Dynamic modelling of the core microbiota in the porcine gut

Application Probiotic and prebiotic manipulation of the porcine gut core microbiota could have impacts on microbial short chain fatty acids (SCFAs) fermentation, affecting positively host health and performances.

Introduction Porcine core microbiota (CM) is a bacterial community present in the porcine gut microbiota, suggested to be independent of diet or breed (Holman et al., 2017). It could therefore represent an optimal and standardised probiotic target allowing modifications of microbiota-derived SCFA patterns, with benefits for host health and performance. Considering both in vitro and in vivo limitations, we developed a mathematical CM model to uncover, in silico, its possible role through comparison with the whole microbiota SCFA pattern found in literature, and to explore, in silico, the possibility of using CM genera and/or their substrates in the context of probiotic/prebiotic intervention.

Material and methods The porcine CM identified by a meta-analysis study (Holman et al., 2017), is composed of up to 20 genera, depending on the CM gut location. To develop our CM mathematical model, we modelled each genus as a bacterial unit (BU), summarising their species and pathways complexity, and modelling each BU as if it were a singular microorganism. The BU models were first validated comparing their output with experimental data from the study on which the BU stoichiometry is based. Where no cultural information was found, the stoichiometry assumption was based either on the theoretical description of the genus or on the phylogenetically closest genus with known data. Therefore, it is assumed that merging the BU models in a CM model, the latter output might be as similar as possible to a culture composed of the same genera. The mathematical formulation of the model consists of a system of ordinary differential equations (ODEs) based on a previous model (Kettle et al., 2015), characterising different kinds of substrates related to the different type of bacterial growth. The environment modelled is an in vitro fermenter, either batch or continuous culture, since we did not include any biotic factors in the system of ODEs. The model was solved using the microPop package in R (Kettle et al., 2017). The major SCFA and lactate concentrations from the model are compared with three different studies, which used faecal, caecal or colonic inoculum in either batch or continuous culture (Lin et al., 2011; Tanner et al., 2014; Ding et al., 2015). In each case, the model simulated the conditions described in the study, through the modification of the starting concentration value of inoculum and substrates. The model was also used to simulate the effect of probiotic and prebiotic approaches targeting each BU and their substrates. In this case, probiotic/prebiotic simulations were compared with the output of a model simulating the non-enhanced CM under the same experimental conditions.

Results The modelled SCFA pattern shows a great similarity to experimental data not used for developing the model. This indicates that the CM is largely responsible for the SCFA pattern produced by the whole microbiota. The small differences in acetate and butyrate production may indicate that CM provides acetate to the non-CM bacteria (Figure 1). The probiotic simulation showed that not all the BU enhancements had an appreciable effect on the SCFA pattern, whereas the prebiotic simulations showed, an increasing SCFA concentration, with all the substrates used as a prebiotic (data not shown). Thus, we simulated a therapy combining Faecalibacterium BU, Treponema BU and resistant starch, in a continuous culture.

Conclusion Although a validation through in vitro/in vivoexperimentation is necessary, our modelling data showed that the CM forms the basis of microbial SCFA production despite the limited number of CM genera, and it could provide acetate, used by the non-CM genera (e.g. producing butyrate). Furthermore, the model shows that a possible CM genera/substrate enhancing therapy could increase the total SCFA concentration with a five-fold increase in butyrate concentration.