This study analyzed data from blood samples collected on young, healthy piglets that were subsequenty entered into a natural polymicrobial disease challenge to: (1) estimate genetic parameters of antibody levels (NAb and total IgG) in plasma of young, healthy pigs; (2) evaluate the potential of these NAb and IgG levels as indicator traits to select for disease resilience; and (3) conduct GWAS analysis for these antibody levels. To the best of our knowledge, this represents the first report on the genetic basis of NAb in the blood of pigs and their relationship with disease resilience.
NAb levels of two isotypes (IgG and IgM) associated with four important PAMPs (KLH, LPA, LTA and PDG) were quantified. NAb are non-inducible, meaning that their levels do not increase following exposure to key epitopes. Piglets are, however, exposed to multiple PAMPs during early life, and potentially in utero and during birthing. Three of the antigens used in this study (LPS, LTA, PDG) have PAMP epitopes that are common to Gram-negative and Gram-positive bacteria. However, KLH has PAMP epitopes derived from the inedible mollusc, Megathura crenulate, which commercial piglets are very unlikely to have been exposed to, although KLH NAb could cross-react with other antigens that have a similar structure. In spite of this apparent lack of specificity, which is common to IgM and the earliest IgG antibodies produced, the fact that these antigens are readily available and their respective antibodies can be easily measured in blood, makes them attractive targets for as indicator traits for disease resilience.
Genetic parameters of NAb
Genetic parameters for NAb in blood have been studied in several other livestock species but not in pigs. In poultry, Siwek et al. [29] estimated the heritability for NAb binding LPS and LTA in blood from different chicken lines and at different ages to range from 0.09 to 0.23 and from 0.03 to 0.42, respectively; Wijga et al. [25] estimated a heritability of 0.23 for the titer of NAb binding rabbit red blood cells in chickens. Thompson-Crispi et al. [24] estimated the heritability of IgM and IgG NAb binding KLH to be 0.32 and 0.18, respectively. Klerk et al. [22] found that the heritability of NAb was higher for IgM (0.25) than for the IgG isotype (0.15). Ploegaert et al. [14] estimated intra-herd heritabilities for NAb titers in milk of Dutch Holstein-Friesian cows to range from 0.10 to 0.53. In our study, estimates of the heritability of NAb titers in young healthy pigs ranged from 0.12 to 0.53, which is within the published range of estimates from previous studies [14, 24, 30]. In our study, heritability estimates were low for IgG NAb but moderately high for IgM NAb. Estimates of the proportion of variance explained by maternal effects for IgG and IgM NAb showed the opposite trend, being moderately high for IgG NAb and low for IgM NAb.
Natural antibodies consist of three isotypes: IgM, IgG and IgA. A previous study found that immunoglobulins IgG, IgM and IgA can transfer to lacteal secretions in the parturient sow and be absorbed by neonatal piglets [31]. In humans, IgG has been found to be the only isotype that has the ability to significantly cross the placental barrier to provide immunity to the fetus [32]. Previous studies in cattle have also shown that colostrum contains IgG NAb [33]. Because IgG transfers between the dam and fetus in pigs, part of the IgG NAb measured in piglet plasma likely originated from the colostrum from the dam, which explains the high litter effects for NAb IgG and IgG-T. Porter [31] found that more than 60% of whey protein in milk of sows was immunoglobulin, with the IgG isotype accounting for the majority (79.7%), while the IgM isotype content was only 6.27%. The contents IgG and IgM in colostrum are, however, different, which might lead piglets to get more IgG and less IgM from colostrum, explaining the higher litter effects for IgG NAb than for IgM NAb.
Ploegaert et al. [14] showed that the genetic correlation between titers of total NAb (IgG, IgM, and IgA) binding the same four antigens as used in our study ranged from 0.45 to 0.99 in milk of Dutch Holstein-Friesian cows and the titers of IgG NAb binding KLH in milk was highly correlated with the corresponding titer in serum (0.70). The estimates of the genetic correlations between NAb (IgG and IgM) binding the four antigens obtained in our study (0.41 to 0.99) are in agreement with these literature estimates. In dairy cows, Thompson-Crispi et al. [24] estimated a negative genetic correlation (− 0.41) between IgG NAb and IgM NAb titers in milk, which contrasts with the moderate positive estimates in our study.
Our results show that a given NAb isotype binding different antigens has a common genetic basis and selection of any of the four antigens is expected to result in correlated responses in the same direction for antibodies to the other PAMPs. Hence, there may not be a need to measure NAb levels for all four antigens in genetic selection programs. The KLH NAb were strongly correlated with NAb to the other three antigens, and therefore, may be a reliable measure of NAb, given its unique nature.
Few studies have assessed the genetic and phenotypic relationships of NAb with total IgG titers in blood. In the current study, total IgG had lower genetic and phenotypic correlations with IgM NAb than with IgG NAb, except for LTA binding IgM NAb. A recent study by Berghof et al. [17] in chickens estimated the phenotypic and genetic correlations of KLH binding IgG NAb with total IgG to be 0.08 and − 0.61, respectively, but with a high standard error for the latter (0.55). Our estimate of the genetic correlation of IgG-T with KLH-G was strong and positive (0.81) and the phenotypic correlation was also positive (0.51). Possible reasons for the differences between these estimates may be the species studied, differences in age and prior exposure to pathogens, and the large standard errors of the estimates.
Relationships of antibody levels with resilience
NAb levels prior to challenge had weak phenotypic correlations with subsequent disease resilience and performance traits after challenge, although some were significant and in the expected direction. For example, some NAb levels were negatively correlated with mortality, the number of parenteral treatments, and day-to-day variation in duration at the feeder (VARDUR), and positively correlated with growth rate. Consistent with these correlations, we also found that pigs that died had significantly higher NAb for some antigens and isotypes than pigs that survived. Estimates were, however, too small to serve as reliable phenotypic predictors of resilience. Star et al. [27] reported that the levels of NAb binding KLH and LPS were phenotypically positively correlated with the probability that chickens survive during the laying period. Sun et al. [34, 35] also revealed that IgG and IgM NAb levels in blood, especially IgM binding KLH, were phenotypically positively correlated with survival rate to around 20 weeks of age in both purebred and crossbred laying hens.
In the present study, several NAb traits had moderately high estimates of genetic correlations with some disease resilience and performance traits following challenge. All IgM NAb levels had negative genetic correlation estimates with mortality. While in the expected direction, none of these estimates were significantly different from zero. Estimates of the genetic correlation of IgG NAb and IgG-T with resilience and performance traits following challenge did not have a consistent direction and had high standard errors, likely because of the lower heritability for IgG NAb compared to IgM NAb. All IgG NAb traits had positive genetic correlation estimates with the number of parenteral treatments, with the estimate of PDG-G with the number of treatments in the challenge nursery being significantly different from zero. The estimate of the genetic correlation of IgM NAb with the number of treatments was not significantly different from zero. The direction of estimates of genetic correlations of NAb levels with growth rate were also mixed, but the significant correlations were positive (i.e. of growth rate in the challenge nursery with KLH-G and LPS-G), as expected.
Putz et al. [6] showed that smaller day-to-day variation in feed intake (VARFI) or duration (VARDUR) and the proportion of off-feed days based on negative residuals for 5% quantile regression of feed intake (OFFFI) or duration (OFFDUR) on age were genetically correlated with improved disease resilience and performance under disease challenge in the NDCM. In the present study, most estimates of the genetic correlation of these traits with IgG NAb and IgG-T were positive, opposite to expectations. The estimate of the genetic correlations of VARDUR with KLH-G was significant and similar to the unexpected positive correlation of KLH-G with growth rate in the challenge nursery, as well as the estimate for VARDUR with total IgG.
In general, compared with IgG NAb and total IgG, IgM NAb had a relatively consistent genetic correlation estimates with resilience and performance traits under challenge, especially with mortality, and their direction was consistent with expectations.
Genomic regions associated with antibody levels
Several genomic regions that were significantly associated with antibody levels were identified in the GWAS and numerous positional candidate genes were identified for these regions (Table 4). None of the positional candidate genes identified by the GWAS analysis in this study have been reported in previous studies on NAb, except the CD14 gene on chromosome 2, which will be discussed in the following.
Functional annotation analysis of these positional candidate genes for IgG NAb did not identify annotated functional clusters. However, 13 genes related to the immune process were identified. In poultry, Berghof et al. [17] identified a significant genomic region associated with KLH-M and total IgM concentration that contains the TLR1A gene. Although the TLR gene was not in the significant genomic regions identified in our study, the significant regions we identified on chromosome 2 around 143 Mb for average IgG NAb and around 148 Mb for LPS-G, KLH-M and LPS-M are close to the CD14 gene. This gene has been reported to encode proteins involved in inflammatory and innate immune responses and, together with TLR-4 and MD-2, acts as a common receptor for bacterial LPS [36]. The TLR signaling pathway is critical for innate immunity and provides the first line of defense against antigens.
The 148 Mb region of chromosome 2 also contains the ankyrin repeat and KH domain-containing protein 1 (ANKHD1) gene, which is involved in the immune process. Miles et al. [37] found that a novel splice variant of ANKHD1 may play a role in the apoptosis and cell survival pathway. Machado-Neto et al. [38] found that the ANKHD1 gene is involved in the Hippo signaling pathway and can promote cell growth and cell cycle progression by upregulating Cyclin A. In an acute leukemia human study, Traina et al. [39] found that the ANKHD1 gene may be related with the abnormal phenotype of leukemia cells as a scaffold protein. It was also found that the ANKHD1 gene can reduce cell growth and delay cell cycle progression in the S phase.
Three immune genes were detected in the significant genomic region for KLH-G (at 76 Mb on chromosome 7), i.e. T cell receptor delta chain C region (TRDC), interferon regulatory factor 9 (LRD9), and interleukin 25 (IL25). TRDC participates in antigen recognition [40], especially for a variety of self and foreign non-peptide antigens that are frequently expressed at the epithelial boundaries between the host and external environment. Once the antigen is recognized, a rapid, innate-like immune response is produced to participate in pathogen clearance and tissue repair [41, 42]. LRD9 is a transcription factor that can mediate signaling by type I interferons. After type I interferons bind to cell surface receptors, Jak kinases (TYK2 and JAK1) are activated, which leads to tyrosine phosphorylation of signal transducer and activator of transcription 1 (STAT1) and signal transducer and activator of transcription 2 (STAT2). IRF9/interferon stimulates transcription factor 3 gamma (ISGF3G) to associate with the phosphorylated STAT1:STAT2 dimer to form a complex called the ISGF3 transcription factor, which can enter the nucleus. Interferon stimulates transcription factor 3 (ISGF3) binds to the interferon stimulated response element (ISRE) to activate the transcription of interferon stimulated genes, which drives the cell into an antiviral state [43]. IL25 is involved in cytokine binding and inflammatory response. and encodes cytokine proteins, which can induce NF-kappaB activation and stimulate the production of interleukin 8 [44]. Another important function for IL25 is promoting the development of a Th2 immune response, which can protect against bowel infection by helminths [45, 46].
A series of beta defensin genes are located in the significant genomic region for IgG at 35 Mb on chromosome 17. Beta-defensin belong to the mammalian defensin family and are involved in the first line of defense of innate immunity [47]. Defensin helps to protect against a variety of microbes, including Gram-positive and Gram-negative bacteria, fungi, yeast and enveloped viruses [48]. Defensins may be involved in linking innate and adaptive immune response, and have been found to act as signal molecules of the immune system and a chemokine for T lymphocytes and immature dendritic cells [49].
The NFIL3 gene is located in a significant genomic region at 2 Mb on chromosome 14 associated with IgG NAb. The NFIL3 gene regulates the transcription of interleukin-3 (IL3), which controls IgE class switching [50]. Kashiwada et al. [51] found that NFIL3 was the key regulator of type 2 helper cells, which are important for the development of allergic immune responses. The spleen tyrosine kinase (SYK) gene is located in the same genomic region (at 2 Mb on chromosome 14) that was found to be associated with IgG NAb. This gene is involved in multiple biological processes, including innate immune recognition, cell adhesion, platelet activation, and vascular development [52].
Although functional annotation analysis did not identify gene clusters that were associated with the identified positional candidate genes for IgM NAb, four genes associated with immune response were identified in the genome region chromosome 14 around 48 Mb. These include the leukemia inhibitory factor (LIF), macrophage migration inhibitory factor (MIF), oncostatin M (OSM), and sushi domain containing 2 (SUSD2) genes. LIF is a multifunctional cytokine that mediates neuronal, hepatic, endocrine, inflammatory, and immune responses in autocrine and paracrine manners [53]. MIF is involved in the host antimicrobial alarm system and stress response, which can promote pro-inflammatory functions [54]. The MIF proteins are released by stimulated white blood cells and produce an acute immune response by binding to CD74 on immune cells such as macrophages, lymphocytes, dendritic cells, and endothelial cells [55]. The OSM gene has been reported to be an inflammatory mediator, similar to cytokines, but its exact effect on the immune system is unknown [56]. The SUSD2 gene is involved in the invasiveness of breast cancer cells and tumor evasion [57]. The CSF3R gene, which is associated with cytokine binding and cytokine receptor activity processes, is located in the genome region around 92 Mb on chromosome 6. This gene can encode the receptor protein for colony stimulating factor 3, which controls the production, differentiation, and function of granulocytes. After ligand binding the CSF3R, the receptor undergoes a conformational change, which can active the downstream pathways including JAK/STAT, PI3K/AKT, and MAPK/ERK [58].
Comparison with other innate immunity and immunocompetence traits in pigs
Several previous studies have estimated genetic parameters of immune traits in blood and provided a genetic framework for potential immune options in pigs [59,60,61]. Clapperton et al. [59, 60] showed that several traits measured in blood (white blood cells and peripheral blood mononuclear leucocyte subsets) that are related to the innate and adaptive immunity, are heritable and genetically negatively correlated with growth performance under different health status conditions. This suggests that these immune traits could be the potential genetic predictors of performance under different health conditions, which agrees to some extent with our results for NAb. In a review on piglet survival, Heuß et al. [62] argued that it is necessary to investigate the relationship of immune parameters with the robustness, survival, and performance. Our study offers insights into the genetic relationships between immune traits measured in young healthy pigs and disease resilience and production traits under disease, including survival and performance.
Limitations of natural polymicrobial disease challenge
The disease challenge model used here was designed to mimic a severe disease challenge on commercial farms with multiple pathogens, to maximize the expression of disease resilience. Because of seasonal effects, necessary veterinary interventions, and the dynamic nature of natural transmission of pathogens in a barn, pathogen exposure was not constant from batch to batch, similar to a commercial situation, where pathogen profiles varies widely between farms and over time within a farm. However, by evaluating pigs from multiple batches, the relationships identified in this study are expected to be robust to the specific level and nature of pathogen exposure. Also, because of the polymicrobial dynamic disease pressure, there was little value in determining the cause of each death.