Functional maturation of the gut microbiota at weaning is influenced by maternal environment in piglets

The objective of this study was to analyze in piglets the impact of weaning on the production of metabolites by gut bacteria and to determine whether early life environment influences the functional maturation of the gut microbiota. Fecal metabolome and microbiome were analyzed in piglets raised in two separate maternity farms and mixed at weaning. In piglets from both maternity farms, the relative abundance of Lactobacillus and of the predicted function “Fucose degradation” decreased after weaning while the relative abundance of Ruminococcus 2 and of the predicted function “Starch degradation” increased. In piglets from the first maternity farm, the relative concentration of biogenic amines and the relative abundance of Escherichi-Shigella decreased after weaning while the relative concentration of short chain fatty acids and the relative abundance of Christensenellaceae R-7 group and Ruminococcaceae UCG-002 increased. These changes were not observed at weaning in piglets from the second maternity farm probably because they already had high relative concentration of short chain fatty acids and higher relative abundance of Christensenellaceae R-7 group and Ruminococcaceae UCG-002 during the suckling period. In conclusion, the functional maturation of the microbiota at weaning is highly dependent on the maternal environment in piglets. ORIGINALITY - SIGNIFICANCE STATEMENT Bacterial metabolites are key molecular intermediates between the gut microbiota and host cells. Our study in piglets reveals that the metabolic activity of the gut microbiota shifts at weaning, a key developmental period for intestinal and immune health. We also show that this functional maturation of the gut microbiota is strongly influenced by maternal environment. Thus, targeting early life environmental factors is a promising strategy to program health trough the production of beneficial bacterial metabolites at the suckling-to-weaning transition.


SUMMARY 23
The objective of this study was to analyze in piglets the impact of weaning on the production of 24 metabolites by gut bacteria and to determine whether early life environment influences the functional 25 maturation of the gut microbiota. Fecal metabolome and microbiome were analyzed in piglets raised 26 in two separate maternity farms and mixed at weaning. In piglets from both maternity farms, the 27 relative abundance of Lactobacillus and of the predicted function "Fucose degradation" decreased 28 after weaning while the relative abundance of Ruminococcus 2 and of the predicted function "Starch 29 degradation" increased. In piglets from the first maternity farm, the relative concentration of biogenic 30 amines and the relative abundance of Escherichi-Shigella decreased after weaning while the relative 31 concentration of short chain fatty acids and the relative abundance of Christensenellaceae R-7 group 32 and Ruminococcaceae UCG-002 increased. These changes were not observed at weaning in piglets 33

INTRODUCTION 41
The gut microbiota is a major regulator of animal physiology and health. After initial colonization at 42 birth by microorganisms originating both from the mother and the environment, maternal milk shapes 43 the gut microbiota composition through nutrients (e.g. lactose, milk oligosaccharides), 44 immunoglobulins and antimicrobial compounds (e.g. lactoferrin, lysozyme) (Macpherson et al., 2017). identified yet, emerging evidences suggest that bacterial metabolites may play a key role as 54 intermediates between the microbiota and its host at the suckling-to-weaning transition (Al Nabhani 55 and Eberl, 2020). Therefore, identifying the metabolites produced by gut bacteria in young mammals 56 across this dietary shift might be useful for the development of health-promoting strategies based on 57 the control of the metabolic activity of the microbiota in early life. 58 Piglets raised in commercial conditions represent an attractive animal model to study the effects of 59 weaning on the microbiota since the separation from the sow occurs 3 to 4 weeks after birth (versus 8 60 to 14 weeks in natural conditions) which results in an abrupt cessation of suckling when piglets milk 61 intake is still high while solid feed intake very low (Newberry and Wood-Gush, 1985). Moreover, the 62 results obtained in this model are relevant both for the pig industry (e.g. management of post-weaning 63 diarrhea in piglets) and for human health due to the anatomical, functional and microbial similarities 64 between piglet and human infant gastrointestinal tract (Heinritz et al., 2013;Gresse et al., 2017).

Maternal environment and weaning influence the predicted functionality of the gut microbiota 136
We predicted the functional capacity of gut bacteria by using inference based on 16S rDNA amplicons 137 sequences (PICRUSt2). Heatmap representation of the relative abundance of all predicted pathways 138 suggested that the functional potential of the microbiota was influenced mainly by age and to a lower 139 extent by maternal environment (figure 6A). Indeed, there was a significant effect of weaning for 62% 140 of the predicted pathways (199/320) and a significant effect of maternal environment for only 5% of 141 them (15/320) (table S8). We focused on two key microbial pathways involved in the degradation of 142 dietary substrates available for the gut microbiota: fucose (a monosaccharide contained in milk 143 oligosaccharides) and starch (a major plant carbohydrate). 144 The relative abundance of the predicted pathway "Fucose degradation" decreased after weaning in 145 piglets from both maternity farms (figure 6B). The 5 OTUs contributing the most to this predicted 146 function were affiliated to Enterobacteriaceae, Lachnospiraceae and Ruminococcaceae (figure 6B and 147 table S9). At day 13, the relative contribution of OTU8 (Ruminococcaceae) to this function was more 148 important in piglets from maternity 2 than in piglets from maternity 1. In piglets from maternity 1, the 149 relative contribution of OTU1 (Enterobacteriaceae) to the fucose degradation pathway decreased after 150 weaning while the opposite was observed for the OTU33 (Lachnospiraceae). In piglets from maternity 151 2, the relative contribution of the 5 OTUs contributing the most to this pathway did not change after 152 weaning. At day 23, the relative contribution of OTU8 (Ruminococcaceae) was still more important in 153 piglets from maternity 2 compared to piglets from maternity 1 while the opposite was observed for the 154 relative contribution of OTU19 (Lachnospiraceae). 155 The relative abundance of the predicted pathway "Starch degradation V" increased after weaning in 156 piglets from both maternity farms (figure 6C). The 5 OTUs contributing the most to this function were 157 affiliated to Ruminococcaceae, Lachnospiraceae and Clostriadiaceae 1 (figure 6C and table S10). At 158 day 13, the relative contribution of OTU8 (Ruminococcaceae) to this function was more important in 159 piglets from maternity 2 compared to piglets from maternity 1. The relative contribution of OTU3 160 (Ruminococcaceae) increased after weaning in piglets from both maternity farms. At day 23, the 161 relative contribution of OTU8 (Ruminococcaceae) to this function was still more important in piglets 162 from maternity 2 compared to piglets from maternity 1 (as observed at day 13).

9
It is important to consider that the relative contribution of OTUs to these two predicted pathways was 164 strongly and positively correlated with the OTUs relative abundances (table S7, ρ>0.86 for both  165   pathways). Moreover, all the OTUs contributing the most to the predicted pathways had a high relative 166 abundance (>1.4% in at least one group, table S7). Overall, our results show that weaning and, to a 167 lower extent, maternal environment influence the predicted functionality of the gut microbiota and the 168 relative contributions of bacterial OTUs to these functions. 169

DISCUSSION 170
Our study reveals that the production of metabolites by gut bacteria shifts at weaning in piglets in an 171 early life environment-dependent manner. Since bacterial metabolites which concentration was altered 172 at weaning (e.g. SCFA and biogenic amines) are known to regulate the homeostasis of host cells (Koh 173 et al., 2016;Tofalo et al., 2019), the metabolic shift of the microbiota observed at weaning might have 174 major implication for health. Moreover, the influence of maternal environment on the functional 175 maturation of the microbiota at weaning in piglets suggests that controlling early life environment 176 might be a promising strategy to promote the production of beneficial metabolites at weaning, and thus 177 to program long term health (Al Nabhani and Eberl, 2020). 178 During the suckling period, the microbiota of piglets raised in the two separate maternity farms 179 produced different levels of several bacterial metabolites. For instance, the relative concentration of 3-180 (4-hydroxyphenyl)propionate was higher in piglets from maternity 1 during the suckling period. This 181 metabolite is produced by the microbiota through degradation of the aromatic amino acid tyrosine 182 (Oliphant and Allen-Vercoe, 2019) which concentration was similar in the feces of piglets from both 183 maternities. Thus, rather than substrate availability, differences in microbiota composition and 184 functional potential might drive the differential production of 3-(4-hydroxyphenyl)propionate. We also 185 observed a high concentration of succinate in fecal samples of piglets from maternity 1, despite this 186 metabolic intermediate usually do not accumulate since it can be rapidly converted to SCFA by gut 187 bacteria (Oliphant and Allen-Vercoe, 2019). Accordingly, the concentration of the SCFAs propionate 188 and isobutyrate were lower in piglets from maternity 1 than in piglets from maternity 2. Since 189 propionate is produced mainly through plant derived polysaccharide degradation (Louis and Flint,190 2017), our results suggest that carbohydrate fermentation might be more active in suckling piglets 191 from maternity 2. The different metabolic activity of the microbiota of piglets from the two maternity 192 Despite the effects of weaning on the production of bacterial metabolites were specific to each 202 maternity farm, some of the effects of weaning on the microbiota diversity and composition were 203 observed in piglets from the two maternity farms. There was an increase in OTU richness after 204 weaning, which is in agreement with previous studies in piglets (Frese et  The taxonomic changes observed after weaning in piglets from both maternity farms were associated 220 with a major shift in the predicted functionality of the microbiota. Indeed, the relative abundance of 221 62% of the predicted pathways was altered after weaning. Among them, the relative abundance of the 222 "Fucose degradation" pathway decreased in piglets from the two maternity farms. Fucose is a 223 monosaccharide present in milk oligosaccharides (fucosylated oligosaccharides represent 9.1% of total 224 milk oligosaccharides in sow) (Salcedo et al., 2016). Thus, the predicted decrease in the relative 225 abundance of the "Fucose degradation" pathway might be driven by suckling cessation at weaning. 226 The OTU contributing the most to this predicted pathway was classified in Enterobacteriaceae 227 (dominant family during the suckling period). This result is in agreement with a study using shotgun 228 metagenomics in suckling piglets showing that the majority of genes related to fucose utilization were 229 assigned to Enterobacteriaceae (Salcedo et al., 2016). Another interesting result was the increased 230 relative abundance of the predicted pathway "Starch degradation V" after weaning in piglets from both 231 maternity farms. Starch is an important source of carbohydrates for the microbiota of weaned piglets 232 study showing that bacterial genes coding for starch degrading-enzymes were more abundant after 234 weaning in piglets (Frese et al., 2015). We also found that after weaning the OTU contributing the 235 most to the "Starch degradation V" predicted pathway was assigned to Ruminoccocaceae. 236 Accordingly, a member of this family, Ruminococcus bromii, was demonstrated to be a key stone 237 species for starch degradation in the human microbiota (Ze et al., 2012). Overall, our results suggest 238 that suckling cessation and the shift to a plant based diet drive the maturation of the microbiota 239 composition and predicted functionality observed after weaning in piglets. 240 In contrast to the shared effects of weaning on the microbiota composition and predicted functionality 241 observed in piglets from the two maternity farms, the alteration of bacterial metabolites production 242 was specific to piglets from each maternity farm. In piglets from maternity 1, there was a reduction of 243 the relative concentrations of biogenic amines (cadaverine, tyramine and 5-aminovalerate) after 244 weaning. These metabolites are produced by the gut microbiota from amino acids degradation (lysine, Strikingly, most of the modification observed after weaning in piglets from maternity 1 were not found 276 in piglets from maternity 2. This might be linked to the more "mature state" of the microbiota 277 composition and metabolic activity observed already during the suckling period in these piglets, as 278 discussed above (e.g. higher relative abundance of Ruminococcaceae UCG-002 and 279 Christensenellaceae R-7 group and higher relative concentration of SCFA). Interestingly, some 280 differences observed during the suckling period (e.g. higher relative abundance of Ruminococcaceae 281 UCG-002) were still found after weaning, showing a persistent effect of the maternal environment. 282 Thus, our study highlights the importance of early life environment on the microbiota composition and 283 functionality during the suckling period but also after weaning. 284 In conclusion, our study shows that the functional maturation of the microbiota at weaning is 285 influenced by early life environment in piglets. Since this alteration of the microbiota functionality at 286 the suckling-to-solid food transition might play a key role in intestinal and immune maturation 287

Animals 296
The experiments were performed in two separate maternity farms (maternity 1 and maternity 2) and 297 one post-weaning farm, all located in Morbihan, France. Piglets (Piétrain x Large white x Landrace) 298 were housed in maternity 1 (n=14 piglets/10 sows) or maternity 2 (n=14 piglets/9 sows) from birth to 299 day 21 (weaning day) (figure 1). Chemical composition of gestation and lactation diets of sows from 300 the two maternity farms are shown in table S1. Suckling piglets had access to creep feed from day 1 in 301 maternity 1 and from day 7 in maternity 2 (chemical composition of piglets diets shown in table S1). 302 At weaning (day 21), piglets from the two maternity farms were moved to a single post-weaning farm 303 and mixed in the same pens in one room. After weaning, all piglets had access to the same diet (table 304 S1). None of the piglets received antibiotic treatment. Fecal samples were collected during the 305 suckling period (day 13) and two days after weaning (day 23) from the same piglets and stored at -306 80°C until analysis.   predicted pathways in each group. The color represent the Z-scores (row-scaled relative abundance) 563 from low (blue) to high values (red). Predicted pathways (rows) were clustered by the average method. 564 B -"Fucose degradation" predicted pathway. C -"Starch degradation V" predicted pathway. Left 565 panels: relative abundance of the predicted pathway (mean + SEM). Right panels: relative contribution 566 of the 5 OTUs contributing the most to the predicted pathway (means). After log transformation of 567 bacterial predicted pathway relative abundances or OTU relative contribution to the pathway, a linear 568 mixed model was used with age and maternity as fixed effects and sows and piglets as random effects. 569 The means of groups associated with different letters are significantly different. 570