Characterization of the abomasal transcriptome for mechanisms of resistance to gastrointestinal nematodes in cattle

The response of the abomasal transcriptome to gastrointestinal parasites was evaluated in parasite-susceptible and parasite-resistant Angus cattle using RNA-seq at a depth of 23.7 million sequences per sample. These cattle displayed distinctly separate resistance phenotypes as assessed by fecal egg counts. Approximately 65.3% of the 23 632 bovine genes were expressed in the fundic abomasum. Of these, 13 758 genes were expressed in all samples tested and likely represent core components of the bovine abomasal transcriptome. The gene (BT14427) with the most abundant transcript, accounting for 10.4% of sequences in the transcriptome, is located on chromosome 29 and has unknown functions. Additionally, PIGR (1.6%), Complement C3 (0.7%), and Immunoglobulin J chain (0.5%) were among the most abundant transcripts in the transcriptome. Among the 203 genes impacted, 64 were significantly over-expressed in resistant animals at a stringent cutoff (FDR < 5%). Among the 94 224 splice junctions identified, 133 were uniquely present: 90 were observed only in resistant animals, and 43 were present only in susceptible animals. Gene Ontology (GO) enrichment of the genes under study uncovered an association with lipid metabolism, which was confirmed by an independent pathway analysis. Several pathways, such as FXR/RXR activation, LXR/RXR activation, LPS/IL-1 mediated inhibition of RXR function, and arachidonic acid metabolism, were impacted in resistant animals, which are potentially involved in the development of parasite resistance in cattle. Our results provide insights into the development of host immunity to gastrointestinal nematode infection and will facilitate understanding of mechanism underlying host resistance.


Introduction
Gastrointestinal (GI) nematode infections in ruminants remain a major impediment to the efficient production of both meat and dairy products, and therefore, represent a major constraint on global food availability. These GI infections have a significant economic impact on the U.S. cattle industry, with an estimated annual cost of~$2 billion per year in lost productivity and increased operating expenses. Nematode infections of the GI tract impact numerous production traits. Among the most prominent effect is a reduction in weight gain that may cause decreased bodyweights of up to 14% [1]. Although the impact is particularly evident in young calves, substantial evidence suggests that infection produces long lasting effects on the productivity of adult cattle [2]. In dairy cows, parasitic infections reduce milk yield between 1.2 and 2.2 kg milk/cow per day [3]. Infections also negatively impact carcass quality and reproductive performance, including calving rate and calf mortality [4]. Potential economic loss resulting from GI nematode infections is clearly recognized by producers and veterinarians, as evidenced by the fact that approximately 99% of feedlots and 69% of dairies use a parasiticide in their operations [5].
Among 41 bovine GI nematodes, species from the genera Ostertagia, Cooperia, and Nematodirus are arguably the most important cattle parasites in temperate regions of the world, as assessed by their negative economic impact [6]. Development of protective immunity and resistance to these GI nematodes relies upon the precise control of expression of the host genome. It is evident that the evolution of regulatory programs controlling the transcriptome occurs at a rapid rate comparable to that of other genomic processes. Understanding these regulatory elements is crucial towards unraveling their functional relevance. Comparative transcriptomic analysis has emerged as a promising means for unraveling the molecular basis and regulatory networks underlying complex traits such as host resistance. While recent progress has been made with regard to genes associated with nematode resistance in small ruminants [7][8][9], an in-depth comparison and characterization of transcriptomic responses of cattle populations that harbor varying degrees of resistance to parasitic nematodes is not yet available. Sequencing steady-state RNA in a biological sample (RNA-seq technology) using next-generation sequencing platforms (e.g. Illumina) overcomes many limitations of previous technologies, such as microarrays and real-time PCR. Most importantly, RNA-seq has been shown to elucidate previously inaccessible complexities in the transcriptome, such as allele-specific expression and involvement of novel promoters and isoforms [10], detection of alternative splicing [11], RNA editing [12], novel transcripts [13], all in conjunction with quantitative evaluation of transcript abundance [14]. In this study, we utilized RNA-seq technology to characterize the bovine transcriptome response of nematode-resistant and nematode-susceptible heifers to identify molecular mechanisms that underlie host resistance to GI nematodes in cattle.

Animals and parasitology
Six 12-month-old Angus heifers that differed with regard to susceptibility to GI nematode infection (3 resistant and 3 susceptible) were used in this experiment. These heifers were from a selective breeding program for parasite resistance that was initiated at our facilities in 1991, using parental stock originating from the Wye Angus herd [15,16]. Once the initial breeding females were identified, semen from high and low EPG (eggs per gram of feces) bulls was used to produce calves of desired phenotypes. Calves were kept with their dams on pastures with extremely low numbers of parasites prior to weaning. When the median age of the contemporary group was 205 days, calves were weaned and placed on pastures infected with the two most common nematode parasites of cattle, Ostertagia ostertagi and Cooperia oncophora. The calves were monitored weekly for a number of parasitologic and immunologic parameters along with selected measurements of animal growth. The calves were kept pastured for a minimum of 120 days. Replacement animals were selected for secondary challenge experiments, while parasitologic and immunologic parameters were collected from animals chosen for slaughter. This program resulted in resource populations selected for the fecal egg trait (high or low fecal eggs counts or eggs per gram, EPG; high or low parasite resistance, respectively). Based on actual weekly EPG counts and sire expected progeny difference (EPD) values for EPG, a total of six heifers were selected for this study. Three heifers used in this study were classified as susceptible (high EPG EPD value) and the remaining 3 were classified as resistant (low EPG EPD value).
At the end of the grazing season, all heifers were treated with a combination of 10 mg fenbendazole and 0.5 mg moxidectin per kg of body weight to remove existing GI parasites transmitted from the infected pastures. After resting for 30 days on concrete to preclude further parasite exposure, the heifers were orally infected with a single dose of combined O. ostertagi and C. oncophora infective L3 larvae (8.5 × 10 4 and 1.0 × 10 5 O. ostertagi and C. oncophora L3 larvae, respectively, per animal) and housed on concrete for an additional 20 days, allowing the experimental infection to progress. EPG was monitored during the resting period to ensure that the drug treatment eliminated all pre-existing parasites. The infective L3 larvae were obtained from cultures maintained at the USDA-ARS Beltsville facilities. The heifers were handled according to a protocol approved by the Beltsville Agricultural Research Center Animal Care and Use Committee, following Institutional Animal Care and Use Committees (IACUC) guidelines. The heifers were sacrificed at 20 days post infection (dpi). Worms were counted from the contents of the abomasum and small intestine. The full-thickness of folds from the fundic abomasa were collected, minced into 1-2 cm pieces, and snap frozen in liquid nitrogen prior to storage at -80°C until total RNA was extracted.

RNA extraction and sequencing using RNA-seq technology
Total RNA was extracted using Trizol followed by DNase digestion and Qiagen RNeasy column purification as previously described [17]. The RNA integrity was verified using an Agilent Bioanalyzer 2100 (Agilent, Palo Alto, CA, USA) with a RIN value > 7.0. High-quality RNA was processed using an Illumina RNA-seq sample prep kit following the manufacturer's instruction (Illumina, San Diego, CA, USA). Final RNA-seq libraries were validated and sequenced at 36bp/sequence read using an Illumina GAIIx sequencer at a depth of approximately 23.7 million sequences per sample (mean ± SD = 23 723 620 ± 7 447 499 per sample).

Data analysis and bioinformatics
23 632 bovine genes in the Bovine Official Gene Set version 2 (OGS2.0) [18] were first mapped to the bovine reference genome (Btau4.0) using Genomic Short-read Nucleotide Alignment Program or GSNAP [19]. The best mapping position of each gene (≤ 10 kb intron span, ≥ 95% identity, ≥ 90% coverage, and minimum tail length of 5% of coding sequences) was extracted. Accordingly, 20 809 of the 23 632 bovine genes were uniquely mapped. After removing ambiguously mapped genes, 18 834 genes were used for RNA-seq data analysis.
Raw sequence reads were then checked using several layers of quality control filtering to remove low-quality reads. Raw reads with ≥ 2 ambiguous nucleotides (N) were discarded. Trimming removed approximately 1% of input raw reads and led to the retention of 99% of raw reads. The input reads after cleansing were mapped to the reference genome with gene coordinates using Bowtie (v0.12.7), an ultrafast and memory-efficient shortread aligner using a Burrows-Wheeler index [20]. Approximately 68.5% of trimmed reads mapped to the bovine genome (mean ± SD = 68.54% ± 2.48%). Only reads with one unique best match in the reference genome were used for subsequent analyses. The read depth of each gene was computed based on the coordinates of mapped reads and gene locations in the reference genome and was normalized using a method that corrects for biases introduced by RNA composition and differences in the total numbers of uniquely mapped reads in each sample [21]. Only genes having ≥ 20 uniquely mapped reads (mean of all 6 samples) were further analyzed. The R package edgeR was used to test the null hypothesis that expression of a given gene is not different between the two groups [21]. The normalized read counts were also analyzed using the DEGseq algorithm [22]. The DEGseq built-in function "samWrapper" that is recommended for testing RNA-seq data with biological replication was used to detect differential expression. Candidate genes were first sorted based on P value (P < 0.05) and fold change (2-fold as a cutoff). Genes identified as candidates for differential expression were further filtered with a false discovery rate (FDR) of < 5% to account for multiple testing. Differentially-expressed genes in the transcriptome were further analyzed using Gene Ontology (GO) analysis (GOseq). Over-representation of certain GO terms was determined based on Fisher's exact test. A multiple correction control (permutation to control false discovery rate [23]) was implemented to set up the threshold to obtain the lists of significantly over-represented GO terms. The candidate genes were analyzed using IPA v9.0 for pathways (Ingenuity Systems, Redwood City, CA, USA).
Tophat (v1.2.0) was used to map input reads to the reference genome [11] and identify potential splice junctions or splicing variants. Only reads with one unique match in the reference genome were used for subsequent analyses. The maximum allowed intron size was 5kb (a conservative parameter to avoid a high false discovery rate). At each potential splice junction, spanning reads were counted. Potential splice junctions were compared to annotated splice sites. To identify differential splice junctions between two groups, normalized read counts of splice junctions were required to be eight times different between resistant and susceptible groups. An unpaired t-test was performed on normalized sequence read counts. Splice junctions at P ≤ 0.05 were considered candidates junctions that were differentially regulated between resistant and susceptible groups.

Real-time RT-PCR
Real-time or quantitative RT-PCR (qPCR) was performed as previously described [17]. Briefly, the cDNA synthesis was performed using an iScript cDNA Synthesis kit (Bio-Rad, Hercules, CA, USA). Real-time RT-PCR analysis was carried out with an iQ SYBR Green Supermix kit (Biorad) using 200 nM of each amplification primer and the 1 st -strand cDNA (100 ng of the input total RNA equivalents) in a 25 μL reaction volume as described. The amplification was carried out on an iCycler iQ™ Real Time PCR Detection System (BioRad) with the following profile: 95°C-60 s; 40 cycles of 94°C-15 s, 60°C-30 s, and 72°C-30 s. A melting curve analysis was performed for each primer pair. The gene encoding for phospholipase A2, group IVA (cytosolic, calciumdependent) (PLA2G4A), which has a relatively constant expression level across all experiment samples, was used as an endogenous control. Relative gene expression data was calculated using the 2 -ΔΔCT method. The fold change was normalized against the susceptible group.

EPG and worm counts
Mean weekly EPG values were 8.8 ± 1.6 (Mean ± SD) and 31.3 ± 10.4 for low-EPG (resistant) and high-EPG (susceptible) heifers, respectively (N = 3) during the 6month grazing season ( Figure 1). In accordance with the experimental design, this difference was statistically significant (P < 0.05). Although a temporal fluctuation in weekly EPG values was evident, resistant heifers shed constantly fewer parasite eggs in feces than did susceptible heifers. Over the grazing period, resistant heifers gained more weight (P < 0.05) than susceptible heifers (159.0 ± 32.8 vs 92.1 ± 20.4 lbs). Similarly, resistant heifers displayed a numerically, but not statistically, greater (P > 0.05) gain in hip height over the experiment period (Table 1). Serum pepsinogen levels between both resistant and susceptible groups were statistically indistinguishable. The mean number of total parasite worms (both O. ostertagi and C. oncophora) recovered from resistant heifers (5 200 ± 3 191) after the experimental challenge were numerically less than those of susceptible heifers (5 923 ± 3 203), but not significantly less (P > 0.05). These worm burden data were not unexpected because the population under study has never been selected for worm burdens as an indicator trait in the applied breeding program. Additionally, the worm counts were obtained from an experimental infection with a high dose of infective larvae that is not typically encountered by calves under nature exposure.

Differentially expressed genes
Normalized sequence counts were analyzed using both edgeR and DEGseq algorithms. A total of 203 genes met 2 criteria: unadjusted P value < 0.05 and 2-fold difference in normalized read counts between resistant and susceptible animals (Additional file 1). These candidate genes were further filtered with a stringent cutoff (FDR < 5%). Sixty four genes were significantly different between resistant and susceptible animals at FDR < 5% (Table 2). These genes had a significantly higher ratio of normalized sequence counts between resistant and susceptible groups. For example, common salivary protein BSP10, form A (BT12506 or SPLUNC2A) was expressed 57.68 fold higher in resistant heifers. Similarly, the sequence counts of intelectin (ITLN2) in the abomasum of resistant heifers were 51.98 times higher than in that of susceptible heifers. Mucin 12 (MUC12) and fatty acid binding protein 6, ileal (FABP6, intestinal bile acid-binding protein or gastrotropin) were significantly over-expressed in resistant animals. Several apolipoproteins (APOA1, A4, B100, and C2) were also over-expressed in resistant heifers. Transcripts for alphainducible protein 27 and 27-like 2 (IFI27 and IFI27L2) were more abundant in resistant than in susceptible animals.

Splicing variants
A total of 94 224 potential splice junctions spanned by ≥ 1 sequence read were identified using TopHat. Among them, 139 junctions displayed a significantly different number of sequence reads between resistant and susceptible groups (FDR < 5%). These 139 junctions were distributed on 28 autosomes and the X chromosome. There were no junctions on BTA12 that had significantly different numbers of sequence reads between the Figure 1 Weekly mean fecal egg counts (eggs per gram, EPG or FEC) of resistant and susceptible Angus heifers grazing on infected pasture between April to October. Y-axis represents mean weekly fecal egg counts. * indicates a significant difference in EPG between resistant and susceptible groups (P < 0.05). ** P < 0.01; *** P < 0.001.  two groups. However, the distribution of these 139 junctions did not appear to be random, and the number of these junctions was not proportional to the physical length of the chromosomes. The vast majority of these junctions were unique to one of the two groups.  Table 3.

Gene Ontology (GO) and pathway analyses
Over-representation of GO terms was determined based on Fisher's exact test and filtered further using a multiple correction control at FDR < 5%. As Table 4 shows, the GO enrichment of genes under study was predominantly associated with lipid metabolism.
To gain insights into pathways involved in the development of parasite resistance, we analyzed the differentially expressed genes using Ingenuity pathways analysis software, IPA. Among the 7 regulatory networks identified (data not shown), the primary function of 3 networks was related to lipid metabolism. The primary function of the 3 rd regulatory network was involved in antimicrobial response and inflammation. A total of 12 pathways were significantly impacted (P < 0.05) and possibly involved in the development of host resistance to parasitic infection in cattle (Table 5). FXR/RXR activation was the pathway most significantly impacted in resistant heifers (P = 3.66E-07) with at least 8 of the 203 differently expressed genes involved, including APOA1, APOB, APOC2, APOC3, FABP6, and IL18. The other pathways significantly impacted in resistant animals during parasitic infection included LXR/RXR activation (P = 2.78E-04), LPS/IL-1 mediated inhibition of RXR function (P = 8.80E-04), and arachidonic acid metabolism (P = 4.68E-03). In addition, acute phase response signaling was also impacted in resistant heifers (P < 0.05).

Real-time RT-PCR confirmation
The expression of 10 genes at mRNA level in the fundic abomasum was examined using real-time RT-PCR. The mRNA level of cholecystokinin B receptor (CCKBR, NM_174262), a receptor for gastrin, was extremely low in the fundic abomasum. The mRNA levels of gastrin (GAST, NM_173915) and pepsinogen 5, group I (pepsinogen A) (PGA5, NM_001001600) were reliably detected. PGA5 expression level appeared to be higher in resistant animals. However, the difference was not statistically significant due to a large variation while no changes in gastrin mRNA level between susceptible and resistant animals were detected. The expression of MUC12 was barely detectable (≤ 40Ct). The mRNA levels of glucosaminyl (N-acetyl) transferase 3, mucin type (GCNT3, NM_205809), mucin 2 (MUC2, NM_001245997), and galectin 15-like (LGALS13, XM_593263) were moderately abundant but no differences were detected between susceptible and resistant animals, consistent with the RNAseq results. Expression levels of BPI fold containing family A, member 2A (BPIFA2A or SPUNC2A, NM_174803), bovine putative ISG12(a) protein (IFI27, NM_001038050), lectin, galactoside-binding, soluble, 3 (LGALS3, NM_001102341), and bovine collectin-46 (CL46, NM_001001856), were significantly higher in resistant animals, consistent with RNAseq data (Figure 2).

Discussion
The parameters of resistance to GI nematode infections in cattle while yet to be precisely defined, generally include decreased worm establishment and reduced parasite fecundity. It has long been known that host genetic factors play a significant role in determining susceptibility and resistance. Among eight factors determining EPG variation, additive genetic variation is predominant and accounts for~30% of the variation in EPG [24]. Estimates of heritability for parasite indicator traits in ruminants are phenotype-dependent. In small ruminants, the heritability of adult worm length at the end of the first grazing season is very strong at 0.62 [24], whereas the heritability of EPG is moderate, ranging from 0.14 to 0.33 in Creole goats [25]. In cattle, the heritability of EPG released during the 1 st grazing season is approximately 0.30 [26]. The ability of calves to recognize parasitic antigens is also under the control of host genetics [27]. Several studies suggest that there exist significant differences in the ability of cattle to resist GI nematode infections, and 3 major responder types can be readily identified in outbreed cattle populations [28,29]. Worm establishment (worm burden) is predominantly influenced by host responder types. The ability of intermediate and high responders to mount a  more effective and rapid immune response compared to low responders is sustained after secondary infection, providing more evidence that genetics may play an important role in regulating host resistance. The finding that the different responder types, based on parasitological variables, also feature a different immune response is very interesting since this also provides the opportunity to study the influence of genetic components of the host immune response [28]. These observations have spurred efforts to develop resource populations and identify genes and QTL that underlie the resistance trait and to develop criteria for selective breeding [30]. A vigorous and effective mucosal immunity is essential for resistance to GI nematode infection in ruminants. The resistant phenotype is often manifested in the host transcriptome. For example, resistant sheep breeds are able to more rapidly up-regulate Th2 cytokines than susceptible breeds [31]. In Angus cattle, our evidence suggests that resistant heifers can better maintain inflammatory responses at the sites of infection, especially during early stages of infection [16]. In the current study, we conducted an in-depth transcriptomic analysis to identify molecular mechanisms that underlie the development of host resistance in cattle, taking advantage of a resource population developed via selective breeding. Our results suggest that among the 94 224 splice junctions identified, 133 were uniquely present in either resistant or susceptible cattle, possibly representing novel splicing variants that have implications in the development of host resistance. We identified 203 candidate genes that displayed significantly different numbers of sequences between resistant and susceptible animals at a combined cutoff value P < 0.05 and 2-fold.
The transcripts from 16 genes, including gastrinreleasing peptide (GRP) and macrophage-stimulating 1 (MST1), had a significantly higher number of sequences in susceptible cattle. GRP has been reported to be down-regulated by parasitic infection in a helminthmouse system. Our results suggest that parasitic infection in susceptible cattle may have a negative impact on the host enteric nervous system that extends beyond its role in modulating normal functions of host epithelial, immune, and muscle cells [32].
Among the 187 genes with more abundant transcripts in the abomasum of resistant heifers, a notable feature was the up-regulation of various lectins. At least 4 lectins, such as bovine-specific collectin 46 (CL-46), C-type lectin domain family 12 member A (CLEC12A), galectin 3 (LGALS3), and intelectin 2 (ITLN2), had significantly more abundant transcripts in the abomasum of resistant cattle, which confirmed our previous study utilizing high-density DNA oligo arrays [33]. ITLN2 and several C-type lectins, such as collectin 11 (COLEC11), cattlespecific collectin-46, and conglutinin, as well as galectins were strongly up-regulated in the abomasal mucosa of immune cattle developed using multiple rounds of drugattenuated infections [6,33]. ITLN2 expression is regulated by Th2 cytokine IL-4 [34]. Its elevated expression is observed in the sheep abomasums in response to Teladorsagia circumcincta infection, Dictyocaulus filaria natural infection [35], and Haemonchus contortus infection [36]. Most pertinent, this gene is naturally deleted in the genome of the susceptible mouse strain, C57BL/ 10, but is present in the genome of a nematode-resistant mouse strain, BALB/c, suggesting that this gene may serve a protective role in the innate immune response to Trichinella infection [37]. Cattle-specific collectin-46 has been suggested to provide the first line of defense against pathogens without eliciting a general inflammatory reaction [38]. Galectins also play an important role in innate immunity, including serving as receptors for pathogen-associated molecule patterns (PAMP), which is integral in recognizing carbohydrate moieties on the cell surface of parasites, activating various immune cells, participating in cytotoxicity, modulating innate immunity via binding to IgA, and promoting the reconstruction of damaged tissues as receptors for damageassociated molecular patterns (DAMP) [39]. Together, our results suggest that lectins may play an important role in invoking effective host immune responses and in the development of host resistance.
Our evidence also indicates that alterations in lipid metabolism may be necessary to the development of host resistance. The top function of 3 of the 7 regulatory networks identified was associated with lipid metabolism. GO terms associated with genes that were differently expressed between resistant and susceptible animals were also predominantly related to lipid metabolism (Table 4). Lipid metabolism is significantly regulated in the bovine small intestine. In response to C. oncophora infection, lipid balance in the GI tract during parasitic infection may be disrupted [17]. Polyunsaturated fatty acids (PUFA), especially those in omega-3 (n-3) and omega-6 (n-6) families, such as arachidonic acid and linoleic acid, have long been known to have strong immunomodulatory effects [40] and may serve as a potent inhibitor for Th1 response. In cattle, dietary supplementation with fish oil (omega-3 PUFA) results in a 24% reduction in EPG in calves that are infected with O. ostertagi and C. oncophora [41]. The treatment also leads to an increased percentage of immature parasites, indicating that PUFA may enhance protective immunity against parasitic infections. Interestingly, arachidonic acid metabolism was among the pathways most significantly impacted in resistant animals (Table 5). Arachidonic acid (AA) is one of the important PUFAassociated membrane phospholipids. When liberated from the plasma membrane, AA can be oxidized, via a series of enzymatic steps, to a variety of eicosanoids, including prostaglandins, thromboxanes, prostacycline, and leukotrienes. Eicosanoids act as signaling molecules and stimulate a variety of responses in their target cells, such as innate immune responses [42], inflammation, and smooth-muscle contraction. Dietary n-3 PUFA has been used to attenuate tissue AA levels and subsequent eicosanoid formation. Recently, worm killing activities of AA have been demonstrated [43]. In mice, a single oral dose of AA led to a significant reduction of total worm burden of Schistosoma. AA-mediated parasite killing is suggested to be due to excessive activation of parasite neutral sphingomyelinase, leading to sphingomyelin hydrolysis into ceramide and phosphorylcholine [43]. In addition, products of the 5-lipoxygenase pathway, a part of AA metabolism, are important mediators of inflammation. 5-lipoxygenase plays a major role in controlling parasite burden of Trypanosoma cruzi in mice [44]. Detailed link between lipid metabolism and the development of protective immunity and host resistance to parasitic infections in cattle is worthy of further investigation.
The three most significant pathways impacted in resistant animals are associated with retinoid X receptor (RXR). These pathways included FXR/RXR activation, LXR/RXR activation, and LPS/IL-1 mediated inhibition of RXR functions. RXR acts as a master coordinator of numerous signaling pathways [45] via dimerizing with other nuclear receptors, such as liver X receptor (LXR), farnesoid X receptor (FXR), and vitamin D receptor (VDR). This partnership exerts transcriptional control and leads to distinct functions ranging from cell proliferation and differentiation to lipid metabolism. In addition, RXR can bind to a variety of natural and synthetic ligands, including omega-3 unsaturated fatty acids [45], which in turn stimulate transcriptional activation by RXR partners. While retinoic acid receptors (RAR) bind all-trans retinoic acid (RA) and its 9-cis isomer (9-cis RA), which convey most of the activity of RA, only 9-cis RA and docosahexaenoic acid (DHA) are suggested to be endogenous RXR ligands [46]. RA can inhibit cytokine expression including reduction of TNFα, iNOS, IL-6, and IL-1β at the mRNA level [47]. Recently, it has been observed that LPS-specific regulatory networks in which NF-B plays a critical role in the mouse mucosa overlap with the LPS/IL-1β mediated inhibition of RXR functions [48]. The importance of RXR in cattle during C. oncophora infection has also recently been recognized [17]. Our future work will focus on the mechanistic link between RXR-related signaling pathways and the development of host resistance to GI nematode infection in cattle.

Additional material
Additional file 1: Genes with significantly different read counts between resistant and susceptible cattle in response to parasitic infection. A total of 203 genes met 2 criteria: unadjusted P value < 0.05 and 2-fold difference in normalized read counts between resistant and susceptible animals.