Investigation of the associations between the gastrointestinal microbiome and animal host performance and health

The projected human population growth drives a worldwide increased pressure to improve animal production systems from both the economic and environmental perspectives. Bovine ruminants are one of the most interesting sources of high-quality protein, due to their ability to convert human-undigestible plant biomass and utilize it in the production of meat and milk. This ability is due to symbiotic associations with their rumen microbiome, i.e., the collection of microorganisms (bacteria, protozoa, fungi, and archaea) inhabiting the rumen, and their microbial genes (i.e., the metagenome). Whereas mammals do not produce the necessary enzymes to breakdown and digest complex polysaccharides in fibrous plants, bacteria, protozoa, and fungi are able to ferment these into volatile fatty acids, microbial proteins, and vitamins, which are utilized by the ruminant host for maintenance, growth, and development. An excess of hydrogen produced during the fermentation process is utilized by archaea, leading to the production of methane, a greenhouse gas with 28 times higher warming potential than carbon dioxide, which is released mostly by eructation into the environment, and thus contributing to climate change. This thesis focuses on the investigation of the host animal-microbiome symbiosis and how it impacts host animal performance traits, including appetite, growth, and feed conversion efficiency. We also analysed the symbiotic association with the host’s health, by considering the impact of the presence of a parasitic nematode on the gastrointestinal microbiome of cattle. In the first chapter of this thesis, I present an overview of the state-of-the-art knowledge, addressing why these microbiome-focused studies are crucial for the future development of more efficient bovines with lower environmental impact, summarizing the main findings in the field so far, and underlining the current challenges. The second chapter of this thesis includes a comparison between taxonomic compositions obtained from processing 16S rRNA amplicon sequences derived from caecum, colon, and faecal pig samples, using two different bioinformatics pipelines, the MetaGenome Rapid Annotation using Subsystem Technology (MG-RAST) and the Quantitative Insights Into Microbial Ecology 2 (QIIME2). The results suggested that the microbiota profiles differ significantly according to the bioinformatics pipeline applied, with consequences on the subsequent statistical analyses; for example, at family level the richness and evenness of the samples was higher when samples were processed in QIIME2. We also compared different data cleaning and filtering methods (e.g., application of a minimum relative abundance threshold of 1%), which led to inconsistent results, depending on the pipeline used to identify microbial taxa. When using the whole datasets in partial least squares discriminant analyses (PLS-DA) to discriminate between sample collection sites, MG-RAST-derived data led to more accurate results, whereas QIIME2 was more accurate when a minimum relative abundance threshold was applied. This study was published in the Journal of Microbiological Methods in May of 2021. This first study provides a substantial insight into the challenges of taxonomic characterization of samples using 16S rRNA sequencing data, which could be at least partly reduced by using whole metagenome sequencing data. The high resolution of whole metagenome sequencing methods provides the opportunity to identify functional microbial gene orthologs (e.g., using the Kyoto Encyclopedia of Genes and Genomes, KEGG), allowing for the comprehensive understanding of the biological and biochemical networks involved in the digestive processes closely associated with the host animal performance and health. Therefore, in the third chapter, we investigated rumen metagenome data derived from whole metagenome sequencing of rumen samples taken from beef cattle at slaughter, and their association with host performance traits including feed conversion ratio, average daily gain, residual feed intake and daily feed intake (FCR, ADG, RFI and DFI, respectively). Our analyses based on partial least squares (PLS) models identified sets of 20, 14, 17, and 18 microbial genes whose relative abundances explained 63, 65, 66, and 73% of the variation of FCR, ADG, RFI, and DFI, respectively. This research was the first to investigate the association between rumen microbial genes and beef cattle performance, and it provides very detailed information about the functional role of the microbiome on cattle digestive processes. For example, microbial genes associated with cellulose and hemicellulose degradation, vitamin B12 synthesis, and amino acid metabolism were identified as biomarkers for enhanced feed conversion efficiency, whereas biomarkers for inefficient feed conversion were associated with functions such as pathogen lipopolysaccharide synthesis, cationic antimicrobial peptide resistance, and degradation of toxic compounds. This study was published in the journal Frontiers in Genetics in August of 2019. Most studies focused on associations between the microbiome and the bovine host are based on taxonomic and metagenomic compositions derived from samples taken shortly after slaughter, but whether these samples are representative of the microbiome found in the rumen at different stages of the host animal’s life is still unclear. In chapter 4, a study on the longitudinal stability of the rumen microbiome of beef cattle throughout the finishing growth phase is presented. Samples were collected from 20 animals before a nitrate-or oil-based additive was included in their basal diets (pre-additive), at the start, mid and end of a 56-day testing period (during which the performance traits FCR, ADG, DFI, and RFI were measured), after the animals left the respiration chamber in which they were tested individually for methane production, and at slaughter. Our results highlight that the microbiome compositions are stable throughout the finishing growth period, both at the microbial genera and microbial genes levels. We used partial least squares models to predict each performance measure (FCR, ADG, DFI, and RFI), and methane production (CH4 production in g/day and CH4 yield in g/kg dry matter intake) trait based on the microbiota and metagenomic datasets derived from each timepoint and compared the variable importance in projection (VIP) scores and the regression coefficients extracted from each PLS model. The results showed substantial consistency throughout the timepoints, indicating that data collected from any timepoint during the finishing growth phase would be suitable for prediction of host traits. Additionally, we found that rumen microbial biomarkers previously identified in other studies based on samples taken at slaughter are also informative to predict host performance traits based on rumen samples taken at earlier stages of the animal’s life, underlining not only the stability of the microbiome but also our confidence in the results from previous studies. The gastrointestinal microbiome is not only highly associated with host performance traits, as shown in chapters 3 and 4, but also with host health traits. We investigated the influence of the presence of the abomasal parasitic nematode Ostertagia ostertagi, and of a vaccine against the nematode, on the rumen and caecum microbiomes of dairy cattle, and these studies are presented in chapters 5 and 6, respectively. A total of 24 calves were included in the experiment, of which 4 were left unvaccinated and uninfected throughout the whole experiment (UNF), 10 received a native vaccine against O. ostertagi, and 10 were injected with adjuvant only (positive control). The vaccinated and positive control groups were then subjected to the infection challenge that consisted of oral administration of 1000 infectious L3 larvae per day for 25 days. The animals were evaluated based on their cumulative faecal egg count (cFEC), leading to the identification of 4 vaccinated animals (with average cFEC, VAC) and 8 animals from the positive control group (4 extremely low and 4 extremely high cFEC, CLE and CHE, respectively) for whole metagenome sequencing and microbiome profiling at both taxonomic and microbial genes levels. The results indicate that the parasitism by this abomasal nematode substantially impacted the rumen and caecum microbiomes, and that the impact on the microbiome differs according to the parasitism severity. For example, in comparison to UNF, the rumen of CLE animals was depleted of microbial genes associated with valine, leucine, and isoleucine metabolism, whereas the rumen of CHE animals was depleted of microbial genes associated with bacterial chemotaxis and flagellar assembly. Additionally, in comparison to UNF, the rumen of VAC animals was enriched in microbial genes associated to metabolism of vitamin C and vitamin B12 synthesis. In the caecum of infected animals (CLE and CHE) we observed enrichment of several genera belonging to the phylum Actinomycetia, e.g., Cellulomonas, and depletion of several genera in phyla Gammaproteobacteria and Bacilli with pathogenic potential, such as Legionella and Melissococcus. At the functional level, the analyses revealed enrichment of microbial genes associated with uptake of urea, and degradation of aromatic compounds in the caecum of infected, in comparison to uninfected animals. The rumen and caecum of vaccinated animals was enriched in several fungi, most of which with potential pathogenicity, e.g., Colletotrichum and Botrytis in the rumen, and Pneumocystis and Malassezia in the caecum, in comparison to uninfected animals. The final chapter of the thesis includes a summary of the results of all previous chapters, and a general discussion on the symbiotic relationships between the gastrointestinal microbiome and host animal performance and health traits.