Performing Department
Molecular Biomedical Sciences
Non Technical Summary
Porcine reproductive and respiratory syndrome (PRRS) is an endemic viral disease impacting pork production worldwide. PRRS results in decreased reproductive performance in sows and reduced growth performance in finisher pigs. The economic impact of PRRSV on the US swine industry alone is $664 million per year (Holtkamp et al., 2013). Efforts to control or prevent PRRSV infection have been largely unsuccessful. Challenges include the repeated failure of biosecurity measures and the limited success of vaccines due to rapid mutation of the virus. Hence, PRRS outbreaks are an unavoidable reality faced by the pork industry. Alternative or complementary control strategies are sorely needed to address this challenge. Our long-term goal is to establish a quantitative framework within which genetic resistance to Porcine Reproductive Respiratory Syndrome virus (PRRSV) can be improved in reproductive sows using antibody immune response as an indicator trait for selection.One approach to the challenge of endemic infections is selective breeding for increased host resistance to infection or disease. Recent studies indicate that selection for improved performance under PRRSV infection may be feasible in sows. However, the success of genetic selection depends on both a definition of the resistance phenotype and the identification of heritable traits associated with disease resistance. While growth is a moderately heritable outcome under the selection pressure of PRRSV, reproductive performance has moderate to low heritability, ranging from 1 to 10%, in sows during a PRRS outbreak (Serão et al., 2014a). Thus, genetic progress for these traits is slow. Fortunately, Serão et al. also observed that antibody response to PRRSV was both highly heritable and predictive of reproductive performance. PRRSV sample-to-positive (S/P) ratio, a semi-quantitative measure of PRRSV-specific antibody, had high heritability (45%) and a high positive genetic correlation (0.72) with number of piglets born alive. Furthermore, Serão et al. (2014b, 2015a) showed that genomic information is predictive of S/P ratio in PRRSV-infected or vaccinated gilts across swine populations. Intriguingly, this appeared to be true regardless of whether antibody expression was a consequence of wild-type PRRSV infection or PRRS vaccination.We propose to overcome the limits associated with selecting directly for a desirable trait with low heritability by pursuing genetic selection through a correlated trait exhibiting high heritability. Specifically, we propose that S/P ratio, a measure of immune response, can be used as a genetic indicator trait to select for reproductive performance during PRRSV infection. Further, we posit that controlled exposure to PRRSV via vaccination is sufficient to allow for immune-response based genetic selection. The huge impact of PRRS on local industry and the as yet intractable nature of this widespread and rapidly evolving pathogen motivate our interdisciplinary, integrated computational and experimental research effort.
Animal Health Component
(N/A)
Research Effort Categories
Basic
100%
Applied
(N/A)
Developmental
(N/A)
Goals / Objectives
Aim 1: To compare the immune response profiles of PRRS-exposed pregnant gilts subjected to vaccination vs. infection, we will vaccinate and/or infect pregnant commercial naïve gilts sourced from a commercial multiplier and collect blood samples across time to quantify viral load, antibody response, and cytokine levels. This data will allow testing of our hypothesis that immune response to PRRS vaccination and infection elicit the same resistance phenotype. The data also provides for model development and validation in Aim 3. The four planned exposure groups will include (1) a control, (2) vaccination, (3)infection and (4) vaccination with subsequent heterologous infection.Aim 2: To characterize the genetic response to vaccination and infection, we will quantify total gene expression in the blood of gilts during PRRS exposure. This will allow us to (1) characterize the transcriptomic response to PRRS in pregnant gilts, (2) assess the differential expression of genes between the PRRS exposure groups, and (3) identify the pathways controlling genetic response to PRRS. This data will allow testing the hypothesis that response to PRRS vaccination and infection are genetically the same traits.Aim 3: To utilize a novel mathematical model to assess variability in immune response as dictated by host-pathogen interactions, we will develop an integrative model of host-pathogen interactions in PRRSV infection. The model will explicitly address the dynamic interplay of infection, viral replication, cytokine expression, and humoral and cellular adaptive responses. The model will allow us to interrogate the time-dependence of the resistance phenotype and identify associated causative mechanisms. Differential innate and adaptive responses to vaccination vs. infection will be used to interrogate: (1) whether vaccination is likely to be adequate to stimulate a significant immune response in the context of a heterologous infection with variable virulence; (2) critical timing for genetic evaluation; (3) perturbations of immune processes likely to be responsible for delayed adaptive and innate responses.
Project Methods
We will use crossbred naïve gilts sourced from a commercial multiplier in NC. The study will be performed at NC Swine Evaluation Center (NCSEC) at Clayton, NC. At arrival, animals will be weighed, ear tagged, and randomly assigned to one of 4 rooms. Animals will have 40 days of acclimation to mimic standard commercial acclimation procedures. On day 21 after entry, gilts will be estrus-stimulated and synchronized as outlined in Ladinig et al. (2014). After acclimation, all animals will be weighed and artificially inseminated (day 0 of gestation). Pregnancy will be confirmed ultrasonically. The study will start with 52 gilts (13/group) with a target of 40 gilts successfully impregnated. Animals kept in the study will be individually penned with free access to water and standard gestation diet. Body weight and body condition scores of gilts will be recorded at entry at the farm, after acclimation, and at weaning. Gilts will be monitored periodically for PRRS infection signs and for any other abnormalities. At farrowing, the number of piglets born alive and born dead will be recorded. Piglets will be weighed at birth and at weaning. Weaning mortality will be recorded. The study will be performed from February 2017 to August 2017 to allow for farrowing during the summer.Biosecurity: Careful strategies will be employed to avoid cross-contamination between vaccinated vs. infected groups. The proper feeding and sampling strategies have been discussed with personnel from NCSEC. Strategies include: isolation of rooms, feeding animals based on the exposure groups, use of one scale per exposure group, etc. The center has previously hosted disease challenge studies of this nature.Aim 1: To compare the immune response profiles of PRRS-exposed pregnant gilts subjected to vaccination vs. infection, we will vaccinate and/or infect pregnant naïve gilts and collect blood samples across time to quantify viral load, total and neutralizing antibody response, and cytokine levels. Animals in each room at the NCSEC will be subject to one of four PRRS-exposure strategies:CTRL: Control group - pregnant gilts will be injected with saline.VX75: Vaccination with a commercial PRRS vaccine - Animals will be vaccinated with a standard commercial modified-live virus (MLV) vaccine at 75 days of gestation. Seventy-five days is the beginning of the third trimester, when pregnancy outcomes are most sensitive to PRRSV.INF: Infection with PRRSV - Animals will be inoculated intramuscularly and intranasally with the highly virulent and new NC PRRSV strain (RFLP 1?7?4) at 75 days of gestation.VX0+INF: Vaccination followed by infection - Animals will be vaccinated at day 0 and inoculated with NC PRRSV at 75 days of gestation.The first 3 groups will allow comparison of the immune responses in animals subject to vaccination and infection (VX75 vs. INF) as well as distinguishing between responses at early and late stages of gestation (VX0 vs. Vx75). Blood samples will be taken at entry to determine if animals are negative to PRRSV. Subsequent sampling will take place at 8 time points between days 0 and 35 (roughly corresponding to peak Ab) and 8 time points between 75 and 110 days of gestation (d.o.g.). These time points will allow for the quantification of PRRSV RNA (detectable by ~5 days post exposure), anti-viral and pro-inflammatory cytokines (IFNα, TNFα, IL1β, IL6, IL8), immunoregulatory cytokines (IFNγ, IL10, IL12), and total and neutralizing antibodies (evident in most animals by ~14 days post exposure) thereby characterizing viral infectiousness and innate (early) and adaptive (late) host immune responses. Additional sampling between days 35 and 75 will allow monitoring for unwanted cross-contamination. Blood samples will be stored in an -80°C freezer.Aim 2: To characterize the genetic response to vaccination and infection, we will quantify total gene expression in the blood of gilts during exposure. Using the next-generation RNA sequencing (RNA?seq) methodology, blood collected at the same time points detailed in Aim 1 will be used for gene expression profiling. Data collection, management and statistical analyses will be performed using the standard bioinformatics and statistical tools outlined in Koltes et al. (2015). We will specifically assess the different contrasts between infection and vaccination. The CTRL group will be used to control for the standard gene expression profile of pregnant gilts. The list of genes showing differential expression (DE) between groups will be used for gene enrichment analysis (Mi et al. 2013) to identify the biological processes associated with PRRS response in gestating gilts. In addition, we will assess the complex interactions between DE genes using pathway and co-expression network analyses (Watson-Haigh et al., 2010).Aim 3: To utilize a novel mathematical model to assess variability in immune response as dictated by host-pathogen interactions, we will develop an integrative model of host-pathogen interactions in PRRSV infection. Two principle challenges inherent to developing the model are the need to decide which elements of a complex network of interdependent physiological processes to include in the model and the need to assign values to the parameters describing these processes. The first challenge is often met by starting with a minimal model (e.g. cells susceptible to infection and free virus causing infection) and adding complexity only as required to capture physiological outcomes of interest (e.g. anti-viral cytokines promoting viral clearance). The challenge of parameter estimation can be addressed by fitting the model to experimental data characterizing outcomes of interest. Such data is frequently unavailable when modeling efforts and experimental investigations proceed independently. This leaves model developers wrangling with incomplete datasets from multiple sources differing in critical variables such as viral strain, animal model, cytokines measured, or sampling time-points.As a starting point in addressing the first challenge, we have implemented a preliminary model based on the work of Go et al. (2014). The algorithm, implemented in MATLAB, predicts the dynamics of 18 states via differential equations based on 39 fluxes that represent processes of recruitment, activation, inactivation, production, and decay. The state variables include 4 macrophage states, free viral particles, natural killer cells, 9 cytokines, and 3 effectors of the adaptive immune response (cellular, humoral - including antibodies, and regulatory). A total of 32 model parameters makes parameter estimation a nearly insurmountable challenge given current datasets. The iterative and integrated experimental and computational study we propose will overcome this challenge. The preliminary mathematical model will be used to suggest blood sampling timepoints of high value. Quantification of cytokine and antibody levels as described in Aim 1 will provide multiple dimensions of immune response data from a single study involving controlled infection using a single viral strain. Differential parameterization of the model based on data from the 4 exposure groups will permit numerical simulation of immune dynamics associated with both vaccination and infection. Furthermore, gene expression (Aim 2) offers an alternative to numerical sensitivity analyses as a means to identify critical elements of the host response. Results from our preliminary analysis indicate the importance of 9 flux rate constants in determining the evolution of the humoral response (normalized model sensitivity >20%). Of these, one parameter corresponds to PRRSV strain infectiveness and one to activation of the cellular component of the adaptive immune response, i.e., cytotoxic T-cells. The remainder are almost exclusively associated with the innate immune response, split between amplifying and inhibiting effects on the humoral effector dynamics.