Epigenetic Inheritance of Maternal Disease Status in Dairy Cattle and Effect on Offspring Performance

Animals and Performance Data: The proposed project will use Holstein at the VTD which houses 175 lactating/dry cows, births 110 female calves, and raises 80 heifers annually.  While this is a teaching dairy, it is operated as a commercial dairy herd that markets milk through a cooperative. Cows average 85 lbs/day and are housed in a free-stall facility. Electronic records of birth, breed, production, breeding, health, and culling are regularly maintained in DairyComp305 and will be used in this study to classify disease status and performance outcomes. The VTD has a 12% incidence of metritis, 6% retained placenta, 13% ketosis, 1% milk fever, 2% displaced abomasum, and 27% incidence of mastitis. All animals in the herd are genotyped and have genomic evaluations available that will be used to compare genetic merit. As part of this study, additional performance data that is not routinely collected by the farm will be recorded with help from our research staff moving forward. This includes monthly weight collections and weekly body condition scores as heifers and calfhood disease incidence such as pneumonia and scours. Data will be collected by farm and research staff as appropriate, and procedures will follow IACUC protocols.

Cohort Definitions and Study Design: This project relies on two methodologies: a retrospective whole-herd record analysis and DNA methylation sequencing on a subset of animals. In both methods, the study subjects will be assigned to two cohorts based on their dam’s incidence of disease in the lactation preceding their birth. Any animal whose dam had at least one recorded incidence of metritis, retained placenta, ketosis, milk fever, displaced abomasum, or mastitis will be assigned to the ‘diseased dam’ cohort and any animal whose dam has no record of any of the previously mentioned or other diseases will be assigned to the ‘healthy dam’ cohort. Due to cohort definitions, only animals born to multiparous dams will be included in the analysis. For the first methodology, differences in first lactation performance and health will be analyzed between cohorts utilizing historic records (Obj. 1).  For the second methodology, a subset of the herd will be enrolled in two cohorts (6 offspring and their respective dams each) and differences in DNA methylation will be analyzed between cohorts (Obj. 2-4). The study design for the second methodology is illustrated in the attached figure.  

Biological Samples and Epigenetic Data: Biological samples will only be utilized for the second methodology of the study. As part of our larger investigation, all female calves born at the VTD and their dams will undergo routine blood sample collections within 24 hours of parturition starting in May 2022. DNA will be extracted from these samples and frozen. This will create a library of samples for the larger project and allow enrollment into this project based on both dam and offspring disease status. For this project, 12 study subjects and their dams (n=12 offspring; 12 dams) will be used. Enrollment will be dependent on both the dam’s history of disease and the offspring’s disease incidence within the first 4 weeks of their 1^st lactation. We will have 3 subjects from ‘diseased dams’ who also had their own disease event, 3 from ‘diseased dams’ without their own disease event, 3 from ‘healthy dams’ with their own disease event, and 3 from ‘healthy dams’ without disease. This sampling strategy is illustrated in the submitted figure. For the enrolled animals, an additional blood sample will be taken 4 weeks into their 1^st lactation (n = 12 offspring) following IACUC protocols. The genome wide methylation profile will be analyzed via reduced representation bisulfite sequencing (RRBS). RRBS is cost effective as only fragments pertaining to CpG-rich regions of the DNA (1-5% of genome, mostly promoter regions) are sequenced. Hyper and hypo methylated regions and differentially methylated regions (DMR) between groups will be identified. Gene ontology and KEGG pathway enrichment analysis will be performed to functionally characterize the differentially methylated regions and genes in terms of what is known about the biological function of genes.

Objective 1: Investigate associations between dam history of disease and offspring’s susceptibility to disease, production, fertility, longevity, and genetic merit using a retrospective study design of the VTD.

Performance Outcomes: For this objective, the whole herd will be split into ‘diseased dam’ and ‘healthy dam’ cohorts to analyze differences in 1^st lactation health and performance of their respective offspring. The presence or absence of any clinical disease and the number of disease events in 1^st lactation will be used as health outcomes in the analysis. Total 1^st lactation milk, fat, and protein production, heifer and 1^st lactation insemination and pregnancy rates, and culling rate and reasons will be used as performance outcomes in the analysis. Additionally, the difference between genomic predicted production and realized first lactation production will be used as an outcome to be compared between cohorts in order to assess if maternal disease status may contribute to phenotypic variability from genomic predictions.

Statistical Analysis: Analyses similar to the Carvalho study will be used to analyze differences in offspring health and performance between cohorts using R version 4.1.1 (Carvalho et al., 2020). ANOVA will be used for continuous outcomes, logistic regression will be used for binary outcomes, and Cox proportional hazard regression will be used for intervals to events. Data such as season, age, and parity of the data will be collected and assessed as covariates in the model.  

Objective 2: Characterize associations between DNA methylation and disease status.    

Dam Disease Comparison: Differences in DNA methylation will be identified between dams with and without a history of clinical disease based on the samples collected at calving. Differentially methylated regions (DMR) between the diseased and non-diseased dams could stem from two possible options.  One, these DMRs could indicate differences existing before the disease incident that contributed to either their susceptibility or resistance to disease or the DMRs could be differences the animals acquired in response to the disease. 

Pre- and Post- Offspring Disease Comparison: In order to determine the source of the DMR found between dams, methylation patterns for individual offspring at birth will be compared to their methylation patterns 4 weeks into their first lactation and to their dams. The peak incidences of all the diseases of interest are within the first 4 weeks of lactation. Therefore, changes in methylation between birth and lactation present only in offspring that had their own disease event, not in offspring that did not have disease, could be considered changes in methylation as a response to disease. In contrast, DNA methylation patterns of offspring at birth in relation to whether they end up having a disease event in their first year of life, suggests DNA methylation provides either resistance or susceptibility to the disease. These DMRs identified in the offspring at birth or as a response to disease will be compared to the patterns found in the dams.

Objective 3: Characterize DNA methylation inheritance with a focus on dam disease status and the association with offspring disease susceptibility.

Offspring Cohort Comparison: DMR in offspring at birth will be identified between cohorts of dam disease status. DMR identified here would represent differences in methylation in offspring due to maternal disease status near gestation. The diseases studied in this project are most common in early lactation meaning they occur before or around the time of insemination. The DMR between the two cohorts of offspring would have either been the result of in-utero programming of the offspring’s methylation pattern due to its own exposure to the disease or inherited from the dam based on her methylation pattern. The analysis for objective 1 investigated the source of the methylation patterns of the dam.

Inheritance Comparison: The DMR found between the two cohorts of offspring will be compared to the DMR between sick and healthy dams. If similar regions are found, this supports that either offspring and dam experienced the same re-programming event as a response to their exposure to the disease or the dam passed her methylation pattern on to the offspring. Without more frequent sampling, it will be hard to determine which scenario is the case, but we will know the outcome relationship between offspring methylation and disease.

Objective 4: Identify differentially methylated regions with putative biological functions relevant to health, growth, production, fertility, and longevity relationships observed

Performance Analysis: A similar performance outcome analysis (excluding disease incidence) as conducted for objective 1, will be conducted for the 12 methylation study subjects. While the power of this analysis will be much lower than the whole-herd analysis, it will be conducted to assess whether the direction of trends is the same in the subset of animals as in the whole-herd. Heifer average daily gain and calfhood disease incidence will be included as additional performance outcomes. This is not something that is routinely collected by the farm but will be collected for the study subjects of this project. Previous studies that investigated the impact of maternal disease status on offspring performance, proposed poor heifer growth as a possible reason for the increased cull rate of heifers born to diseased dams but did not collect the data to investigate that claim (Carvalho et al., 2020).   

Functional Analysis: Genes near the identified DMRs in Objectives 2-4 will be identified. RRBS focuses on methylation patterns near promoters which are regions that impact gene expression. Genes found near DMR could indicate differentially expressed genes between groups. Gene ontology and KEGG pathway enrichment analysis will be conducted to functionally categorize the nearby genes. Gene ontology and KEGG pathway analyses use databases having genes cross-referenced with functional research to provide annotation as to what the biological function is of a gene.  This should reveal genes that have biological functions that support the phenotypic relationships between maternal disease status and offspring health and performance previously identified (in the whole herd and subset).

Methods adjustments (January 2023): Since beginning the study, a few adjustments have been made to the protocol. Health evaluations for calves have been implemented as a way of monitoring calf health between birth and weaning. These are conducted three times a week and include rectal temperatures and eye, ear, nose, cough, fecal, and body condition scores. An additional blood sample timepoint was added at 48 hours after birth. This sample is used for a serum total protein estimation using a digital refractometer. Serum total protein is a good estimator of serum IgG levels which are indicative of the success of passive transfer of immunity through colostrum. As failure of passive transfer could be a confounder for future growth, disease susceptibility and productive outcome, we can now control for this in future models. This change was implemented on 7/1/22. The first 4 calves enrolled in the study do not have serum total protein readings. Due to logistical concerns with the scale, bodyweights are now being estimated using a weigh tape once a week during the time from birth to weaning. This change was implemented on 9/21/22. The 16 calves born since implementation and all future calves will have complete pre-weaning growth information. Enrollment of bull calves has stopped as of 9/14/22.  An updated version of the study design schematic has been added as figure 2.

Methods adjustments (January 2024 Report):

As explained in the previous annual report, we were struggling to enroll multiparous dams with a history of disease at the VTD. To rectify this, we decided to conduct targeted sampling at a different farm, the Cornell University Ruminant Center (CURC). CURC has a larger herd (~500 head) making it easier to find the animals we need. However, this herd requires payment per animal used so we developed a more targeted approach. Figure 3 displays the adjusted study design. In part A, blood samples will be collected from primiparous heifers within the week before calving and then again at 30 DIM with the goal of getting 6 heifers who have a TD event in the first 30 DIM and 6 calving-date matched heifers that do not experience a disease event. All disease diagnosis was done by farm staff and electronically recorded. This will allow us to find DMR in heifers before calving that may be risk indicators for TD by comparing the pre-calving samples from those that got TD to those that did not get TD. We can also identify DMR that may be acquired in the individuals that experience disease by comparing the pre- and post-calving samples of an individual. This supports objective 2 and 4. In part B of Figure 3, 6 cows due with their second calf that had a TD event in their first lactation and 6 due-date and parity matched cows without a history of TD will be identified pre-calving. Blood will be collected from their calves within 48 hours of birth. From these samples we will be able to identify DMR between calves born to dams with a TD history and those born to dams without a history of TD. This applies to objective 3 and 4. Our ability to track these calves’ future progress is limited because we had to enroll male and female calves in order to reach our sample size. All males and some females are sold soon after birth and we will not have the ability to assess their future performance. Of the female calves that were chosen to remain in the herd, we did get an additional blood sample at 48 hours of birth to use for serum total protein analysis, a measure of passive transfer of immunity. We do plan to look at their future performance, but our sample size will be smaller than what we require to draw concrete conclusions.

In the original proposal, we were planning on using reduced representation bisulfite sequencing. We have since decided to do whole genome bisulfite sequencing as the price has reduced since the project proposal was submitted. This will give us whole genome coverage instead of just the CpG-rich regions. This is the gold standard for DNA methylation sequencing.