulation and its effect on fertility are well-understood (Vasconcellos et al., 2001), the growth rate of the DF is less studied.For example, in one of few studies relating the growth rate of DF in dairy cows with their nutrition, the supplementation of Met downregulated pro-inflammatory transcripts in follicular cells of the first DF postpartum but did not affect Means within a row with different superscripts differ (P ≤ 0.05). 1 Dietary treatments included at top-dress with rumen-protected Lys (RPL) prepartum and postpartum (PRE-L POST-L), with RPL prepartum and without RPL postpartum (PRE-L POST-C), without RPL prepartum and with RPL postpartum (PRE-C POST-L), and without RPL prepartum and postpartum (PRE-C POST-C) in a carrier of 300 g of dried molasses.
2 Greatest value of SEM within treatment.
3 Consists of the main effect of RPL prepartum (PRE), the main effect of RPL postpartum (POST), and their interactions.time to first of ovulation or growth rate of the first DF (Acosta et al., 2017).In our study, cows that did not receive RPL postpartum tended to have a greater growth rate of DF than cows that received RPL postpartum.The growth of the DF depends on endocrine stimuli and after deviation, the DF relies more on LH than on FSH for growth (Aerts and Bols, 2010).Changes in the endometrium because of infection and inflammation disturb this endocrine function (Zerbe et al., 2002;Sheldon et al., 2006), as GnRH and LH release are impaired as a consequence of uterine inflammation (Sheldon et al., 2009).Nevertheless, because no differ-ence was evident regarding the follicle size at ovulation time when RPL was fed or not, we suggest that this difference in growth rate might be of limited biological significance.

Vaginal Discharge, Cytology of the Uterine Endometrium, and Morphology of Uterine Glands
Cows in PRE-L POST-L had lesser PMN percentage at 28 DIM; however, there was no difference in the mRNA expression of IL1β, IL6, and IL8.Furthermore, a lesser PMN percentage in the uterus at the fourth week postpartum was previously associated with improved uterine immunity (Stella et al., 2018).A possible factor contributing to the lesser PMN % in the uterus of cows in PRE-L POST-L could be the contribution of Lys to production of natural antimicrobial peptides (Li et al., 2020).Antimicrobial peptides have a cationic characteristic which is conferred by AA residues, particularly Lys and Arg (Guaní-Guerra et al., 2010).In the present study, feeding RPL prepartum tended to increase DMI postpartum (Fehlberg et al., 2020).Thus, we suggest that feeding RPL through the transition period had a positive but indirect effect on the immune response in the early postpartum period, through the increase of DMI.
Uterine glands are crucial for fertility, as demonstrates in Gray et al. (2001b), which reported failure in the conceptus development at early stages of pregnancy in uterine gland knockout sheep.Adenogenesis, or the process of uterine gland development, can be affected by nutritional factors, such as insufficient colostrum ingestion (Bartol et al., 2013;Vallet et al., 2013).Huang et al. (2012) investigated the cellular mechanisms of regeneration of the endometrium at different time points of the postpartum period of wild-type mice and detected cell proliferation occurring in the glandular epithelium.However, cell death signals were only identified in luminal epithelium and stroma (Huang et al., 2012).This implies regional differences in cell proliferation and death during the postpartum period in mice.Safeguarding the obvious limitations in drawing a comparison between studies using different species, it is possible for this mechanism to be true for the bovine uterine epithelium as well.Thus, a greater number of cells per uterine gland as a response to RPL feeding could be indicative of greater proliferation of the glandular epithelia.This is endorsed by the modulation in the expression of uterine transcripts, such as HGF and FGF10, which will be discussed subsequently.Changes in uterine glandular epithelia occur through the estrus cycle but also with inflammation, which results in glandular atrophy (Ohtani et al., 1993).Cows that received RPL postpartum had greater numbers of cells per uterine gland, which indicates cell proliferation.Cell proliferation is part of the tissue healing process.Uterine involution can be similar to wound healing, as it also comprises a dynamic process of inflammation, cell proliferation, and tissue remodeling (Eming et al., 2007).Li et al. (2020) reported that Lys-derived carbon quantum dots (CDQ) promoted growth of typical mammalian cells [NIH 3T3 (highly contact-inhibited cell line developed from NIH Swiss mouse embryo cultures) and red blood cells] and when injected into infected wounds in mice, they accelerated wound heal-ing.The authors suggested that Lys-CQD may have signaling agents that promote cell proliferation, which could be related to intracellular oxidative damage (Li et al., 2020).One possible explanation is that Lys may contribute to mitogenic signaling activity, promoting cell proliferation.However, further research is needed to elucidate this action.

Expression of mRNA Transcripts from Uterine Samples
Overall, our results indicate downregulation of transcripts involved in inflammatory processes at 28 DIM for cows that were fed RPL in comparison with those that were not.Cows with uterine inflammation have greater expression of toll-like receptor (TLR) transcripts (Herath et al., 2006;Gabler et al., 2010;Kasimanickam et al., 2014), which corroborates our findings.Furthermore, the upregulation of AHCY transcript, along downregulation of SOD1, indicates increased supply of homocysteine that is probably being used for antioxidant synthesis.This was also previously reported by Osorio et al. (2014), which observed increased hepatic mRNA abundance of AHCY and decreased abundance of SOD1 in dairy cows' liver upon supplementation with Met.The upregulation of MAT2A would support the supply of S-adenosylmethionine, which links methionine and one-carbon metabolism to mechanistic target of rapamycin complex 1 (mTORC1) through SAMTOR protein (Gu et al., 2017;Coleman et al., 2020).Thus, the upregulation of AHCY and MAT2A transcripts suggest that feeding RPL also affects Met metabolism in uterine tissue.The mechanism (or mechanisms) responsible for the effect on increased supply of RPL on methionine metabolism is not clear.One possible explanation is the theory behind limiting AA, where the protein synthesis depends on the supply of the most limiting AA and the efficiency of its use (Wolfe, 2017).Thus, increasing the supply of intestinally available Lys could indirectly affect Met metabolism and protein synthesis.However, further research is needed to confirm this hypothesis.Chapwanya et al. (2009), when evaluating gene expression of uterine tissue from cows at 2 wk postpartum, reported an increase in the expression of NFKB1 and TLR4 in comparison with cows in late postpartum (9 wk).Additionally, the expression of NFKB1 was related to the level of inflammation of the uterine tissue at 2 wk postpartum, which was classified as mild, moderate or severe according to uterine cytology and histological characteristics.With this, the authors proposed that the predominant uterine immune response in the early postpartum period in dairy cows was mediated by TLR/ NFκB.However, recent studies in mice suggest a different role and function for NFKB1.Best et al. (2019) reported that the NFKB1 deletion in mice increased NFκB activation and signaling during tendon healing, leading to increased macrophage recruitment and general inflammation.The transcription factor NFκB is involved in innate and adaptive immune responses and is a mediator of inflammation (Liu et al., 2017).The canonical pathway is one of the activation routes of the NFκB pathway and is activated upon cellular exposure to inflammatory cytokines (such as IL-1β) or in response to LPS, upon its binding to TLR (Perkins, 2007;Didonato et al., 2012).The NFKB1, however, encodes a protein that is a transcription inhibitor, being a suppressor of inflammation (Cartwright et al., 2016).Thus, the upregulation of NFKB1 in cows that received RPL prepartum is consistent with the downregulation of TLR4 for the same animals.
Moreover, the upregulation of MUC1 in the uteri of cows that received RPL postpartum is consistent with the upregulation of NFKB1, as it was previously reported that MUC1 increases the phosphorylation and degradation of IκBα (nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, α), thus blocking apoptosis process (Ahmad et al., 2007).Both MUC1 and MUC4 contribute to cell proliferation, through activation of extracellular signal-regulated kinases 1 and 2 in the case of MUC1 (Schroeder et al., 2001); and activation of growth factor receptors in the case of MUC4 (Bafna et al., 2010).Both these transcripts were upregulated for cows that were fed RPL postpartum, which reflected in greater cell proliferation, as exemplified by the tendency for greater number of cells per uterine gland in cows that received RPL postpartum.The upregulation of APOL3 in dairy cows fed RPL prepartum can also point out to a lesser inflammatory environment in their uteri, because it was proposed that the expression of apolipoprotein L family transcripts are negatively correlated with inflammation (Akl et al., 2017).This would be due to the involvement of these apolipoproteins in apoptotic processes, but neutrophils actually tend to increase their life span from the onset of inflammation (Kolaczkowska and Kubes, 2013).Moreover, cows in PRE-L POST-L had lesser percentage of PMN in the uterus at 28 DIM, which support this premise.
The EHMT2 gene encodes the histone methyltransferase enzyme, also known as G9a (Lee et al., 2006).Its role and function, although still not completely understood, has been more studied in mice models than in ruminants, thus we cautiously tried to find a possible explanation for the downregulation of EHMT2 transcript at 28 DIM in the uterus of dairy cows fed RPL during prepartum.Antignano et al. (2014), when studying murine intestinal inflammation, reported that the expression of EHMT2 is necessary for development of intestinal inflammation.This happens through regulation of T helper-17 and regulatory T cells responses, by an increase differentiation toward T helper-17 response, which promotes inflammation through production of cytokines and recruitment of neutrophils (Littman and Rudensky, 2010).This immune response begins with the recognition of pathogen-associated molecular patterns by receptors, such as TLR (Janeway and Medzhitov, 2002).Thus, it is possible that the downregulation of EHMT2 is following the downregulation of TLR4 in the uterus of dairy cows that were fed RPL prepartum.
The HGF transcripts were downregulated in the uterus of cows in PRE-L POST-L and PRE-L POST-C, but upregulated in those of cows in PRE-C POST-L, when comparing with cows in PRE-C POST-C.This growth factor is involved in cell proliferation and morphogenic activity of several epithelial cells, including endometrial cells (Barros et al., 1995).Yoshida et al. (2004) reported a modest reduction in growth of endometrial stromal cells cultured in serum-free medium with supplementation of anti-HGF antibody, suggesting a role of endogenous HGF in stimulating endometrial stromal cell proliferation.Additionally, HGF stimulated epithelial morphogenesis in the ovine uterus (Chen et al., 2000a).Thus, and if the same stimulation also happens in glandular cells, it is possible that HGF was involved in the morphological alterations of cells from uterine glands observed here.Fibroblast growth factor 10, as well as HGF, are considered stromal-derived growth factors and their expression were reported in the female ovine reproductive tract (Chen et al., 2000a,b).The expression of FGF10 transcript in our study follows the expression of HGF transcript, which corroborate previous studies that reported a role of these growth factors in epithelial growth and differentiation of endometrium (Rubin et al., 1995;Chen et al., 2000a,b).

CONCLUSIONS
Feeding RPL around parturition altered the expression of transcripts involved in inflammatory and immune responses.The downregulation of TLR4, PT-GES3, SOD1, and EHMT2; and the upregulation of APOL3, NFKB1, MUC1, and MUC4, in conjunction with the lesser uterine PMN percentage, are indicatives of a potentially less severe inflammatory process by wk 4 postpartum.Additionally, a stimulus of cell proliferation is suggested by the tendency of RPL to increase the number of glandular epithelial cells.There was no effect of feeding RPL on the size of the first ovulatory follicle nor days to first ovulation.Increasing intestinal availability of Lys throughout the transition period improved several indicators of uterine health.
Guadagnin et al.:  RUMEN-PROTECTED LYSINE AND IMMUNITY milk production (11,363 ± 1,860 kg), expected calving date, and BCS during the far-off period (3.76 ± 0.84).Each block had 4 cows in it, with exception one block that contained 6 cows.Cows were then assigned to 1 of 2 dietary treatments [TMR with or without RPL (AjiPro-L Generation 3; 42% L-Lys-HCl; Ajinomoto Heartland Inc.)] for the prepartum period.Prepartum (−28 d to calving), animals were fed a diet with RPL [PRE-L (n = 38); 0.54% RPL of dietary DMI] or without RPL [PRE-C (n = 37)] top-dressed in a carrier of 300 g of dried sugarcane molasses.After calving, half of the cows from each prepartum treatment group were assigned to a diet with RPL [PRE-L POST-L (n = 18); PRE-C POST-L (n = 19); 0.40% RPL of dietary DMI] or a diet without RPL [PRE-C POST-C (n = 18); PRE-L POST-C (n = 20)] until 28 DIM, in a randomized, complete block design, using a 2 × 2 factorial arrangement of treatments.Cows not fed with RPL received 300 g of dried sugarcane molasses only.

Figure 1 .
Figure 1.Survival curves for dietary treatments included as topdress with rumen-protected Lys (RPL) before calving (PRE) and after calving (POST) across days to first ovulation.Treatments consisted of a crossover design by RPL inclusion PRE or POST in which cows consumed a top-dress either with (L) or without (C) RPL for 4 wk prepartum and for 4 wk postpartum [with RPL prepartum and postpartum (PRE-L POST-L), with RPL prepartum and without RPL postpartum (PRE-L POST-C), without RPL prepartum and with RPL postpartum (PRE-C POST-L), and without RPL prepartum and postpartum (PRE-C POST-C)] in a carrier of 300 g of dried molasses.The y-axis represents the overall probabilities.The x-axis represents the number of days postcalving.Log-rank test indicates no differences in the survival experiences observed between treatments (P = 0.28).
ftp: / / ftp .ncbi.nlm.nih.gov/blast/ db/ ) at National Center for Biotechnology Information and the University of California Santa Cruz Cow (Bos taurus) Genome

Table 1 .
Least squares means and associated SEM for the ovulation dynamics of the dominant follicle of the first follicular wave 2Greatest value of SEM among treatments.3

Table 2 .
Guadagnin et al.:RUMEN-PROTECTED LYSINE AND IMMUNITY Least squares means and associated SEM for the proportion of polymorphonuclear cells (PMN) in the uterus and for glandular morphology of uterine endometrial tissue samples from Holstein cows Guadagnin et al.: RUMEN-PROTECTED LYSINE AND IMMUNITY