Abstract

Objective

   A study was conducted to identify metabolic biochemical differences
   between two chicken genotypes infected with Eimeria acervulina and to
   ascertain the underlying mechanisms for these metabolic alterations and
   to further delineate genotype-specific effects during merozoite
   formation and oocyst shedding.

Methods

   Fourteen day old chicks of an unimproved (ACRB) and improved (COBB)
   genotype were orally infected with 2.5 x 10^5 sporulated E. acervulina
   oocysts. At 4 and 6 day-post infection, 5 birds from each treatment
   group and their controls were bled for serum. Global metabolomic
   profiles were assessed using ultra performance liquid
   chromatography/tandem mass spectrometry (metabolon, Inc.,). Statistical
   analyses were based on analysis of variance to identify which
   biochemicals differed significantly between experimental groups.
   Pathway enrichment analysis was conducted to identify significant
   pathways associated with response to E. acervulina infection.

Results

   A total of 752 metabolites were identified across genotype, treatment
   and time post infection. Altered fatty acid (FA) metabolism and
   β-oxidation were identified as dominant metabolic signatures associated
   with E. acervulina infection. Key metabolite changes in FA metabolism
   included stearoylcarnitine, palmitoylcarnitine and linoleoylcarnitine.
   The infection induced changes in nucleotide metabolism and elicited
   inflammatory reaction as evidenced by changes in thromboxane B2,
   12-HHTrE and itaconate.

Conclusions

   Serum metabolome of two chicken genotypes infected with E. acervulina
   demonstrated significant changes that were treatment-, time
   post-infection- and genotype-dependent. Distinct metabolic signatures
   were identified in fatty acid, nucleotide, inflammation and oxidative
   stress biochemicals. Significant microbial associated product
   alterations are likely to be associated with malabsorption of nutrients
   during infection.

Introduction

   Eimeria spp. are apicomplexan parasites that share many metabolic
   pathways with their animal hosts [[32]1]. The genus Eimeria is the
   largest of the Eimeriidae family and is responsible for coccidiosis
   disease in poultry with global economic losses in excess of $3 billion
   annually [[33]2]. Upon ingestion by a host, the parasite undergoes a
   period of asexual reproduction (schizogony) resulting in the production
   of merozoites which proceeds to a sexual reproduction phase
   (gametogony) producing macro- and micro-gametocytes. Fertilization of
   gametocytes is followed by the formation of oocysts which are shed in
   the feces [[34]3]. Eimeria spp. exhibit a high degree of host and site
   specificity in the gastrointestinal tract, as well as distinct
   morphology of oocyst and pathogenic effects [[35]4]. The most common
   parasite species found in chickens are Eimeria (E.) acervulina, E.
   maxima and E. tenella and they are usually localized in the lower
   duodenum and upper ileum, mid-ileum near the Meckel’s diverticulum, and
   caeca, respectively [[36]4,[37]5].

   Infection of meat-type chickens (broilers) is often characterized by
   weight loss, poor feed utilization efficiency, intestinal lesions,
   diarrhea and ultimately death, depending on the degree of pathogenesis
   [[38]2,[39]6]. Efficient control of coccidiosis therefore requires
   rapid and accurate detection and identification of the Eimeria spp,
   however, current diagnosis of infection relies heavily upon post-mortem
   site specific enteric lesions and observed morphological
   characteristics during necropsy [[40]4,[41]7,[42]8]. Eimeria spp.
   infection has been identified as one of the predisposing factors of
   necrotic enteritis in poultry [[43]2–[44]5], therefore control of
   coccidiosis in poultry is paramount in the control of other enteric
   disease. Although DNA-based techniques for diagnosis using oocysts in
   fecal samples have been reported, there are differences in the
   sensitivity and specificity in detecting different Eimeria spp.
   [[45]9–[46]11].

   Progress in omic science has led to significant improvement in the
   understanding of the molecular and cellular mechanisms that underlie
   several disease pathologies. Metabolomics is increasingly being used in
   biomarker discovery, characterization of metabolites and metabolic
   changes, and development of specific biochemical fingerprints
   (signatures) for different cellular processes [[47]12,[48]13].
   Continuous advances in omics will undoubtedly lead to the development
   of better diagnostic tools and the discovery of new therapeutics for
   efficient control of coccidiosis and other diseases in chickens and
   other species.

   Despite the large body of work exploring the metabolome in several
   species, there has been no comprehensive study of the metabolome of
   chickens infected with Eimeria spp. We hypothesized that infecting
   different chicken genotypes with E. acervulina could identify metabolic
   changes that are likely to be mechanistically involved in the host
   infection response. We further assessed the metabolomic alterations at
   two time points (merozoite formation and oocyst shedding) post
   infection and ascertained genetic background-specific changes and
   effects.

Materials and methods

   All protocols were approved by the University of Georgia Institutional
   Animal Care and Use Committee. This research was conducted according to
   the guidelines approved by the institutional animal care and use
   committee of the University of Georgia. Two hundred and eighty male
   chicks comprising of equal numbers of Athens Canadian Random Bred
   (ACRB) and Cobb500 (COBB) genotypes were raised under standard
   husbandry practices in coccidian free rooms from hatch until 14 days of
   age. The ACRB is an unselected control population established in 1955
   and has been compared with the modern commercial broiler in several
   studies. An extensive review of the history and comparative studies
   involving the ACRB has been previously published [[49]14]. Cobb 500 is
   a commercial meat-type strain developed by Cobb Vantress Inc., in 1983.
   Although originally selected for breast meat, the Cobb 500 has been
   continuously selected for feed efficiency and growth [[50]15]. At 14
   days of age, the chicks were randomly assigned to 4 treatments groups
   in a 2 x 2 factorial design, with 7 replicates per treatment, and 10
   birds per replicate. The treatment effects (TX) consisted of bird
   genotypes (ACRB or COBB) and two infection levels (oral gavage of 2.5 x
   10^5 sporulated E. acervulina oocysts or distilled water (CTRL)). Birds
   were fed on a standard non-medicated grower diet, containing 20% crude
   protein and 12.92 MJ/kg metabolizable energy. Feed and water were
   provided ad libitum throughout the experiment. Birds were monitored
   twice a day for any adverse clinical signs. At 4 and 6 days post
   infection (dpi), 5 birds from each treatment group were randomly
   sampled and bled according to the University of Georgia animal care
   protocol. Serum from the blood samples was stored at -86°C. All birds
   were individually weighed at 0, 4, 6 and 7 dpi.

Metabolome profiling of chicken serum

Sample preparation

   The serum metabolomic quantification was performed by Metabolon, Inc.
   (Durham, NC, USA). The samples were deproteinized by dissociating small
   molecules bound to protein or trapped in the precipitated protein
   matrix by precipitating with methanol under vigorous shaking for 2
   minutes (Glen Mills GenoGrinder 2000) followed by centrifugation. The
   resulting extract was divided into five fractions for four different
   ultra-performance liquid chromatography-tandem mass spectrometry
   (UPLC-MS/MS) methods: (1) two fractions were analyzed by two separate
   reverse phase (RP)/UPLC-MS/MS methods with positive ion mode
   electrospray ionization (ESI); (2) one for analysis by RP/UPLC-MS/MS
   with negative ion mode ESI; (3) one fraction for analysis by
   HILIC/UPLC-MS/MS with negative ion mode ESI, and (4) one fraction was
   reserved for backup. Samples were placed briefly in a TurboVap®
   (Zymark, Palo Alto, CA, USA) to remove the organic solvent. The sample
   extracts were stored overnight under nitrogen before preparation for
   analysis.

Ultra-high performance liquid chromatography-tandem mass spectroscopy
(UPLC-MS/MS)

   For all the methods, a Waters ACQUITY ultra-performance liquid
   chromatography (UPLC) and a Thermo Scientific Q-Exactive high
   resolution/accurate mass spectrometer interfaced with a heated
   electrospray ionization (HESI-II) source and Orbitrap mass analyzer
   operated at 35,000 mass resolution was used. To ensure injection and
   chromatography consistency, sample extracts were dried and
   reconstituted in solvents compatible to each of the four methods. (A):
   the first aliquot was analyzed using acidic positive ion conditions,
   chromatographically optimized for more hydrophilic compounds. Using
   this method, the extract was gradient eluted from a C18 column (Waters
   UPLC BEH C18-2.1x100 mm, 1.7 μm) using water and methanol, containing
   0.05% perfluoropentanoic acid (PFPA) and 0.1% formic acid (FA). (B):
   the second aliquot was also analyzed using acidic positive ion
   conditions, however it was chromatographically optimized for more
   hydrophobic compounds. The extract was gradient eluted from the same
   aforementioned C18 column using methanol, acetonitrile, water, 0.05%
   PFPA and 0.01% FA and was operated at an overall higher organic
   content. (C): the third aliquot was analyzed using basic negative ion
   optimized conditions with a separate dedicated C18 column. The basic
   extracts were gradient eluted from the column using methanol and water,
   however with 6.5mM Ammonium Bicarbonate at pH 8. (D): the fourth
   aliquot was analyzed via negative ionization following elution from a
   HILIC column (Waters UPLC BEH Amide 2.1x150 mm, 1.7 μm) using a
   gradient consisting of water and acetonitrile with 10mM Ammonium
   Formate, pH 10.8. The MS analysis varied between MS and data-dependent
   MS^n scans using dynamic exclusion. The scan ranged was from 70 to 1000
   m/z.

Bioinformatics

   Raw data files were stored at the Metabolon Laboratory Information
   Management System (LIMS). The data was extracted and analyzed using
   peak-identification software. The software has data processing tools
   for quality control and compound identification, and a collection of
   information interpretation and visualization tools. Metabolites were
   identified using Metabolon library which is based on authenticated
   standards that contain the retention time/index (RI), mass to charge
   ratio (m/z), and chromatographic data (including MS/MS spectral data)
   on all molecules present in the library using software developed at
   Metabolon [[51]16,[52]17]. Over 3,300 commercially available purified
   standard compounds are registered into LIMS for analysis on all
   platforms for determination of their analytical characteristics.
   Area-under-the-curve was used to identify peaks. Whenever necessary,
   the data was normalized to account for differences in metabolite levels
   due to differences in the amount of material present in each sample.
   Data was log transformed and missing values were imputated using the
   minimum observed value for each compound ([53]S1 Table).

Statistical analyses

   Statistical analysis was based on a model that included the main
   effects of genotype, treatment and time post infection and their
   interaction using ArrayStudio ([54]www.omicsoft.com/array-studio).
   Analysis of variance contrasts were used to identify biochemicals that
   differed significantly between experimental groups. Statistically
   significant (p<0.5), as well as those approaching significance
   (0.05<p<0.10) metabolites are reported. An estimate of the false
   discovery rate (q<0.05) was calculated to account for multiple
   comparisons.

Random forest

   Random forest (RF) is a supervised ensemble and learning technique used
   to build predictive models for classification and regression problems
   [[55]18]. In the current study, RF plots using 30 biochemicals were
   used to discriminate between genotypes, time post infection and
   treatment. An accuracy greater than 50% was considered to be better
   than random chance.

Principal components analysis (PCA) and hierarchical clustering analysis
(HCA)

   PCA analysis was performed using the R package ‘pcaMethods’ with single
   decomposition [[56]19]. Unsupervised HCA was performed using complete
   linkage and Euclidian distance where each sample was used as a vector
   with all the metabolite values to determine sample similarity.
   Hierarchical clustering analysis was performed by using pheatmap
   package in R [[57]20]

Pathway enrichment analysis

   Each contrast was subjected to permutation testing to identify putative
   pathways containing more significant metabolites than would be expected
   by random chance alone using proprietary software (Metabolon Inc.,
   Durham, NC, USA). Statistically significant metabolites that were
   identified were mapped onto general biological pathways using Pathview
   software package [[58]21].

Results

Growth performance

   The birds did not show any adverse clinical effects for the duration of
   the experiment. The growth performance of the ACRB and COBB genotypes
   are presented ([59]Fig 1).

Fig 1. Body weight of chicken genotypes (ACRB and COBB) infected with Eimeria
acervulina (Treatment) and their uninfected control; *P<0.05.

   [60]Fig 1
   [61]Open in a new tab

   A: represent the Athens-Canadian Random Bred (ACRB) genotype; B:
   represent the Cobb genotype.

   There were no significant differences in growth among genotypes at 0
   and 4 dpi. However, at 6 and 7 dpi, there were significant differences
   (P<0.05) between the treatment (parasite infection) and control birds
   within each genotype.

Metabolome composition

   A total of 752 metabolites were characterized across genotype,
   treatment and time post infection classes. The detected metabolites
   were mapped onto biochemical pathways and grouped into classes by
   prevalence of lipids (44%), followed by amino acids (27%), xenobiotics
   (10%), nucleotides (6%), Cofactors/vitamins (4%), carbohydrates (4%),
   peptides (3%), and energy (2%) ([62]S2 Table). The statistical analysis
   demonstrates that E. acervulina infection resulted in the most
   significant changes among serum samples in the ACRB genotype. In fact,
   312 out of the 752 named biochemicals were detected in the ACRB
   genotype compared to only 184 biochemicals for COBB genotypes at 4 dpi.
   Additional time point comparisons (6 versus 4 dpi) for samples
   collected from infected animals revealed 340 significant changes for
   ACRB and 320 for COBB genotypes indicating similar level of response.
   Out of the 340 biochemicals that changed in the ACRB treatment group,
   286 increased while 54 decreased. In the COBB genotype, 153
   biochemicals increased and 167 decreased ([63]Fig 2).

Fig 2. Number of significant metabolite changes by genotype (ACRB and COBB),
infection with Eimeria acervulina (TX and CTRL), time (4 and 6 days post
infection) and their interactions.

   [64]Fig 2
   [65]Open in a new tab

   The complete statistical summary of the number of altered biochemicals
   for the different genotypes, treatment and time post infection are
   provided ([66]S1 Table).

Metabolic differences across genotype, infection and time

   Principal component analysis conducted using all serum samples showed
   wide distribution with defined separation of infected sera collected 4
   dpi from ACRB and COBB genotypes (with partial overlap with 4 day
   collected control samples) as indicated on [67]Fig 3 (left).

Fig 3. Principal component analysis of metabolic profiles indicating the
impact of genotype (ACRB and COBB), infection (TX and CTRL) and time (4 and 6
days post infection with Eimeria acervulina).

   [68]Fig 3
   [69]Open in a new tab

   Analyzing the data by genotype shows clustering of genotype and time.

   Surprisingly, sera collected at 6 dpi from inoculated birds was similar
   to the controls. Additional assessment restricting the analysis to
   samples collected either from ACRB ([70]Fig 3, right bottom) or COBB
   ([71]Fig 3, right top) genotype demonstrate better separation for 6 dpi
   Cobb collected samples (infected versus control) and 4 dpi ACRB serum
   samples. The hierarchical clustering analysis based on stepwise
   clustering method grouped metabolically similar samples close to each
   other, and, in agreement with PCA, showed limited sub-clustering based
   on time post infection and treatment ([72]Fig 4).

Fig 4. Hierarchical clustering analysis of metabolic profiles indicating the
impact of genotype (ACRB and COBB), infection (TX and CTRL) and time (4 and 6
days post infection with Eimeria acervulina).

   [73]Fig 4
   [74]Open in a new tab

   Under a given treatment and time combination, red indicates metabolite
   abundance higher than the mean and blue respresent metabolite
   abundances lower than the mean. The actual normalized levels of the
   metabolites were listed in [75]S1 Table.

   The top-level split separates samples collected at 4 dpi from other
   samples. Secondary split by genotype was also present, with time post
   infection as a confounding variable because there were untreated
   (control) and infected groups. The RF classification uniquely
   identified biomarkers differentiating classification groups. RF
   analysis using biochemical data derived from ACRB and COBB chickens
   resulted in predictive accuracy of 100% compared to 50% expected by
   random chance alone ([76]Fig 5).

Fig 5. Random forest plot of top 30 biochemical present in serum samples of
chicken genotypes (ACRB and COBB) infected with Eimeria acervulina.

   [77]Fig 5
   [78]Open in a new tab

   Compounds contributing the most to the accuracy of disease
   classification were identified as those having the highest mean
   decrease accuracy (MDA). A higher MDA value indicates a greater
   predictive value.

   RF analysis also generates an accompanying list of the top metabolites
   contributing most to the separation of the genotypes. The top 30
   metabolites separating ACRB and COBB genotypes pointed heavily to amino
   acids, nucleotides, xenobiotics and carbohydrates. The same analysis
   using biochemicals identified from samples collected at 4 and 6 dpi
   resulted in predictive accuracy of 95% ([79]Fig 6), with the top
   discriminating metabolites pointing towards lipid metabolism (e.g.
   13-HODE+9-HODE), nucleotides, peptides and amino acids.

Fig 6. Random forest plot of top 30 biochemical present in serum samples of
chickens infected with Eimeria acervulina for 4 (T4) and 6 (T6) days.

   [80]Fig 6
   [81]Open in a new tab

   Compounds contributing the most to the accuracy of disease
   classification were identified as those having the highest mean
   decrease accuracy (MDA). A higher MDA value indicates a greater
   predictive value.

   Additionally, RF analysis using samples collected from control and
   infected animals resulted in predictive accuracy of 90% ([82]Fig 7)
   with the main discriminating metabolites including carbohydrates,
   cofactors and vitamins and lipid metabolism.

Fig 7. Random forest plot of top 30 biochemical present in serum samples of
chicken infected with Eimeria acervulina (Treatment) and their control.

   [83]Fig 7
   [84]Open in a new tab

   Mean decrease was used as a measure of variable importance. Compounds
   contributing the most to the accuracy of disease classification were
   identified as those having the highest mean decrease accuracy (MDA). A
   higher MDA value indicates a greater predictive value.

Changes in lipid metabolism due to E. acervulina infection

   A metabolic signature of altered fatty acid mobilization and
   β-oxidation was observed as a function of infection status (E.
   acervulina infection vs control) and with time post infection (6 vs 4
   dpi) as evidenced by significantly elevated levels of long chain-,
   monosaturated-, polyunsaturated, and dicarboxylate- fatty acids in
   serum samples collected 6 versus 4 dpi in the ACRB (infected and
   controls) and COBB (control only) (blue box, [85]Fig 8), suggesting a
   potential interaction between time post infection and treatment for
   these compounds. The COBB genotype responded with significant changes
   in lipid utilization due to E. acervulina infection, where levels of
   most of the long chain and dicarboxylate fatty acids were significantly
   elevated at 4 dpi and decreased 6 dpi post infection (purple box,
   [86]Fig 8).

Fig 8. Changes in fatty acid metabolism in chicken genotypes (ACRB and COBB)
infected (TX and CTRL) with Eimeria acervulina and their interactions at 4
and 6 days post infection.

   [87]Fig 8
   [88]Open in a new tab

   There were significant decreases in lysophospholipids
   (2-arachidonolyl-GPA (20:4), 1-palmitoyl-GPE (16:0), 1-stearoyl-GPE
   (18:0)) in the COBB treatment group compared with the controls at 4 dpi
   ([89]S2 Table). Additional changes in the levels of metabolites
   associated with fatty acid β-oxidation, including carnitine conjugated
   fatty acids such as stearoylcarnitine (C18), palmitoylcarnitine (C16),
   oleoylcarnitine (C18:1) and linoleoylcarnitine (C18:2)* along with
   ketone body 3-hydroxybutyrate ([90]Fig 9) is suggestive of alterations
   in serum free fatty acid (FFA) utilization, and β-oxidation as a
   response to E. acervulina infection (in relation to control) and time
   post infection (6 vs 4 dpi).

Fig 9. Differences in abundance of lipid metabolites in chicken genotypes
(ACRB and COBB) infected with Eimeria acervulina (TX) and their control
(CTRL) at 4 and 6 days post infection.

   [91]Fig 9
   [92]Open in a new tab

   The adjusted P-values for significant comparison are presented in
   [93]S2 Table. The upper whiskers represent the maximum, and the lower
   whiskers the minimum values. The plus-signs indicate the mean values,
   while the median values are represented by the black line within the
   boxes.

Infection related changes in inflammation and oxidative stress

   Inflammation is a highly regulated process with multiple inputs,
   including oxidized lipids (eicosanoids, the products of LOX/COX
   activation) and other regulatory compounds. Infection led to a strong
   decrease in the eicosanoids, thromboxane B2 and 12-HHTrE for both
   genotypes at 4 dpi and recovered by 6 dpi ([94]Fig 10).

Fig 10. Differences in abundance of inflammation and oxidative stress
metabolites in chicken genotypes (ACRB and COBB) infected with Eimeria
acervulina (TX) and their control (CTRL) at 4 and 6 days post infection.

   [95]Fig 10
   [96]Open in a new tab

   The adjusted P-values for significant comparison are presented in
   [97]S2 Table. The upper whiskers represent the maximum, and the lower
   whiskers the minimum values. The plus-signs indicate the mean values,
   while the median values are represented by the black line within the
   boxes.

   This reduction was more pronounced in the ACRB strain. Finally, in
   contrast to the decline in eicosanoids, increases in itaconate (an
   inhibitor to bacterial growth) were observed at both 4 and 6 dpi.

Nucleotide metabolism

   Based on the hierarchical clustering presented in [98]Fig 4, it is
   reasonable to infer that the time post infection (6 vs 4 dpi) had a
   dramatic impact on the level of several metabolites reporting on these
   processes with the significant increase in most of the detected purine
   (inosine, xanthine, guanine, 2’deoxyguanosine) and pyrimidine (uracil,
   cytosine, thymidine and thymine) metabolites ([99]S2 Table). In
   particular, infection with E. acervulina led to a decline in purine
   metabolism in the ACRB genotype. However, the level of sera nucleotides
   increased from 4 to 6 dpi ([100]Fig 11).

Fig 11. Differences in abundance in serum nucleoties in chicken genotypes
(ACRB and COBB) infected with Eimeria acervulina (TX) and their control
(CTRL) at 4 and 6 days post infection.

   [101]Fig 11
   [102]Open in a new tab

   The adjusted P-values for significant comparison are presented in
   [103]S2 Table.

Discussion

   In the current study, an untargeted serum metabolome of two chicken
   genotypes infected with E. acervulina was performed at 4 and 6 dpi to
   characterize the joint biochemical and metabolic changes that occurred
   during the infection. We provide the first mechanistic insight into the
   metabolic alterations that occur after E. acervulina infection during
   merozoite formation (4 dpi) and oocyst shedding (6 dpi) in ACRB and
   COBB genotypes. The Cobb 500 has been used in other studies involving
   E. acervulina [[104]22–[105]24] and the ACRB was used to study the
   genetic variability to E. acervulina and E. tenella [[106]25]. Data on
   comparative analysis of the responses of the Cobb and ACRB genotypes to
   Eimeria spp is scant.

   The PCA analysis by genotype pointed towards a differential time
   response. The ACRB genotype was developed in the 1950s and has been
   maintained as an unselected control population [[107]14], whereas the
   COBB genotype is a commercial strain under constant genetic
   improvement. Even though the body weights of both genotypes at the time
   of infection were significantly different ([108]Fig 1), they were both
   infected with the same dose of E. acervulina sporulated oocysts. Thus,
   potential parasitic load could be responsible for the differential time
   response as indicated by the PCA analysis. Infecting the two genotypes
   at similar body weight to ensure equivalent parasitic load could be
   done, however, the genotypes would be at significantly different ages
   which could introduce other sources of variation. The differential
   amino acids, nucleotide, xenobiotics and carbohydrate metabolism
   ([109]Fig 5) as predicted by the RF analysis could also contribute
   towards the differential time response of these two genotypes.

   The main biochemical changes between infected and control birds are
   differentiated by carbohydrates, cofactors, vitamins and lipid
   metabolism as indicated by the RF analysis ([110]Fig 6). In the current
   study, the serum glucose levels were unchanged between infected and
   uninfected groups regardless of the genotype ([111]S2 Table). This is
   not surprising given the known ability of chickens to maintain a stable
   blood glucose level under a variety of stressors [[112]26]. Sucrose
   levels significantly changed in the infected birds. Monosaccharide
   transporters, e.g. GLUT2, GLUT5 and SGLT1 have been shown to be
   downregulated in the gut epithelia during E. acervulina infection
   [[113]27]. However, the mRNA expression of the putative animal sucrose
   transporter (SLC45) [[114]28] in chickens remains unknown. Other
   biochemicals that changed in the infected group compared to the
   controls included alpha-tocopherol, retinal, carotene diol,
   2-aminobutyrate, itaconate and N-acetyltryptophan. Chickens infected
   with E. acervulina, which primarily affects the lower duodenum and
   jejunal mucosas [[115]29], suffer nutrient malabsorption including that
   of carotenoids [[116]30,[117]31]. The physical damage to the absorptive
   mucosa could be responsible for the reduced absorption of vitamins A,
   E, carotenoids [[118]27,[119]31,[120]32] and other cofactors leading to
   their concomitant reductions in the serum ([121]S2 Table).

   The physical disruption to the intestinal mucosa villi and epithelium
   [[122]33] and alterations in lipid transport within mucosal epithelial
   cells [[123]34] have been previously associated with lipid
   malabsorption. Young chickens infected with E. acervulina have been
   shown to have a reduction in the pH of the intestinal content [[124]35]
   which may cause an impairment of lipid hydrolysis [[125]36]. According
   to Webb et al. [[126]37], increased levels of unabsorbed lipids may
   cause diarrhea due to bacterial production of C-18 fatty acids. The
   dynamics of fatty acid metabolism depended on the genotype. In the ACRB
   birds, there was a significant reduction in long chain monosaturated
   fatty acids (LCMFA) at 4 dpi, but the opposite was observed in the COBB
   genotype. There were significant increases in LCMFA, long chain
   polyunsaturated fatty acid and fatty acid dicarboxylate levels in the
   COBB infected birds at 4 dpi compared to the ACRB birds, however,
   between 4 and 6 dpi, the COBB infected chickens showed a reduction in
   LCMFA, long chain polyunsaturated fatty acids and fatty acid
   dicarboxylate pointing to a better fatty acid absorption in the COBB
   genotype compared to their ACRB counterparts. These
   carnitine-conjugated fatty acids are transferred from the cytosol to
   the mitochondrial matrix to undergo β-oxidation ([127]Fig 9), supplying
   acetyl-CoA to the TCA cycle to generate energy. Induction of
   β-oxidation to generate energy during E. acervulina infection could be
   a mechanism to compensate for the reduced energy and protein retentions
   [[128]34]. Further, ketone body 3-hydroxybutyrate can regulate cellular
   processes directly through the inhibition of histone deacetylase,
   binding to cell surface receptors and by altering acetyl-CoA,
   succinyl-CoA, and NAD+ [[129]38]. Previous research on the effect of E.
   acervulina infection on lipid metabolism have also shown a significant
   reduction in high-density lipoproteins, low-density lipoproteins, and
   very low-density lipoproteins in infected chickens [[130]39,[131]40]
   which were all attributed to nutrient malabsorption and liver changes.
   Reduced proteins are associated with deficiency in lipotropic factors
   which are important for the synthesis of phospholipids and lipoproteins
   [[132]41]. Disruption of the epithelia of lower jejunum and upper ileum
   by E. acervulina and its negative effect on absorption of nutrients may
   be a major factor contributing to the decrease in growth of infected
   chickens ([133]Fig 1).

   Cyclooxygenase products thromboxane B2 (TXB2) and
   12-Hydroxyheptadecatrienoic acid (12-HHTrE), products of inflammation
   sites [[134]42,[135]43], decreased in both genotypes and infection
   groups at 4 dpi. However, both TXB2 and 12-HHTrE serum levels
   significantly increased at 6 dpi suggesting an early repression of
   inflammation signaling with infection. This reduction was strongest in
   the ACRB genotype. Finally, in contrast to the decline in eicosanoids,
   increases in itaconate ([136]S2 Table) were observed at both 4 and 6
   dpi. Itaconate is an anti-inflammatory metabolite that regulates
   macrophage function and is required for the activation of
   anti-inflammatory transcription factor Nrf2 [[137]44–[138]46]. Taken
   together, the changes in the aforementioned metabolites suggest a
   potential modulation of inflammation by the host in response to
   Eimeria.

   Endocannabinoids, a group of endogenous bioactive lipids possess
   immunomodulatory effects and are frequently associated with
   inflammation and pain [[139]47], often serving as a brake on
   inflammation via cannabinoid or vanilloid receptors [[140]48–[141]50].
   The levels of those biochemicals were increased in treatment versus
   control sera 4 dpi of the COBB genotype consistent with potential
   activation of immune system when responding to infection, however, at 6
   dpi, when the infected birds were compared with their control
   counterparts, levels of all detected endocannabinoids were decreased
   (oleoyl ethanolamide, N-linoleoyltaurine*, linoleoyl ethanolamide and
   N-oleoylserine) reaching statistical significance ([142]S2 Table).
   Furthermore, significant increases in levels of kynurenine (6 vs 4 dpi)
   were observed for both genotypes ([143]Fig 10). Since the formation of
   kynurenine through ubiquitous indoleamine-2,3-dioxygenase (IDO) is
   stimulated by pro-inflammatory cytokines, kynurenine may serve as a
   biomarker of inflammation [[144]51,[145]52]. Overall, serum samples
   collected post infection showed significant changes in markers of
   inflammation and oxidative stress that were altered based on time of
   collection and genotype of the bird. It should be pointed out that
   serum biomarkers reflect global metabolism and may be impacted by
   induction of factors limiting inflammatory mediators (in the liver, for
   example, which processes blood draining from the gut).

   During normal growth, nucleic acids are constantly being degraded and
   resynthesized. Purine and pyrimidine nucleotides can be broken down
   into nucleosides and nucleobases. These nucleosides can be broken down
   further or recycled back into nucleotides via salvage pathways
   [[146]53]. Comparing 4 and 6 dpi, it can be inferred that E. acervulina
   infection had a dramatic impact on the levels of several metabolites
   reporting on these processes with the significant increase in most of
   the detected purine and pyrimidine metabolites ([147]S2 Table). This
   unidirectional change is suggestive of a potential increase in DNA/RNA
   turnover and utilization of these nucleotides [[148]54]. Purine
   degradation is regulated, in part, by oxidative stress, and it is
   linked to inflammation [[149]55,[150]56] which could be reflective of
   the response to E. acervulina infection.

   There were changes in microbiome-associated products of histidine
   (imidazole lactate and imidazole propionate), phenylalanine
   (phenylpyruvate, phenylacetate and phenyllactate (PLA), tyrosine
   (phenol sulfate) and tryptophan (indolelactate, indoleacetate and
   indolepropionate) metabolism ([151]Fig 12), which could reflect
   infection associated changes in the gut microbiome, where most of the
   biochemicals are elevated in samples collected at day 6 vs 4 dpi.

Fig 12. Compounds with microbial metabolism origin in chicken genotypes (ACRB
and COBB) infected with Eimeria acervulina (TX) and their control (CTRL) at 4
and 6 days post infection.

   [152]Fig 12
   [153]Open in a new tab

   The adjusted P-values for significant comparison are presented in
   [154]S2 Table.

   The gastrointestinal tract of poultry is densely populated with
   microorganisms which closely and intensively interact with the host and
   ingested feed. In the current study, several gut microbiome derived
   metabolites were altered by infection as well as the time post
   infection, suggesting changes in the composition of microbiota. The
   metabolism of aromatic amino acids, phenylalanine, tryptophan and
   tyrosine occurs, in part, through the involvement of enzymes encoded
   within the microbiome. Collectively, the altered fatty acid signature
   may be attributable to alterations in the metabolic program of the
   intestinal microbiome contributing to differential lipid handling.
   Soluble dietary fibers and resistant starch can be fermented by gut
   microbiota, producing short chain fatty acids (SCFAs), namely acetate,
   propionate, and butyrate. In this study, levels of butyrate were
   significantly decreased due to infection in ACRB chickens compared to
   control samples collected at 6 dpi, further confirming infection
   related changes in gut microbiota. Diet supplementation with
   nutriceuticals, prebiotics and probiotics and lipids aimed at
   ameliorating some of the effects of malabsorption caused by E.
   acervulina could reduce potential dysbiosis and contraction of other
   enteric diseases.

Conclusion

   The global metabolome of chicken serum samples from two genotypes (ACRB
   and COBB) that were infected with E. acervulina and collected 4 and 6
   dpi demonstrated a significant number of changes that were treatment,
   time post infection and genotype dependent. In-depth analysis showed
   significant alterations in fatty acid metabolism that might be
   associated with changes in lipid absorption and utilization. The time
   course comparison showed significant changes in markers of inflammation
   and oxidative stress illustrating changes in response to infection.
   Changes in microbiome-associated products of histidine, phenylalanine,
   tyrosine and tryptophan metabolism were observed, which could reflect
   infection associated changes in the gut microbiome. Global profiling of
   intestinal tissues and contents would provide additional insight into
   the metabolic and microbiome effects associated with E. acervulina
   infection. The intriguing changes in inflammatory metabolites at 4 and
   6 dpi invite data collection beyond 6 dip which could lead to the
   identification of additional factors associated with host response to
   E. acervulina.

Supporting information

   S1 Table. Individual metabolite data.

   Dataset used for analysis.

   (XLSX)
   [155]Click here for additional data file.^ (350.5KB, xlsx)
   S2 Table. Metabolites that differ post infection in chicken genotypes.

   Metabolites with fold change and the level of significance by genotype,
   time and genotype by time interaction.

   (XLSX)
   [156]Click here for additional data file.^ (552.8KB, xlsx)

Data Availability

   All relevant data are within the manuscript and its Supporting
   Information files.

Funding Statement

   This work was partially supported by a cooperative grant 58-6040-8-034
   from the United State Department of Agriculture-Agricultural Research
   Service. The funder had no role in the study design and implementation,
   data collection and analysis, decision to publish or preparation of
   this manuscript.

References