Sequencing approaches to identify distal jejunum microbial community composition and function in broiler chickens fed anti-interleukin-10 during coccidiosis and necrotic enteritis challenge

Abstract

Strategies to counteract interleukin (IL)-10-mediated immune evasion by Eimeria spp. during coccidiosis- like anti-IL-10 antibodies- may protect broiler chicken health and reduce incidence of secondary necrotic enteritis (Clostridium perfringens) via undetermined mechanisms. Objectives were to use sequencing techniques to evaluate jejunal microbial community composition and function in anti-IL-10-fed broilers during coccidiosis and necrotic enteritis. On d0, Ross 308 chicks were placed in 32 cages (15 chicks/ cage) for a 25-d study and randomly assigned to diets ± 0.03% anti-IL-10. Six chicks/ diet were euthanized for distal jejunum content and tissue collection on d 14 (baseline) before inoculating the remainder with saline or 15,000 E. maxima oocysts (M6 strain). Half the chicks challenged with E. maxima were challenged with C. perfringens (1×10⁸ colony forming units) on d 18 and 19. Follow-up samples (6 chicks/treatment) were collected at 7 and 11 d postinoculation (pi) for the E. maxima-only group, or 3 and 7 dpi for the E. maxima + C. perfringens group with 3/7 dpi being designated as peak and 7/11dpi as postpeak challenge. DNA was extracted from digesta for microbiota composition analysis (16S rRNA gene sequencing) while RNA was extracted from tissue to evaluate the metatranscriptome (RNA sequencing). Alpha diversity and genus relative abundances were analyzed using the diet or challenge main effects with associated interactions (SAS 9.4; P ≤ 0.05). No baseline microbial changes were associated with dietary anti-IL-10. At peak challenge, a diet main effect reduced observed species 36.7% in chicks fed anti-IL-10 vs. control; however, the challenge effect reduced observed species and Shannon diversity 51.2-58.3% and 33.0 to 35.5%, respectively, in chicks challenged with E. maxima ± C. perfringens compared to their unchallenged counterparts (P ≤ 0.05). Low sequencing depth limited metatranscriptomic analysis of jejunal microbial function via RNA sequencing. This study demonstrates that challenge impacted the broiler distal jejunum microbiota more than anti-IL-10 while future research to characterize the microbial metatranscriptome may benefit from investigating other intestinal compartments.

INTRODUCTION

Vaccine efficacy, growing Eimeria resistance against approved pharmaceuticals, and reduced reliance on antibiotics in poultry production underlie the persistent issue of coccidiosis in global poultry production, amounting to billions of dollars in annual economic losses (Smith, et al., 2002; Arabkhazaeli, et al., 2013; Noack, et al., 2019; Blake, et al., 2020). Primary infection with Eimeria spp. increases the risk of secondary Clostridium perfringens establishment within the intestine and development of necrotic enteritis (Williams, 2005). For these reasons, both diseases are linked etiologically and interventions effective against the primary infection (coccidiosis) may inhibit necrotic enteritis development. While an important driver for the continued development of anticoccidial compounds, the 2 diseases are rarely studied concurrently.

Current research is dominated by various compounds designed to reduce the negative consequences of coccidiosis with direct and indirect modes of action against that parasite. Immunoglobulin (Ig) Y antibody technology takes advantage of avian mechanisms for maternal-offspring passive immune transfer by vaccinating hens against a target (pathogen or other molecule) and non-invasively harvesting specific antibodies deposited in the egg yolk (Trott, et al., 2009; Cook and Trott, 2010; Bobeck, et al., 2016). Previous work has demonstrated that Eimeria increases interleukin (IL)-10 production by the host, which in turn operates as a regulator of inflammatory responses that effectively enables the parasite to “hide” from the immune system (Hong, et al., 2006; Arendt, et al., 2019). Antibodies against IL-10 preserved body weight gain during mild Eimeria challenges during preliminary testing, but specific mechanisms and implications for severe coccidiosis ± secondary necrotic enteritis have been sparsely documented in published literature (Arendt, et al., 2016; Sand, et al., 2016; Rasheed, et al., 2020).

Although anti-IL-10 antibodies as potential coccidiosis mitigators hold theoretical merit from a host-centric perspective, it is also necessary to consider this cytokine's role in the maintenance of commensal microbial communities within the intestinal environment where dietary antibodies are introduced (Rooks and Garrett, 2016). Previous work in broilers challenged with coccidiosis and necrotic enteritis using a Salmonella Typhimurium- Eimeria maxima- Clostridium perfringens coinfection model demonstrated that anti-IL-10 did not affect the jejunal microbiota of unchallenged birds, nor did it protect highly abundant genera like Lactobacillus and Peptostreptococcaceae from challenge-associated reductions (Fries-Craft, et al., 2024). If anti-IL-10 does not have a protective effect on jejunal bacterial communities during coccidiosis and necrotic enteritis challenge, this could indicate that the antibody's mechanisms of action are host-specific; however, the coinfection model used previously may have introduced confounders when evaluating dietary anti-IL-10. Especially since IL-10-mediated pathways are also used by S. Typhimurium to establish in the chicken intestine (Kogut and Arsenault, 2017). This potential confounder was supported by observations that anti-IL-10 protected E. maxima-challenged broilers from additive performance losses during challenge with secondary C. perfringens in a coinfection model excluding S. Typhimurium (Fries-Craft and Bobeck, 2024a).

Microbiota research is dominated by 16S rRNA gene amplicon sequencing approaches, which are effective at providing the identity and relative abundance of bacterial community members (Johnson, et al., 2019; Marizzoni, et al., 2020). Previous work has established that coccidiosis reduces the abundance of potentially beneficial genera in favor of pathogen-associated Escherichia-Shigella and Enterococcus genera (Jebessa, et al., 2022; Campos, et al., 2023). Necrotic enteritis-associated changes to the microbiota are defined by increased Clostridium sensu stricto 1 relative abundance, which is associated with C. perfringens (Yang, et al., 2019; Pietruska, et al., 2023). Presence alone is not a sufficient indicator of microbial activity and metatranscriptome analysis through RNA sequencing (RNAseq) can provide insight into microbial community gene expression to assess community function within complex gastrointestinal environments (Shakya, et al., 2019). In poultry coccidiosis models, RNAseq approaches have previously been implemented to specifically identify response elements within the avian tissues or characterize Eimeria virulence elements, but have sparsely been used to evaluate microbial functional responses (Walker, et al., 2015; Sandholt, et al., 2021; Chen, et al., 2022). At the same time, previous work has emphasized transcriptome changes within the ceca during E. tenella infection, but necrotic enteritis is more closely associated with E. maxima and E. acervulina that target the duodenum or jejunum (Al-Sheikhly and Al-Saieg, 1980; Williams, 2005). Transcriptomic analysis involving E. maxima is subject to the same limitations as other species, where greater emphasis has been placed on the parasite or host independently rather than microbial community-wide effects (Parreira, et al., 2016; Hu, et al., 2018; Li, et al., 2019; Akerele, et al., 2022).

By implementing metatranscriptome sequencing approaches within the chicken jejunum to assess microbiota function, specific effects of E. maxima challenge ± C. perfringens on microbial community functionality can be elucidated in the location where both pathogens are expected to be the most active. Collectively, these outcomes could then identify microbial pathways to target when developing novel mitigation strategies for addressing coccidiosis and necrotic enteritis. As such, objectives in this study were to evaluate jejunal microbial community composition and their metatranscriptome in broiler chickens fed diets ± anti-IL-10 during challenge with E. maxima ± C. perfringens without S. Typhimurium as a predisposing factor.

MATERIALS AND METHODS

Animals and Sample Collection

The Iowa State University Institutional Animal Care and Use Committee approved all live animal procedures. Microbiota findings herein are part of previously published research (Fries-Craft and Bobeck, 2024a; Fries-Craft and Bobeck, 2024b; Fries-Craft, et al., 2024). In-depth description of animal husbandry conditions, sample collection timlines, feed formulation, and challenge inocula has been described by Fries-Craft and Bobeck (2024a). Ross 308 chicks were placed in 32 wire-floor cages (15 chicks/ cage) and assigned to diets ± 0.03% anti-IL-10 for 25 d. On d 14, chicks were orally gavaged with sterile phosphate-buffered saline ± 15,000 sporulated oocysts of E. maxima field strain M6 (Martin, et al., 1997). On d 18 and 19, half the Eimeria-challenged chicks were inoculated with 1 × 10⁸ colony forming units of C. perfringens TAMU strain (Ausland, et al., 2020). Digesta and 3 cm sections of the distal jejunum near Meckel's diverticulum from 6 chicks/ treatment were collected into separate 1.5 ml cryotubes and snap frozen in liquid nitrogen at baseline prior to inoculation (d 14), peak challenge (7 dpi with E. maxima/ 3 dpi with secondary C. perfringens), or post-peak challenge corresponding to (11 dpi E. maxima/ 7 dpi C. perfringens; Figure 1). Snap frozen digesta and tissue samples were transferred from liquid nitrogen to long-term storage at -80°C.

Figure 1.

Overview of (A) challenge and sample collection timelines in a study evaluating the composition and function of jejunal microbial communities in Ross 308 broilers fed diets ± 0.03% anti-IL-10 during challenge with Eimeria maxima ± Clostridium perfringens. Chicks were placed in wire-floor brooders on d 0 and assigned to (B) 6 treatment groups. On d14, baseline jejunal contents and tissue were collected for 16S rRNA gene amplicon sequencing and metatranscriptome sequencing, respectively (6 chicks/ treatment), before inoculating chicks with sterile saline or 15,000 sporulated E. maxima oocysts. Half the E. maxima-challenged chicks received 1 × 10⁸ CFU C. perfringens on d 18 and 19. Additional jejunal content and tissue samples were collected at 7 d post-inoculation (pi) with E. maxima/ 3 dpi C. perfringens (peak). Only jejunal content samples were collected at 11 dpi E. maxima/ 7 dpi C. perfrinens (post-peak).

DNA Extraction, 16S rRNA Gene Amplicon Sequencing, and Analysis

Protocols for DNA extraction, amplicon sequencing, and downstream analysis are based on previous publications and the MiSeq standard operating procedure with the use of kits being in accordance with manufactuer specifications (Kozich, et al., 2013; Wiersema et al., 2021; Fries-Craft et al., 2023). Approximately 0.2g of thawed digesta collected at each timepoint were allocated for DNA extraction using the DNeasy PowerLyzer PowerSoil kit (Qiagen, Hilden, Germany). DNA extracted from sterilized water was used for a process control and all extracted DNA samples were quantified by nanodrop (Nanodrop 2000; ThermoFisher Scientific, Waltham, MA) before -20°C storage. All sequencing and library preparation was completed by the Iowa State University DNA facility (Ames, IA) on samples diluted to 30 ng/μL before submission. PCR amplification of the 16S rRNA gene V4 variable region (515F and 806R; Caporaso, et al., 2011; Caporaso, et al., 2012) was completed before paired-end sequencing (Illumina MiSeq). Raw sequences were uploaded to the National Center for Biotechnology Information (NCBI) sequencing read archive (SRA) and are publically available (bioproject ID: PRJNA 1055975).

Quality-screening was accomplished using Mothur (V.1.43.0) in accordance with the MiSeq standard operating procedure (Schloss, et al., 2009; Kozich, et al., 2013). Briefly, merged sequences were filtered to a minimum read length (239 bp) and removed from the data set if they contained more than 0 ambiguous base pairs or a minimum of 8 bp homopolymers. Taxonomic alignment of the remaining reads was completed using the SILVA reference database (v138; Quast, et al., 2013) and additional quality screening removed any possible chimeric sequences or those aligned outside of the target 16S gene hypervariable region. 1,695,654 sequences remaining after these steps were taxonomically classified (SILVA) before clustering into de novo operational taxonomic units (OTU) at a 99% similarity threshold. OTU data were were exported to R (version 4.2.2; Phyloseq package version 1.42.0) for final quality assessment including the removal of potential contaminants within process controls, OTUs with < 10 total reads, and samples containing < 5,000 sequences. Community-wide analysis included alpha diversity measures subsampled to the lowest read depth (6,238) and Bray-Curtis distances.

16S rRNA Sequencing Statistical Analysis

Statical analysis of R-generated outputs was performed using SAS 9.4 (SAS Institute, Cary, NC). Alpha diversity measures were analyzed using the following statistical model:

$$y_{ijk}=\mu +D_i+C_j+{\left(D\times C\right)}_{ij}+e_{ijk}$$

y_ijk is the dependent variable, μ is the mean, D_i is the diet main effect, C_j is the challenge main effect, (D×C)_ij is the interaction between diet and challenge, and e_ijk is the random error.

Genus-level relative abundances were analyzed using a similar model:

$$y_{ijkl}=\mu +D_i+C_j+T_k+{\left(D\times C\right)}_{ij}+{\left(D\times T\right)}_{ik}+{\left(C\times T\right)}_{jk}+{\left(D\times C\times T\right)}_{ijk}+e_{ijkl}$$

Notation is similar to the previous model but with the additional effect of timepoint (T) and associated interactions.

Relative abundances were normalized to the trimmed mean of the M-value calculated using the log-transformed sequence counts for each sample. Data were analyzed using a generalized linear mixed model following a negative binomial with bird as the random effect (SAS 9.4, SAS Institute, Cary, NC; Robinson and Oshlack, 2010). Alpha diversity measures were analyzed using generalized linear mixed model with individual bird as the random effect. For both relative abundance and alpha diversity analysis, Tukey's post-hoc test was used to account for multiple comparisons while the false discovery rate was controlled using q-values for relative abundance data only. Outcomes were considered significantly different at P ≤ 0.05 (alpha diversity) or P and q ≤ 0.05 (relative abundances).

RNA Extraction and Metatranscriptome Sequencing

Frozen jejunal tissues collected at baseline and peak timepoints were thawed on ice with RNAlater added to stabilize RNA before extraction from ≤ 0.2 g of tissue using the Invitrogen PureLink RNA MiniKit (Thermo Fisher Scientific, Waltham, MA) according to manufacturer instructions. Extracted RNA was quantified using a Nanodrop 2000 Spectrophotometer. SUPERase•In RNase Inhibitor was added prior to removing any contaminating DNA using the Invitrogen TURBO™ DNase kit according to manufacturer recommendations (Thermo Fisher Scientific). Aliquots of DNAse-treated RNA were snap frozen in liquid nitrogen and stored at -80°C. One aliquot of RNA was thawed for quality evaluation using the Agilent 2100 Bioanalyzer with Prokaryote Total RNA Nano chip (Agilent Technologies, Santa Clara, CA). All samples had an RNA integrity number > 8 and were submitted to the Iowa State University DNA facility (Ames, IA). Prior to RNA sequencing, both eukaryote and prokaryote ribosomal RNA were removed from the samples using NEBNext rRNA depletion kits for both human/mouse/rat and bacteria (New England Biolabs, Ipswich, MA). Equal volumes of the of depletion solution for each kit were combined prior to addition to RNA samples. Library preparation was performed using the NEBNext Ultra II Directional RNA Library Prep Kit (New England Biolabs) before 100-cycle paired-end total RNA sequencing on the Illumina NovaSeq 6000 platform with the S4 flow cell on 4 lanes (2.5 billion reads/ lane; Illumina, San Diego, CA).

Metatranscriptome Sequencing Analysis

Sequence data for each .fastq file were evaluated using FastQC v0.11.9 to evaluate quality scores and confirm the absence of sequencing adapters (Andrews, 2010). Trimmomatic v0.39 was used to quality filter paired-end sequencing reads using the sliding window method (size = 4 bp, required average quality = 15) with minimal quality scores of 3 for both the leading and trailing ends and a minimum read length of 36 bp (Bolger, et al., 2014). Trimmed sequences were aligned to the broiler genome (GenBank Accession GCA_016699485.1) to remove host-associated reads using Bowtie2 v2.4.2 (Langmead and Salzberg, 2012). Residual ribosomal RNA reads were removed from the dataset using SortMeRNA v4.3.6 using the included v4.3 sensitive database for 16S, 18S, 23S, 28S, and 5/5.8S rRNA (Kopylova, et al., 2012). At the conclusion of these steps, approximately 6.8% of the original dataset remained and the total number of sequences removed by each step are presented in Table 1. Quality-screened sequences after removal of host reads and residual ribosomal rRNA are available under BioProject ID PRJNA1060284 in the NCBI sequence sequence read archive. Initial attempts to map and annotate the quality-screened RNAseq data using custom databases within Kraken 2 v2.1.2 (Wood, et al., 2019) and direct mapping approaches to select genomes with BBMap v39.01 (Bushnell, 2014) identified 40-47% unclassified reads with a high percentage of residual host reads (52-60%).

Table 1.

Overview of RNA sequencing reads from the distal jejunum of Ross 308 broilers (n = 48) during initial quality screening.

Quality control step	Total number of reads	Reduction from original1
FastQC2	18,724,232,418	N/A
Trimmomatic3	18,125,504,084	3.20%
Remove Host (BowTie24)	1,342,979,018	92.80%
Remove rRNA5	1,278,274,570	93.20%

¹FastQC provided the original number of sequences obtained from RNA sequencing and the percentage of reads removed from the data set were calculated after each quality control step. N/A = not applicable.

²Andrews, 2010.

³Bolger, et al., 2014.

⁴Langmead and Salzburg, 2012.

⁵Kopylova, et al., 2012.

Trinity De Novo Assembly

Due to the high percentage of residual host contamination and unclassified reads identified by initial Kraken 2 and BBMap strategies, a de novo co-assembly approach was implemented using Trinity (v2.15.1; Haas, et al., 2013). All reads from all samples were pooled to generate one assembly from the obtained sequencing data using the default Trinity settings without read normalization or Salmon. Within Trinity, RSEM and DESEQ2 were used to create read count matrices and identify differentially expressed transcripts between predetermined pair-wise comparisons limited to the effects of feeding control vs. anti-IL-10 within each challenge condition. Within DESEQ2, the Benjamini-Hochberg Test (False Discovery Rate procedure) was applied to account for multiple comparisons. TransDecoder (version 5.5.0; https://github.com/TransDecoder/TransDecoder) was used to predict proteins within Trinity-assembled contigs at a minimum amino acid length = 50 and eggNOG (version 5.0; Huerta-Cepas, et al., 2019) to annotate predicted proteins and assign taxonomy. Merged outputs from RSEM, DESEQ2, TransDecoder, and eggNOG were generated to provide a list of differentially expressed genes with predicted functions and log2fold changes between select pairwise comparisons. Differentially expressed genes were considered significant at P and q ≤ 0.05 .

RESULTS AND DISCUSSION

Microbial Community Composition: Alpha and Beta Diversity

Removing samples from the 16S rRNA dataset with < 5,000 sequences remaining after quality control resulted in 88 samples remaining for data analysis from the original 96 with an average of 18,670 ± 5,769 sequence reads/ sample. Beyond presenting the number of observed species, alpha diversity measures also included the Chao1 index (richness) and the Shannon diversity index (richness and evenness). Dietary anti-IL-10 did not affect baseline alpha diversity measures (Figure 2), a finding which is similar to previous outcomes evaluating anti-IL-10 effects in chicks + d 0 inoculation with S. Typhimurium (Fries-Craft, et al., 2024). This consistent finding across replicate studies indicates that 0.03% anti-IL-10 does not disrupt bacterial communities in the distal jejunum of broiler chicks in the first 14 d of life when housed in wire-floor cages.

Figure 2.

Alpha diversity measures encompassing (A) number of observed species, (B) Chao1 species richness, and (C) Shannon diversity index in distal jejunum bacterial/archaeal communities determined by 16S rRNA gene amplicon sequencing Ross 308 chicks fed diets ± 0.03% anti-IL-10 during challenge with Eimeria maxima ± Clostridium perfringens. Baseline samples were collected on d 14 before chicks were inoculated with sterile saline or 15,000 sporulated E. maxima M6 oocysts. Half the E. maxima-challenged chicks received 1 × 10⁸ CFU C. perfringens on d 18 and 19. Outcomes at the peak timepoint represent 7 d post-inoculation (pi) with E. maxima/ 3 dpi with secondary C. perfringens while post-peak represents 11 dpi with E. maxima/ 7 dpi with C. perfringens. Data represent the mean diversity measurement from 6 chicks/ treatment ± SEM. Bars grouped by brackets with different a,b labels denote significant challenge main effects. Bars labeled with different x,y labels represent a significant diet main effect. All differences were significant at P ≤ 0.05.

At peak challenge (7 dpi E. maxima/ 3dpi C. perfringens), the challenge main effect reduced observed species 51.2 to 58.3%, Chao1 richness 48.8 to 54.5%, and Shannon diversity 33.0 to 35.5% in chicks challenged with E. maxima ± C. perfringens compared to their unchallenged counterparts (P ≤ 0.02; Figure 2). This was somewhat consistent with previous research evaluating anti-IL-10 during coccidiosis and necrotic enteritis challenge using the S. Typhimurium- E. maxima- C. perfringens coinfection model where a similar reduction in the Shannon diversity index was observed in 1 of the 2 replicate studies conducted (Fries-Craft, et al., 2024). Also at peak challenge, the diet main effect reduced observed species 36.7% and Chao1 richness 25.2% in chicks fed anti-IL-10 vs. control (P ≤ 0.05; Figure 2A,B). Apparent anti-IL-10-related effects on alpha diversity during coccidiosis and necrotic enteritis challenge observed herein were not apparent in previous research implementing the S. Typhimurium- E. maxima- C. perfringens coinfection model (Fries-Craft, et al., 2024). As the work presented here represents a study replicate without the use of S. Typhimurium as a predisposing factor for necrotic enteritis, it is likely that this apparent diet effect was observable in the absence of S. Typhimurium as a potential confounder. At the same time, reductions in the observed species and Chao1 richness due to anti-IL-10 were observed in unchallenged and challenged chicks only at peak challenge and only the challenge main effect impacted broiler chicken performance at the same time (Fries-Craft and Bobeck, 2024a). As no alpha diversity changes were observed at the post-peak timepoint (11 dpi E. maxima/ 7 dpi C. perfringens; Figure 2), it is likely that observed alpha diversity reductions due to anti-IL-10 were transient and not of sufficient magnitude to translate to altered bird performance during disease challenge.

Principal coordinates analysis based on calculated Bray-Curtis measures did not demonstrate distinctive clustering patterns based on diet or challenge (Figure 3). This suggests that any changes in microbial community composition due to either challenge or anti-IL-10 were not sufficient to be reflected at the whole-community level. As a result, subsequent analysis of the distal jejunum microbiota was restricted to genus-level changes with OTU-level analysis being reserved to specifically identify changes in challenge pathogen abundance.

Figure 3.

Principal coordinates analysis based on Bray-Curtis distances in distal jejunum bacterial/archaeal communities determined by 16S rRNA gene amplicon sequencing in Ross 308 chicks fed diets ± 0.03% anti-IL-10 during challenge with Eimeria maxima ± Clostridium perfringens. Baseline samples were collected on d 14 before chicks were inoculated with sterile saline or 15,000 sporulated E. maxima M6 oocysts. Half the E. maxima-challenged chicks received 1×10⁸ CFU C. perfringens on d 18 and 19. Outcomes at the peak timepoint represent 7 d post-inoculation (pi) with E. maxima/ 3 dpi with secondary C. perfringens while post-peak represents 11 dpi with E. maxima/ 7 dpi with C. perfringens. Samples were collected from 6 chicks/ treatment at each timepoint, but were removed from analysis if they contained < 5,000 sequences.

16S rRNA Gene Sequencing: Genus-Level Changes

Lists of the 20 most-abundant bacterial genera within each challenge condition at each timepoint are presented as heat maps in Figure 4, Figure 5, Figure 6, Figure 7. Changes in the 20-most abundant genera due to a significant challenge × timepoint interaction effect within the same timepoint were evaluated before examining changes within affected genera associated with a significant timepoint×diet×challenge effect within the same challenge condition. Due to the complex nature of microbiota analysis and the number of possible comparisons within each interaction effect, restricting interaction effects to significant comparisons within the same timepoint and challenge effectively addressed the following questions: 1) does challenge with E. maxima ± C. perfringens affect relative abundance among the 20 most abundant bacterial genera? and 2) does dietary anti-IL-10 change the direction/magnitude of challenge-associated changes? Jejunal microbial communities in this study were dominated by Lactobacillus, which was expected based on previous research (Maki, et al., 2020; Zhang et al., 2022). Feeding anti-IL-10 instead of the control reduced baseline relative abundance of Weissella, Kurthia, and Carnobacterium 184-, 654-, and 15-fold, respectively (P ≤ 0.002; Figure 4B). Average relative abundances for these genera were all < 1%, further supporting that 0.03% anti-IL-10 does not impact highly abundant members of the jejunal microbiota in unchallenged broilers.

Figure 4.

Baseline bacterial genera in distal jejunum communities evaluated by 16S rRNA gene amplicon sequencing in 14 d-old Ross 308 broilers fed diets ± 0.03% anti-IL-10. Panel (A) represents the 20 most abundant bacterial genera in chicks fed control or anti-IL-10 diets with numbers in each heat map cell representing the average relative abundance for each bacterial genus. Panel (B) shows the mean relative abundance of bacterial genera affected by the diet effect in 12 chicks/ treatment ± SEM. Data points represent individual birds and samples were removed from the data set if they contained < 5,000 total sequences. Brackets with * denote significant differences, P ≤ 0.05.

Figure 5.

Bacterial genera in distal jejunum communities evaluated by 16S rRNA gene amplicon sequencing in Ross 308 broilers fed diets ± 0.03% anti-IL-10 and challenged with Eimeria maxima ± Clostridium perfringens at 7 d post-inoculation (pi) with E. maxima or 3 dpi with secondary C. perfringens (peak challenge). Chicks were inoculated with sterile saline or 15,000 sporulated E. maxima M6 oocysts on d 14 and half the E. maxima-challenged birds received 1×10⁸ CFU C. perfringens on d 18 and 19. Panels (A–C) represent the 20 most abundant bacterial genera in unchallenged, E. maxima-, or E. maxima + C. perfringens-challenged chicks fed control or anti-IL-10 diets. Numbers in each heat map cell represent the average relative abundance for each bacterial genus. Panel (D) shows the mean relative abundance of bacterial genera affected by the challenge×timepoint effect at peak challenge from 12 chicks/ treatment ± SEM. Different letter labels within each genus denote significant differences, P ≤ 0.05. Data points represent individual chicks and samples were removed from the dataset if they contained < 5,000 total sequences. Panel (E) represents genera in panel D where a significant diet×challenge effect was observed within the peak challenge timepoint from 6 chicks/ treatment ± SEM. Brackets with * denote significant differences, P ≤ 0.05.

Figure 6.

Bacterial genera in distal jejunum communities evaluated by 16S rRNA gene amplicon sequencing in Ross 308 broilers fed diets ± 0.03% anti-IL-10 and challenged with Eimeria maxima ± Clostridium perfringens at 11 d post-inoculation (pi) with E. maxima or 7 dpi with secondary C. perfringens (post-peak challenge). Chicks were inoculated with sterile saline or 15,000 sporulated E. maxima M6 oocysts on d 14 and half the E. maxima-challenged birds received 1×10⁸ CFU C. perfringens on d 18 and 19. Panels (A–C) represent the 20 most abundant bacterial genera in unchallenged, E. maxima-, or E. maxima + C. perfringens-challenged chicks fed control or anti-IL-10 diets. Numbers in each heat map cell represent the average relative abundance for each bacterial genus. Panel (D) shows the mean relative abundance of bacterial genera affected by the challenge×timepoint effect at peak challenge from 12 chicks/ treatment ± SEM. Different letter labels within each genus denote significant differences, P ≤ 0.05. Data points represent individual chicks and samples were removed from the dataset if they contained < 5,000 total sequences. Panel (E) represents genera in panel D where a significant diet×challenge effect was observed within the peak challenge timepoint from 6 chicks/ treatment ± SEM. Brackets with * denote significant differences, P ≤ 0.05.

Figure 7.

Relative abundance of Clostridium perfringens identified as OTU 13 by BLAST searching in the distal jejunum of Ross 308 chicks fed diets ± 0.03% anti-IL-10 during challenge with Eimeria maxima ± C. perfringens. Baseline samples were collected before inoculating chicks with sterile saline or 15,000 sporulated E. maxima M6 oocysts. Half the E. maxima-inoculated chicks were challenged with 1×10⁸ colony forming units of C. perfringens on d 18 and 19. Peak challenge represents 7 d post-inoculation (pi) with E. maxima/ 3 dpi with C. perfringens. Post-peak challenge represents 11 dpi with E. maxima/ 7 dpi with C. perfringens. Bars represent the average relative abundance of C. perfringens from 6 chicks/ treatment at each timepoint while data points represent the relative C. perfringens abundance observed in individual animals.

At peak challenge, E. maxima ± C. perfringens reduced highly abundant Peptostreptococcaceae 1,662- to 2,308-fold compared to unchallenged chicks (P < 0.0001) with additional 3- to 626-fold reductions observed in Streptococcus, Romboutsia, Staphylococcus, Weissella, Leuconostoc, Staphylococcaeae unclassified, Faecalibacterium, Butyricicoccus, Carnobacterium, Lactoccous, Macrococcus, and Blautia (P ≤ 0.01). Challenging chicks with secondary C. perfringens contributed to 65-fold reduced Staphylococcus and 10-fold reduced Lachnospiraceae unclassified and Leuconostoc relative abundance with undetectable Macrococcus compared to chicks challenged with only E. maxima (P ≤ 0.02). Chicks challenged with E. maxima + C. perfringens had 69- to 987-fold greater Clostridium sensu stricto 1 relative abundance than unchallenged and E. maxima-only chicks at peak challenge (P ≤ 0.0004; Figure 5D). Reductions in Romboutsia and Staphylococcus during both coccidiosis and necrotic enteritis challenge along with the increase in Clostridium sensu stricto 1 were consistent with previous studies and were expected changes due to these challenges (Macdonald et al., 2017, Yang, et al., 2019; Campos, et al., 2023; Fries-Craft, et al., 2024). Challenge-associated reductions in the relative abundance of Peptostreptococcaceae, a highly-abundant member of the unchallenged broiler jejunal microbiota (> 28%), were unique to this study and not previously observed in similar work using the S. Typhimurium- E. maxima- C. perfringens coinfection model (Fries-Craft, et al., 2024). This indicates that challenge models without S. Typhimurium may produce shifts in more abundant members of the distal jejunum microbiota; however, the relevance of this finding is uncertain as models ± S. Typhimurium produced similar challenge-associated performance losses and further research to specifically evaluate the role of S. Typhimurium during necrotic enteritis is needed (Fries-Craft and Bobeck, 2024a).

Among genera affected by challenge at the peak timepoint, feeding anti-IL-10 instead of the control diet reduced Weissella 38-fold, Faecalibacterium 19-fold, and Blautia 23-fold, while making Butyricicoccus undetectable compared to their control-fed counterparts during challenge with E. maxima only (P ≤ 0.002). Similarly, feeding anti-IL-10 during E. maxima + C. perfringens challenged reduced Faecalibacterium 74-fold while making Weissella and Blautia relative abundance undetectable compared to the control (P ≤ 0.002; Figure 5E). These findings demonstrate that whenever changes were observed due to dietary anti-IL-10 during coccidiosis or necrotic enteritis challenge, they were most apparent as a reduction in genus-level relative abundances. This indicates that while anti-IL-10 did not affect microbial communities in unchallenged broilers, it also did not preserve the relative abundance of bacterial genera during enteritic disease challenge. Notably, secondary challenge with C. perfringens increased the magnitude of Faecalibacterium, Weissella, and Blautia relative abundance losses in this study (Figure 5E). While members of the Weissella genus are associated with either probiotic or pathogenic qualities, Faecalibacterium is associated with butyrate production and maintenance of intestinal barrier function (Abriouel, et al., 2015; Polansky et al., 2015). Higher magnitude losses in these genera during secondary C. perfringens challenge may be underlying the observed additive performance reductions that were unique to necrotic enteritis models without S. Typhimurium, but further research is necessary (Fries-Craft and Bobeck, 2024a).

At the post-peak timepoint, unchallenged chicks had greater relative abundance of Staphylococcus and Macrococcus (3- to 24-fold) than chicks challenged with E. maxima ± C. perfringens (P ≤ 0.05). At the same time, chicks challenged with only E. maxima had 16-fold greater Romboutsia and 25-fold greater Negativibacillus than their unchallenged counterparts (P ≤ 0.04). Chicks challenged with E. maxima + C. perfringens had 5-fold greater Romboutsia, 7-fold greater Lachnospiraceae, 5-fold greater Erysipelatoclostridium, and 12-fold greater Clostridia UCG-014 compared to their unchallenged counterparts (P ≤ 0.04; Figure 6D). The increased Romboutsia relative abundance in chicks challenged with E. maxima ± C. perfringens at post-peak challenge could indicate recovery in this genus within the 4 d between sampling timepoints while the difference in fold-change magnitude between chicks challenged with E. maxima vs. E. maxima + C. perfringens (16- vs. 5-fold) suggests that secondary C. perfringens challenge may alter Romboutsia recovery timelines (Figure 6D). Peptostreptococcaceae recovery was also observable by its return to being the 2nd most abundant bacterial genus in the distal jejunum at post-peak challenge after being displaced by genera like Lachnospiraceae and Escherichia at peak challenge in chicks inoculated with E. maxima ± C. perfringens (Figures 5A–5Figure 5, Figure 6–6C). Amongst the altered genera at post-peak challenge, feeding anti-IL-10 instead of control during E. maxima challenge increased Weissella 66-fold, but reduced Clostridia UCG-014 70-fold and made Negativibacillus and Erysipelatoclostridium undetectable (P ≤ 0.01). Feeding anti-IL-10 during E. maxima + C. perfringens challenge contributed to 2-fold greater Erysipelatoclostridium and 29-fold greater Clostridia UCG-014 relative abundance compared to chicks fed control diets at the post-peak timepoints (P ≤ 0.01; Figure 6E).

While OTU-level statistical analyses were not performed, OTU 13 within the Clostridium sensu stricto 1 genus was identified as C. perfringens via NCBI BLAST searching. Clostridium perfringens relative abundance was highest at the peak challenge timepoint in chicks challenged with secondary C. perfringens, as expected. Relative abundance of C. perfringens was not affected by dietary anti-IL-10 (Figure 7).

Metatranscriptome Sequencing Outcomes

Evaluating microbial composition alone does not adequately describe community functions and microbial gene expression in the distal jejunum microbiota was evaluated using metatranscriptome sequencing. After quality screening and host read removal, 93% of the 18.7 billion total sequence reads were removed from the dataset (Table 1). Initial mapping and annotation approaches using Kraken 2 and BBMap allowed direct mapping of the remaining reads to genomes likely present in the chicken jejunum using publicly available tools to minimize computational memory requirements (Wood, et al., 2019; Jurado-Rueda, et al., 2023). The high percentage of unclassified reads after quality control emphasizes that Kraken 2 and direct mapping approaches are significantly limited by the contents of their associated databases and the paucity of information available to functionally assess the microbial metatranscriptome within the chicken intestine. This finding meant that initial reductions in the dataset were greater than initially observed, making it necessary to implement de novo assembly approaches to potentially characterize the approximately 40% unclassified reads remaining.

The de novo assembly generated from all samples in the dataset resulted in a total of 442,842 transcripts with an N50 length of 1,514 bases. Metatranscriptome sequencing reads were mapped to contigs for transcript taxonomic classification, which resulted in 3.23% being identified as Bilateria (potential host reads not initially removed by Bowtie2) with 0.35% attributed to Bacteria and 0.01% classified as Apicomplexa. When considering the metatranscriptome of microbial communities including Bacteria, Apicomplexa, Fungi, Archaea, and “other” reads, this accounts for approximately 33,345,818 total reads (0.4% of the total) with an estimated 694,704 reads/ sample (Table 2). At < 1 million reads/ sample, it was uncertain whether the data would yield high-quality insight into differentially expressed genes with functional annotations or biological relevance in the context of feeding anti-IL-10 during disease challenge. The low sequencing depth of microbial mRNA observed here is thus not sufficient for in-depth characterization of the gene expression profiles of entire microbial communities. A similar metatranscriptome sequencing approach of rumen wall microbial communities obtained much higher microbial sequencing depth retaining around 6% of all reads after quality control and host removal (Mann, et al., 2018). For this study, comparisons of metatranscriptome sequencing data were limited to significant outcomes observed when feeding anti-IL-10 vs. control to unchallenged, E. maxima, or E. maxima + C. perfringens-challenged chicks at the peak challenge timepoint only. Gene annotations were evaluated and grouped into broad categories associated with metabolism, DNA replication, cell wall/membrane/envelope maintenance, transcription, translation-associated elements, signal transduction, defense mechanisms, and unknown functions (Table 3, Table 4, Table 5).

Table 2.

Number of RNA sequencing reads from the Ross 308 broiler chicken distal jejunum (n=48) assigned to each taxonomic designation after de novo assembly.¹

Category	Total reads	Composition2
Initial	∼ 9,000,000,000	N/A
Bilateria3	290,530,681	3.23%
Bacteria	31,720,613	0.35%
Apicomplexa	1,135,205	0.01%
Fungi	133,460	0.0015%
Archaea	335	0.0000037%
Other	3,564,205	0.04%

¹RSEM was used to create abundance matrices by mapping to contigs assembled with Trinity (Haas, et al., 2013). Predicted proteins at a minimum length of 50 amino acids were identified by TransDecoder (version 5.5.0; https://github.com/TransDecoder/TransDecoder) and taxonomy was assigned using eggNOG (version 5.0; Huerta-Cepas, et al., 2019).

²Percentage of initial reads attributed to each taxonomic level were calculated to determine taxonomic composition. N/A = not applicable.

³Bilateria designation broadly encompasses animal taxonomies that include the chicken host. Remaining reads attributed to these sources may represent host sequences that were not removed during initial quality screening steps.

Table 3.

Differentially expressed genes in the distal jejunum microbiota of unchallenged Ross 308 chicks fed diets ± 0.03% anti-IL-10 at 21 d of age.

Category	Taxonomy	Annotation	Gene name	Control mean	Anti-IL-10 mean	Log2 fold-change	P - value1
Energy production and conversion	Comamonadaceae	PFAM acyl-CoA dehydrogenase domain protein	-	1059.13	39.13	4.75	0.05
	Comamonadaceae	Nitrite sulfite reductase, hemoprotein beta-component, ferrodoxin domain protein	nirB	374.47	12.05	4.94	0.04
	Comamonadaceae	Aconitate hydratase	acnA	74.28	1.27	5.77	0.05
	Comamonadaceae	TIGRFAM HAD-superfamily hydrolase, subfamily IA, variant 3	-	1059.13	39.13	4.75	0.05
Amino acid transport and metabolism	Comamonadaceae	Belongs to the peptidase S51 family	pepE	1059.13	39.13	4.75	0.05
	Comamonadaceae	Peptidase M20	-	481.42	14.63	5.03	0.03
	Comamonadaceae	Acyclic terpene utilisation family protein AtuA	-	229.53	6.88	5.04	0.05
	Comamonadaceae	Major facilitator superfamily	-	481.42	14.63	5.03	0.03
Carbohydrate transport and metabolism	Lactobacillaceae	ROK family	scrK	4.48	76.55	-3.95	0.05
Lipid transport and metabolism	Comamonadaceae	Short-chain dehydrogenase reductase sdr	-	1059.13	39.13	4.75	0.05
Inorganic ion transport and metabolism	Lactobacillaceae	CorA-like Mg2+ transporter protein	corA	4.48	76.55	-3.95	0.05
	Comamonadaceae	TIGRFAM nitrite reductase NAD(P)H	nirD	374.47	12.05	4.94	0.04
Coenzyme transport and metabolism	Lactobacillaceae	Catalyzes the condensation of isopentenyl diphosphate with allylic pyrophosphates generating different type of terpenoids	uppS	0.00	79.24	-8.76	0.04
	Comamonadaceae	TIGRFAM acetolactate synthase, large subunit, biosynthetic type	ilvI	241.70	10.24	4.54	0.05
	Comamonadaceae	Belongs to the class-III pyridoxal-phosphate-dependent aminotransferase family	-	268.82	7.55	5.12	0.04
	Comamonadaceae	Mur ligase, middle domain protein	cphA1	428.04	18.80	4.50	0.05
Nucleotide transport and metabolism	Lactobacillaceae	Catalyzes the reversible phosphorylation of UMP to UDP	pyrH	0.00	79.24	-8.76	0.04
Cell cycle control, cell division, chromosome partitioning	Comamonadaceae	SMC domain protein	-	481.42	14.63	5.03	0.03
Cell wall/membrane/envelope biogenesis	Lactobacillaceae	zinc metalloprotease	rseP	0.00	79.24	-8.76	0.04
	Comamonadaceae	RimK-like ATP-grasp domain	cphA2	428.04	18.80	4.50	0.05
	Lactobacillaceae	Belongs to the CDS family	cdsA	0.00	79.24	-8.76	0.04
Replication, Recombination and Repair	Lactobacillaceae	Required for replicative DNA synthesis	polC	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Required for DNA replication, binds preferentially to single-stranded, linear DNA	recF	4.48	76.55	-3.95	0.05
	Lactobacillaceae	Participates in initiation and elongation during chromosome replication	dnaB	4.48	76.55	-3.95	0.05
	Lactobacillaceae	Plays an important role in DNA replication, recombination and repair	ssb	4.48	76.55	-3.95	0.05
	Lactobacillaceae	A type II topoisomerase that negatively supercoils closed circular double-stranded (ds) DNA	gyrA	4.48	76.55	-3.95	0.05
	Lactobacillaceae	A type II topoisomerase that negatively supercoils closed circular double-stranded (ds) DNA	gyrB	4.48	76.55	-3.95	0.05
	Comamonadaceae	ATP dependent DNA ligase domain protein	lig	374.47	12.05	4.94	0.04
	Burkholderiales	Catalyzes the recognition and processing of DNA lesions	uvrA2	268.82	7.55	5.12	0.04
Transcription	Lactobacillaceae	Participates in both transcription termination and antitermination	nusA	0.00	79.24	-8.76	0.04
	Bacteria	sequence-specific DNA binding	MA20_10370	1059.13	39.13	4.75	0.05
	Comamonadaceae	response regulator	-	1059.13	39.13	4.75	0.05
	Comamonadaceae	Transcriptional regulator	-	229.53	6.88	5.04	0.05
	Comamonadaceae	Facilitates transcription termination by a mechanism that involves Rho binding	rho	96.68	1.62	5.81	0.05
Translation, ribosomal structure and Biogenesis	Lactobacillaceae	Belongs to the universal ribosomal protein uS2 family	rpsB	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Required for maturation of 30S ribosomal subunits	rimP	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Associated with the 70S Ribosome	rplV	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Belongs to the universal ribosomal protein uL29 family	rpmC	8.47	205.93	-4.52	0.05
	Lactobacillaceae	This protein binds to the 23S rRNA, and is important in its secondary structure	rplF	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Binds directly to 16S rRNA central domain where it helps coordinate assembly of the platform of the 30S subunit	rpsH	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Binds 23S rRNA and is also seen to make contacts with the A and possibly P site tRNAs	rplP	8.47	205.93	-4.52	0.05
	Lactobacillaceae	One of the primary rRNA binding proteins, it binds specifically to the 5′-end of 16S ribosomal RNA	rpsQ	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Binds the lower part of the 30S subunit head. Binds mRNA in the 70S ribosome, positioning it for translation	rpsC	8.47	205.93	-4.52	0.05
	Lactobacillaceae	One of the proteins that surrounds the polypeptide exit tunnel on the outside of the subunit	rplX	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Binds and probably mediates the attachment of the 5S RNA into the large ribosomal subunit	rplE	8.47	205.93	-4.52	0.05
	Lactobacillaceae	Binds to the 23S rRNA	rplI	4.48	76.55	-3.95	0.05
	Lactobacillaceae	Helps stabilize the platform of the 30S subunit	rpsR	4.48	76.55	-3.95	0.05
	Lactobacillaceae	Binds together with S18 to 16S ribosomal RNA	rpsF	4.48	76.55	-3.95	0.05
	Lactobacillaceae	Protects formylmethionyl-tRNA from spontaneous hydrolysis and promotes its binding to the 30S ribosomal subunits	infB	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Associates with the EF-Tu.GDP complex and induces the exchange of GDP to GTP	tsf	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Catalyzes the attachment of proline to tRNA(Pro)	proS	0.00	79.24	-8.76	0.04
	Lactobacillaceae	Responsible for the release of ribosomes from messenger RNA at the termination of protein biosynthesis	frr	0.00	79.24	-8.76	0.04
	Lactobacillaceae	S4 domain protein YaaA	yaaA	4.48	76.55	-3.95	0.05
Post-translational modification, protein turnover, Chaperones	Comamonadaceae	Belongs to the thioredoxin family	trxA	96.68	1.62	5.81	0.05
Signal transduction mechanisms	Comamonadaceae	Diguanylate cyclase	-	558.52	18.03	4.94	0.04
	Comamonadaceae	Diguanylate cyclase	-	481.42	14.63	5.03	0.03
	Lactobacillaceae	Signaling protein consisting of a modified GGDEF domain and a DHH domain	yybT	4.48	76.55	-3.95	0.05
	Comamonadaceae	Histidine kinase	-	1059.13	39.13	4.75	0.05
	Comamonadaceae	Histidine kinase A domain protein	-	558.52	18.03	4.94	0.04
Defense mechanisms	Comamonadaceae	ABC transporter, transmembrane region	-	428.04	18.80	4.50	0.05
	Comamonadaceae	Rhs element vgr protein	-	268.82	7.55	5.12	0.04
	Lactobacillaceae	Protein of unknown function	ylxR	0.00	79.24	-8.76	0.04
Function unknown	Comamonadaceae	RDD family	-	241.70	10.24	4.54	0.05
	Comamonadaceae	Protein conserved in bacteria	-	558.52	18.03	4.94	0.04
	Comamonadaceae	PFAM conserved	-	229.53	6.88	5.04	0.05

¹P-values adjusted for multiple comparisons using the Benjamini-Hochberg Test (False Discovery Rate procedure) in DESEQ2 within Trinity (v2.15.1).

Table 4.

Differentially expressed genes in the distal jejunum microbiota of Ross 308 chicks fed diets ± 0.03% anti-IL-10 7 d postinoculation with Eimeria maxima.

Category	Taxonomy	Annotation	Gene Name	Control Mean	Anti-IL-10 Mean	Log2 Fold-Change	P – Value1
Energy Production and Conversion	Propionibacteriales	Belongs to the complex I 49 kDa subunit family	nuoD	0.31	9.12	-4.88	0.02
	Methylobacteriaceae	Involved in the catabolism of oxalate and in the adaptation to low pH	frc	13.33	0.00	6.04	0.02
	Methylobacteriaceae	Catalyzes the pyrimidine ring opening between N-3 and C- 4 by an unusual flavin hydroperoxide-catalyzed mechanism to yield ureidoacrylate peracid	rutA	30.07	0.00	22.30	0.00
Amino Acid Transport and Metabolism	Methylobacteriaceae	Catalyzes the transfer of the enolpyruvyl moiety of phosphoenolpyruvate (PEP) to the 5-hydroxyl of shikimate-3- phosphate (S3P)	aroA	28.17	0.00	22.21	5.62E-10
Lipid Transport and Metabolism	Methylobacteriaceae	Involved in unsaturated fatty acids biosynthesis	fabZ	27.12	0.17	6.59	0.01
Inorganic Ion Transport and Metabolism	Methylobacteriaceae	PFAM binding-protein-dependent transport systems inner membrane component	-	20.36	0.00	6.65	0.04
	Methylobacteriaceae	PFAM binding-protein-dependent transport systems inner membrane component	-	20.36	0.00	6.65	0.04
	Methylobacteriaceae	PFAM ABC transporter	-	20.36	0.00	6.65	0.04
	Bradyrhizobiaceae	Permease family	-	24.74	0.00	22.01	7.32E-10
	Methylobacteriaceae	NMT1-like family	-	20.36	0.00	6.65	0.04
	Methylobacteriaceae	NMT1-like family	-	20.36	0.00	6.65	0.04
	Bradyrhizobiaceae	Belongs to the UPF0276 family	-	24.74	0.00	22.01	7.32E-10
Coenzyme Transport and Metabolism	Propionibacteriales	Catalyzes the condensation of para-aminobenzoate (pABA) with 6-hydroxymethyl-7,8-dihydropterin diphosphate (DHPt-PP) to form the immediate precursor of folate derivatives	folP	0.31	9.12	-4.88	0.02
	Propionibacteriales	7,8-dihydro-6-hydroxymethylpterin-pyrophosphokinase	folK	0.31	9.12	-4.88	0.02
	Actinobacteria	Catalyzes the conversion of 7,8-dihydroneopterin to 6- hydroxymethyl-7,8-dihydropterin	folB	0.31	9.12	-4.88	0.02
Nucleotide transport and metabolism	Methylobacteriaceae	Belongs to the cytidylate kinase family	cmk	28.17	0.00	22.21	5.62E-10
Cell wall/membrane/envelope biogenesis	Methylobacteriaceae	Belongs to the membrane fusion protein (MFP) (TC 8.A.1) family	-	13.33	0.00	6.04	0.02
	Methylobacteriaceae	Belongs to the membrane fusion protein family	-	13.33	0.00	6.04	0.02
	Propionibacteriales	OmpA family	-	0.31	9.12	-4.88	0.02
	Methylobacteriaceae	Part of the outer membrane protein assembly complex, which is involved in assembly and insertion of beta-barrel proteins into the outer membrane	bamA	27.12	0.17	6.59	0.01
	Methylobacteriaceae	Involved in the biosynthesis of lipid A, a phosphorylated glycolipid that anchors the lipopolysaccharide to the outer membrane of the cell	lpxA	27.12	0.17	6.59	0.01
	Methylobacteriaceae	Condensation of UDP-2,3-diacylglucosamine and 2,3- diacylglucosamine-1-phosphate to form lipid A disaccharide, a precursor of lipid A	lpxB	27.12	0.17	6.59	0.01
	Methylobacteriaceae	PFAM Glycosyl transferase family 2	-	30.07	0.00	22.30	5.44E-10
	Methylobacteriaceae	PFAM Glycosyl transferase family 2	-	30.07	0.00	22.30	5.44E-10
Transporters	Sarcocystidae	BT1 family	-	14.23	0.00	6.14	0.01
	Sarcocystidae	Multi-pass transmembrane protein	-	11.30	0.27	5.32	0.02
	Bradyrhizobiaceae	DoxX-like family	-	24.74	0.00	22.01	7.32E-10
Transcription	Methylobacteriaceae	PFAM regulatory protein TetR	-	13.33	0.00	6.04	0.02
RNA processing and modification	Sarcocystidae	Helicase associated domain	-	14.72	0.63	4.40	0.02
Translation, ribosomal structure, and biogenesis	Lactobacillaceae	Catalyzes the GTP-dependent ribosomal translocation step during translation elongation	fusA	0.00	170.88	-24.25	1.29E-11
	Methylobacteriaceae	PFAM formyl transferase domain protein	-	28.17	0.00	22.21	5.62E-10
	Methylobacteriaceae	May reduce aminoacrylate peracid to aminoacrylate	rutC	30.07	0.00	22.30	5.44E-10
	Lactobacillaceae	Stabilizes bases of the 16S rRNA that are involved in tRNA selection	rpsL	0.00	170.88	-24.25	1.29E-11
	Lactobacillaceae	One of the primary rRNA binding proteins	rpsG	0.00	170.88	-24.25	1.29E-11
Post-translation modification, protein turnover, chaperones	Propionibacteriales	Acts as a processive, ATP-dependent zinc metallopeptidase for both cytoplasmic and membrane proteins	ftsH	0.31	9.12	-4.88	0.02
	Propionibacteriales	C-terminal, D2-small domain, of ClpB protein	clpC	0.31	9.12	-4.88	0.02
Cell motility	Bacteria	Aspartic-type endopeptidase activity	pppA	0.31	9.12	-4.88	0.02
Cytoskeleton	Sarcocystidae	Dynamitin	-	44.81	12.05	1.87	0.05
	Sarcocystidae	Belongs to the TRAFAC class myosin-kinesin ATPase superfamily	-	26.08	0.70	5.18	0.01
Secondary metabolites biosynthesis, transport, and catabolism	Methylobacteriaceae	Hydrolyzes the ureidoacrylate peracid to avoid toxicity	rutB	30.07	0.00	22.30	5.44E-10
Signal transduction mechanisms	Bradyrhizobiaceae	Adenylyl- / guanylyl cyclase, catalytic domain	-	24.74	0.00	22.01	7.32E-10
Defense mechanisms	Methylobacteriaceae	Belongs to the resistance-nodulation-cell division (RND) (TC 2.A.6) family	-	13.33	0.00	6.04	0.02
	Beijerinckiaceae	ABC transporter transmembrane region	atm1	20.36	0.00	6.65	0.04
Function unknown	Methylobacteriaceae	Protein of unknown function	cdsA2	27.12	0.17	6.59	0.01
	Methylobacteriaceae	Protein of unknown function	-	28.17	0.00	22.21	5.62E-10
	Methylobacteriaceae	Domain of unknown function	-	30.07	0.00	22.30	5.44E-10
	Bradyrhizobiaceae	Protein of unknown function	-	24.74	0.00	22.01	7.32E-10

¹P-values adjusted for multiple comparisons using the Benjamini-Hochberg Test (False Discovery Rate procedure) in DESEQ2 within Trinity (v2.15.1).

Table 5.

Differentially expressed genes in the distal jejunum microbiota of Eimeria maxima-challenged Ross 308 chicks fed diets ± 0.03% anti-IL-10 3 d post-inoculation with Clostridium perfringens.

Category	Taxonomy	Annotation	Gene name	Control mean	Anti-IL-10 mean	Log2 fold-change	P – value1
Energy production and conversion	Comamonadaceae	Belongs to the heme-copper respiratory oxidase family	ccoN	160.54	6.49	4.59	0.05
	Comamonadaceae	C-type cytochrome. Part of the cbb3-type cytochrome c oxidase complex	ccoP	160.54	6.49	4.59	0.05
	Sphingomonadales	Aerobic-type carbon monoxide dehydrogenase, small subunit CoxS	-	0.00	11.25	-5.88	0.03
	Comamonadaceae	PFAM 4Fe-4S ferredoxin iron-sulfur binding domain protein	ccoG	160.54	6.49	4.59	0.05
	Methylobacteriaceae	PFAM cytochrome c class I	cycA	14.98	0.39	4.88	0.003
	Sphingomonadales	Ferredoxin	-	0.00	11.25	-5.88	0.03
	Methylobacteriaceae	Catalyzes the cleavage of beta-carotene at its central double bond (15,15′) to yield 2 molecules of all-trans-retinal	-	58.47	14.20	2.05	0.04
	Methylobacteriaceae	Oxidizes proline to glutamate for use as a carbon and nitrogen source	putA	58.47	14.20	2.05	0.04
	Methylobacteriaceae	PFAM FAD linked oxidase domain protein	-	58.47	14.20	2.05	0.04
	Comamonadaceae	Cytochrome c oxidase, cbb3-type, subunit II	ccoO	160.54	6.49	4.59	0.05
	Bradyrhizobiaceae	Acetamidase/Formamidase family	MA20_29860	25.05	0.00	7.15	0.02
	Bradyrhizobiaceae	Acetamidase/Formamidase family	-	25.05	0.00	7.15	0.02
	Bradyrhizobiaceae	Polyhydroxyalkanoate synthesis repressor	phaR	25.05	0.00	7.15	0.02
	Bradyrhizobiaceae	Polyhydroxyalkanoate synthesis repressor	phaR	25.05	0.00	7.15	0.02
	Sphingomonadales	NADPH-dependent FMN reductase	-	0.00	29.82	-7.28	0.03
	Sphingomonadales	Dioxygenase	-	0.00	11.25	-5.88	0.03
Amino acid transport and metabolism	Sphingomonadales	Overlaps another CDS with the same product name	-	0.00	11.25	-5.88	0.03
	Sphingomonadales	Overlaps another CDS with the same product name	-	0.00	11.25	-5.88	0.03
	Sphingomonadales	Belongs to the acetylglutamate kinase family. ArgB subfamily	argB	0.00	29.82	-7.28	0.03
	Comamonadaceae	TIGRFAM O-acetylhomoserine O-Acetylserine sulfhydrylase	metY	160.54	6.49	4.59	0.05
	Methylobacteriaceae	Bacterial extracellular solute-binding proteins, family 5 Middle	-	58.47	14.20	2.05	0.04
	Methylobacteriaceae	PFAM Anthranilate synthase component I	pabB	0.00	10.17	-5.76	0.0009
	Methylobacteriaceae	TIGRFAM glutamine amidotransferase of anthranilate synthase	pabA	0.00	10.17	-5.76	0.0009
	Methylobacteriaceae	PFAM Aminotransferase class IV	-	0.00	10.17	-5.76	0.0009
Carbohydrate transport and metabolism	Sphingomonadales	Transporter	-	0.00	11.25	-5.88	0.03
	Sphingomonadales	Epimerase	-	0.00	29.82	-7.28	0.03
Lipid transport and metabolism	Bradyrhizobiaceae	Belongs to the thiolase family	phbA	25.05	0.00	7.15	0.02
	Bradyrhizobiaceae	Belongs to the thiolase family	phbA	25.05	0.00	7.15	0.02
	Comamonadaceae	PFAM AMP-dependent synthetase and ligase	matB	160.54	6.49	4.59	0.05
Inorganic Ion transport and metabolism	Methylobacteriaceae	flavoprotein involved in K transport	-	7.02	0.57	4.50	0.04
	Comamonadaceae	heavy metal translocating P-type ATPase	ccoI	160.54	6.49	4.59	0.05
	Methylobacteriaceae	L-lysine 6-monooxygenase (NADPH-requiring)	-	7.02	0.57	4.50	0.04
	Sphingomonadales	Nickel/cobalt transporter regulator	-	0.00	11.25	-5.88	0.03
Coenzyme transport and metabolism	Methylobacteriaceae	Autoinducer synthase	-	9.00	0.46	4.41	0.002
	Sphingomonadales	Involved in the import of queuosine (Q) precursors, required for Q precursor salvage	-	0.00	29.82	-7.28	0.03
Nucleotide transport and metabolism	Sphingomonadales	Catalyzes the oxidation of 5,10- methylenetetrahydrofolate	folD	0.00	29.82	-7.28	0.03
	Methylobacteriaceae	PFAM MazG nucleotide pyrophosphohydrolase	mazG	0.00	10.17	-5.76	0.00
	Methylobacteriaceae	AIR synthase related protein, C-terminal domain	-	7.02	0.57	4.50	0.04
Cell wall/ membrane/ envelope biogenesis	Sphingomonadales	Penicillin-binding Protein	-	0.00	29.82	-7.28	0.03
Transporters	Sphingomonadales	Integral membrane protein	-	0.00	29.82	-7.28	0.03
	Methylobacteriaceae	DoxX-like family	-	0.00	10.17	-5.76	0.0009
	Methylobacteriaceae	Bacteriorhodopsin-like protein	-	58.47	14.20	2.05	0.04
Replication, recombination, and repair	Methylobacteriaceae	PFAM transposase mutator type	-	7.85	0.00	5.64	0.0003
Translation, ribosomal structure, and biogenesis	Methylobacteriaceae	PFAM Amidase	-	14.98	0.39	4.88	0.003
	Methylobacteriaceae	Catalyzes the methylthiolation of an aspartic acid residue of ribosomal protein S12	rimO	0.00	10.17	-5.76	0.001
Posttranslational modification, protein turnover, chaperones	Sphingomonadales	Prolyl 4-hydroxylase alpha subunit homologues.	-	0.00	11.25	-5.88	0.03
	Comamonadaceae	PFAM Cbb3-type cytochrome oxidase component	ccoQ	160.54	6.49	4.59	0.05
Intracellular trafficking, secretion, and vesicular transport	Sphingomonadales	UPF0056 membrane protein	-	0.00	29.82	-7.28	0.03
Defense mechanisms	Sphingomonadales	Transport permease protein	-	0.00	10.11	-5.71	0.05
	Bradyrhizobiaceae	ABC transporter	atm1	8.49	0.00	5.71	0.02
	Sphingomonadales	Beta-lactamase enzyme family	-	0.00	10.11	-5.71	0.05
Function unknown	Sphingomonadales	Belongs to the UPF0235 family	-	0.00	29.82	-7.28	0.03
	Sphingomonadales	Alpha Beta	-	0.00	11.25	-5.88	0.03
	Sphingomonadales	Protein conserved in bacteria	-	0.00	29.82	-7.28	0.03
	Methylobacteriaceae	Domain of unknown function (DUF892)	-	0.00	10.17	-5.76	0.001
	Methylobacteriaceae	Uncharacterized protein conserved in bacteria (DUF2218)	-	58.47	14.20	2.05	0.04
	Methylobacteriaceae	Protein of unknown function (DUF559)	-	9.00	0.46	4.41	0.002
	Methylobacteriaceae	PFAM ThiJ PfpI domain protein	-	14.98	0.39	4.88	0.003

¹P-values adjusted for multiple comparisons using the Benjamini-Hochberg Test (False Discovery Rate procedure) in DESEQ2 within Trinity (v2.15.1)

Many of the differentially expressed genes within each comparison were associated with metabolic pathways or general ‘housekeeping’ functions such as cell wall biogenesis, DNA replication/repair, transcription, translation, and ribosomal structure/biogenesis (Table 3, Table 4, Table 5). These findings are consistent with a published metatranscriptome of rumen wall microbial communities and transcriptomes of bacterial isolate pure cultures, indicating that such results are characteristic of metatranscriptome analysis and emphasizing the complexities of evaluating these datasets for findings that may have biological relevance to host health (Parreira, et al., 2016; Mann, et al., 2018; Anast and Schmitz-Esser, 2020). Within samples collected from unchallenged chicks, transcripts were taxonomically affiliated with Lactobacillaceae (52.3%), Comamonadaceae (44.6%), and Burkholderiales (0.01%; Figure 8A). Amplicon sequencing outcomes for unchallenged chicks at the peak challenge timepoint demonstrated that average Lactobacillus relative abundance as the main genus identified within the Lactobacillaceae family was 59% whereas average relative abundance of any genera within the order Burkholderiales, which includes Comamonadaceae, was too low (0.0002%) to be featured among the 20 most abundant genera (Figure 5A).

Figure 8.

Taxonomies associated with differentially expressed microbial genes in the distal jejunum of 21 d-old Ross 308 broilers fed diets ± 0.03% anti-IL-10. Briefly, Ross 308 chicks were challenged with sterile saline or 15,000 sporulated oocysts Eimeria maxima M6 on d 14 and half the E. maxima-challenged chicks were inoculated with 1 × 10⁸ colony forming units of Clostridium perfringens on d 18 and 19. Distal jejunum samples were collected from 6 chicks/ treatment at 7 d postinoculation (pi) with E. maxima and 3 dpi with secondary C. perfringens (study d 21) for RNA sequencing to evaluate the microbial metatranscriptome. Each panel represents the number of differentially expressed genes attributed to each taxonomic classification when feeding anti-IL-10 instead of the control under each challenge conditions: (A) unchallenged, (B) E. maxima only, or (C) E. maxima + C. perfringens. Totals under each pie chart represent the total number of differentially expressed genes identified within each condition.

The high prevalence of Lactobacillaceae transcripts observed herein corresponds with their observed relative abundance, whereas the percentage of those associated with Burkholderiales/Comamondacaeae is greater than their presence in the distal jejunum microbiota of unchallenged broilers. Functionally, approximately 72% of differentially expressed genes due to feeding anti-IL-10 were associated with housekeeping and unknown function, 26% with metabolic pathways, and 2% with defense mechanisms. When anti-IL-10 increased bacterial gene expression in the jejunal microbiota of unchallenged chicks, these genes were identified as belonging to Lactobacillaceae, whereas anti-IL-10 reduced gene expression in other taxonomies (Table 3). This indicates that feeding anti-IL-10 to unchallenged chicks does not detrimentally affect the metabolic activity of Lactobacillaceae as dominant members of the jejunal microbiota determined by 16S rRNA gene amplicon sequencing within the same samples (Figure 5). Evaluation of different Lactobacillus probiotics in mice and chickens has demonstrated that IL-10-stimulating functions vary between species and strains (Shida, et al., 2006; Brisbin, et al., 2010; Brisbin, et al., 2011; Takano, et al., 2020). This indicates that either the dominant Lactobacillus species/strains in the broiler jejunum do not establish in the host via IL-10 mediated pathways or the current inclusion of anti-IL-10 does not significantly impair these communities in unchallenged broilers.

In samples collected from chicks at 7 dpi with E. maxima, 60% of the microbial genes affected by dietary anti-IL-10 were associated with life-sustaining or unknown functions, 34% with metabolism, and 6% with secondary metabolites or defense mechanisms. Identified bacterial taxonomic groups included Methylobacteriaceae (53.2%), Propionibacterales (12.8%), Bradyrhizobaceae (10.6%), and Lactobacillaceae (6.4%), with Sarcocystidae (10.6%) being the family associated with Eimeria spp. (Table 4; Figure 8). Within the same samples, OTUs associated with either the Methylobacteriaceae or Bradyrhizobiaceae families were not specifically identified using 16S rRNA gene amplicon sequencing, but OTUs associated with their shared order, Rhizobiales (Erlacher, et al., 2015), were present in the dataset with undetectable relative abundances at peak challenge. Propionibacterales-associated OTUs were not identified within the 100 most abundant OTUs in the E. maxima-challenged broiler chicken jejunum at 7 dpi (data not shown), indicating that either low-abundance OTUs contributed to differential gene expression or these findings are an artifact of low sequencing depth.

Like findings in unchallenged chicks, anti-IL-10 increased the expression of genes associated with Lactobacillaceae; however, fewer differentially expressed genes were associated with Lactobacillaceae in the jejunal microbiota of unchallenged vs. E. maxima-challenged chicks (32 vs. 3; Figures 8A and 8B). While the comparisons herein emphasized anti-IL-10 effects within each challenge condition, this finding aligns with observations that Eimeria challenge alters intestinal environments at the expense of dominant genera (Jebessa, et al., 2022; Campos, et al., 2023; Pietruska, et al., 2023). Additionally, anti-IL-10 significantly downregulated the expression of unnamed but annotated cytoskeleton and transporter genes attributed to Sarcocystidae (Table 4). Both cytoskeleton and transporter proteins are important elements of motility in Apicomplexan parasites like Eimeria, making this an intriguing finding for potential anti-IL-10 effects; however, such findings may be better elucidated using in vitro techniques because of the low sequencing depth in this study (Bullen, et al., 2009). Anti-IL-10 may preserve some Lactobacillaceae function in the jejunal microbiota of E. maxima-challenged chickens and affect the parasite itself; however, the gene functions attributed to these taxa were characterized as general housekeeping functions (e.g., translation, ribosomal structure, and biogenesis) and further research is required to better elucidate this finding.

During secondary challenge with C. perfringens, none of the differentially expressed genes were attributed to Lactobacillaceae and were predominantly associated with, Methylobacteriaceae (37.9%), Sphingomonadales (36.2%), Comamonadaceae (13.8%), and Bradyrhizobiaceae (12%; 7C). 16S rRNA amplicon sequencing analysis identified only 1 detectable Rhizobiales-associated OTU, the order containing Methylobacteriaceae and Bradyrhizobiaceae, in the distal jejunum of one broiler challenged with E. maxima + C. perfringens in the control-fed group. Similarly, OTUs among the 100 most abundant identified as Sphingomonadales or Comamonadaceae were not detectable in the distal jejunum of broilers challenged with E. maxima + C. perfringens at the sampled timepoint (3 dpi with secondary C. perfringens; data not shown). While functional members of these taxa captured in the metatranscriptome analysis may be present at lower relative abundances, these findings highlight the potential artifacts introduced by low sequencing depth in the current dataset. Of the microbial genes affected by anti-IL-10 at 3 dpi with secondary C. perfringens, approximately 29% were associated with housekeeping functions, 66% with metabolism, and 5% with defense mechanisms (Table 5). When considered alongside the other challenge condition comparisons, these findings demonstrate that challenge with E. maxima + C. perfringens may be associated with reduced Lactobacillaceae functionality and greater metabolic function in other bacterial genera within the distal jejunum; however, the biological relevance of these findings in the context of dietary anti-IL-10 is unclear at this time.

While some of the findings are intriguing in the context of dietary anti-IL-10 and microbial community function, it is important to consider limitations within the the current work. This research represents investigation of the microbiota within a larger study designed to elucidate anti-IL-10 mechanisms of action within the immune system and investigate performance responses (Fries-Craft and Bobeck, 2024a; Fries-Craft and Bobeck, 2024b; Fries-Craft, et al., 2024). This required balancing adequate numbers for intensive sample collections with experimental goals and pior experience, including retaining acceptable experimental power and retaining adequate numbers for performance evaluation. In turn, 6 chicks/ treatment/ timepoint for metatranscriptome analysis allowed balance with other research considerations within the experimental series. The more significant limitation was the low sequencing depth achieved for evaluating the metatranscriptome of microbial communities associated with the jejunal mucosa. The sequencing strategies implemented were based on methods for evaluating the rumen wall metatranscriptome that have yet to be implemented in poultry (Mann, et al., 2018). Both the rumen wall and chicken jejunal mucosa contain structures that increase surface area, ruminal papillae and intestinal villi, simultaneously increasing absorptive capacity and sites for microbial interaction with the mucosa. While both tissues contain similar structures, the chicken jejunum functions primarily as a site for nutrient absorption whereas the rumen is the site of bacterial fermentation within the gastrointestinal tract of ruminants. Scanning electron microscopy has demonstrated that rumen wall papillae in sheep and cattle are densely covered by morphologically diverse bacterial communities such that underlying tissue structures are obscured (Bauchop, et al., 1975; McCowan, et al., 1978). Similar approaches to visually assess avian intestinal structures demonstrate minimal biofilm associated with the intestinal mucosa (Esmail, 1988; Chichlowski, et al., 2007; Kadhim, et al., 2010). Outcomes in this study suggest that metatranscriptome sequencing approaches are potentially insufficient for non-ruminant livestock due to lower intestinal microbial density and diversity compared to the rumen wall, meaning that fewer bacterial reads can be identified within samples that already contain a high percentage of host reads.

Library preparation for RNAseq in this study did not deplete host-associated reads in order to retain a dataset providing potential host-microbiota insights, which resulted in microbial reads being outnumbered by host-associated sequences. Future research to evaluate the metatranscriptome of chicken intestine mucosal communities with retention of host reads could focus on other sections, namely the ceca which hosts highly diverse luminal and mucosal microbial communities relative to other gastrointestinal compartments (Borda-Molina, et al., 2016). This may potentially overcome limitations associated with low sequencing depth for microbial communities closely associated with host tissues; however, this intestinal location is distal from E. maxima and C. perfringens infection sites and may be of limited relevance for necrotic enteritis outcomes (Williams, 2005; Gong, et al., 2007). Another alternative is to increase the number of sequence reads to overcome the low concentration of microbial sequences relative to host presence; however, such approaches are significantly limited by the cost and computational power necessary (Shakya, et al., 2019).

It is also important to note that sequencing approaches in this study generated a substantial dataset containing over 18 billion reads, suggesting a high sequencing threshold needed to obtain biologically relevant data. While technological advancements may overcome these issues over time, current evaluation of microbial community function in the chicken jejunum may be better accomplished by detecting the known end-products of microbial activity such as short-chain fatty acids. Such metabolites have documented indirect effects on host immunity, but their widespread effect on multiple immune and non-immune cell types means that presence alone may be insufficient to describe the dynamics of host-microbiota interactions (Kim, 2018). Thus, the work herein emphasizes the complex limitations of microbial community -omics research, particularly in non-ruminant model species. While significant limitations were identified for evaluating the metatranscriptome of microbial communities in the chicken jejunum, the dataset generated by this work included host reads and further analysis could identify the potential mechanisms of anti-IL-10 relating to host responses.

DISCLOSURES

The authors declare no conflict of interest.