Faecalibacterium Gut Colonization Is Accelerated by Presence of Older Siblings

Faecalibacterium prausnitzii has been suggested to constitute a key marker of a healthy gut, yet the factors shaping the colonization of this highly oxygen-sensitive, non-spore-forming species in the intestinal environment remain poorly understood. Here, we provide evidence from three separate infant study populations that F. prausnitzii colonization in the gut happens during late infancy and is affected by the number of older siblings in the family. We conclude that Faecalibacterium acquisition is highly likely to be accelerated by contact between siblings. Bearing in mind the immunoregulatory properties of F. prausnitzii and the well-established protective effects against allergic disorders related to the presence of older siblings, early colonization of this species may have profound consequences for child health.

IMPORTANCE Faecalibacterium prausnitzii has been suggested to constitute a key marker of a healthy gut, yet the factors shaping the colonization of this highly oxygensensitive, non-spore-forming species in the intestinal environment remain poorly understood. Here, we provide evidence from three separate infant study populations that F. prausnitzii colonization in the gut happens during late infancy and is affected by the number of older siblings in the family. We conclude that Faecalibacterium acquisition is highly likely to be accelerated by contact between siblings. Bearing in mind the immunoregulatory properties of F. prausnitzii and the well-established protective effects against allergic disorders related to the presence of older siblings, early colonization of this species may have profound consequences for child health.
KEYWORDS Faecalibacterium, infancy, siblings S tarting around birth, the infant gut is colonized by a bacterial community, which gradually increases in richness and diversity (1,2). A number of factors, including mode of delivery, breastfeeding, and introduction of complementary feeding affect this colonization. After approximately 3 years, the gut microbiota of the child is comparable to that of an adult (3).
Faecalibacterium prausnitzii, which is the sole species within the genus Faecalibacterium, is one of the most prevalent and abundant human gut microbes (4), and it has been suggested that it constitutes a marker of gut health due to its lower abundance in patients suffering from inflammatory bowel diseases compared to healthy controls and to its immunoregulatory properties (5,6). Currently, very little is known about factors that impact the colonization of F. prausnitzii during early life and the putative contribution of this species to immune system development (7). Animal studies have shown that prior colonization with other microbes is necessary to obtain robust establishment of F. prausnitzii in germfree rodents (8,9), and it has been suggested that depletion of oxygen in the gut or putatively another type of conditioning of the gut environment by other microbes is required for F. prausnitzii to colonize the gut (7). The sensitivity of F. prausnitzii to low pH and bile salts explains the observation that it predominantly colonizes the distal gastrointestinal tract of humans (6), but it remains to be established how this highly oxygen-sensitive, non-spore-forming bacterium is transmitted between individuals and which factors affect its colonization process. We therefore analyzed 16S rRNA amplicon sequencing data from fecal samples originating from three previous studies of Danish infants entitled SKOT I (10), SKOT II (11), and ProbiComp (12), respectively. Comparison of Faecalibacterium abundance to a number of selected variables allowed identification of factors influencing colonization of Faecalibacterium.
F. prausnitzii is not detected or found at very low levels during the first 4 to 6 months of life, but it increases rapidly during the first years (13)(14)(15). By investigation of three independent study populations, SKOT I (n ϭ 114), SKOT II (n ϭ 113), and ProbiComp (n ϭ 255), we found that the relative abundance and prevalence of Faecalibacterium increased significantly during the period from late infancy (8 to 12 months old) to toddlerhood (14 to 19 months old) ( Fig. 1A to F). In contrast to the SKOT infants, which were sampled as close as possible to 9 and 18 months of age, the infants in the ProbiComp study were sampled over a larger age span (Fig. 1G). This enabled correlation analysis between age and Faecalibacterium abundance and revealed a significant positive correlation at late infancy, but no such correlation at toddlerhood (Fig. 1H), suggesting that Faecalibacterium colonization reaches a plateau (Fig. 1I). To investigate how external factors and host variables may affect the relative abundance of Faecalibacterium during late infancy, we analyzed the association of its abundance with (i) gender, (ii) gestational age at birth, (iii) Caesarean section, (iv) breastfeeding, (v) pets, (vi) older siblings, and (vii) anthropometrics in all three infant study populations. Among these factors, only the presence of older siblings in the home was consistently correlated with the relative abundance of Faecalibacterium across the three study populations ( Fig. 2A). Both the relative abundance and the prevalence of Faecalibacterium were higher during late infancy in infants with older siblings than in infants with no older siblings in the three populations (see Fig. S1 in the supplemental material). These differences were no longer present in toddlerhood, and Faecalibacterium was detected in almost all individuals (Fig. S1). Combination of the three data sets (n ϭ 482) and stratification of individuals with no (n ϭ 270), one (n ϭ 156), or two or more (n ϭ 56) older siblings revealed a stepwise increase in relative abundance (P Ͻ 0.0001 by the Kruskal-Wallis test) and prevalence (P for the trend [P fortrend ] of 0.0002 by the chi-square test) of Faecalibacterium with increasing number of older siblings during late infancy (Fig. 2B). Moreover, age-adjusted partial correlation analysis between relative abundance of Faecalibacterium during infancy and the number of older siblings revealed that this association was highly significant (rho ϭ 0.23; P ϭ 6.4 ϫ 10 Ϫ7 ; n ϭ 482). The association was no longer present when the children became toddlers, suggesting that at this age, children will have acquired Faecalibacterium from other encounters. As infants with older siblings are putatively introduced earlier to solid foods, we investigated diet as a potential confounder, since we have previously found that diet influences the microbiota (11). By analysis of detailed dietary records obtained for the SKOT cohorts (11,16), we found that transition to family foods (rho ϭ 0.16; P ϭ 0.018; n ϭ 217) and dietary intake of rye bread (rho ϭ 0.20; P ϭ 0.003; P value for the corrected false-discovery rate [P FDRcorrected ] of 0.07; n ϭ 217), but not other dietary parameters, were associated with Faecalibacterium (Fig. S2). However, the association between relative abundance of Faecalibacterium and the number of older siblings persisted after adjustment for both transition to family foods (rho ϭ 0.26; P ϭ 0.0001; n ϭ 217) and intake of rye bread (rho ϭ 0.25; P ϭ 0.0001; n ϭ 217). Together, this shows that rather than gender, mode of delivery, breastfeeding, complementary diet, exposure to pets or anthropometric measures, it is the exposure to older siblings that predominantly affects Faecalibacterium acquisition during late infancy.
Study populations and samplings. Data from three infant study populations were included in the present study. The study protocol for ProbiComp was approved by the Committees on Biomedical Research Ethics for the Capital Region of Denmark (H-4-2014-032), and the study was registered at clinicaltrials.org (NCT02180581). The study SKOT II S1 SKOT I S1

protocols for the SKOT cohorts were approved by the Committees on Biomedical Research Ethics for the Capital Region of Denmark (H-KF-2007-0003 and H-3-2010-122).
All parents signed a consent form. The ProbiComp study (n ϭ 290) was a double-blind randomized placebo-controlled study where infants, aged 8 to 13 months and starting daycare within 12 weeks after the start of the study were randomly assigned to receive a combination of the two probiotics, BB-12 (Bifidobacterium animalis subsp. lactis BB-12 strain) and LGG (Lactobacillus rhamnosus GG), or placebo during the 6-month study period (17). Inclusion criteria were single birth and starting daycare at 8 to 14 months of age. Exclusion criteria were low birth weight (Ͻ2,500 g), gestational age at birth of Ͻ36 weeks, chronic illness, regular medication, antibiotics within 4 weeks prior to the start of the study, and non-Danish speaking parents. Fecal samples were obtained before (n ϭ 265; age 8 to 13 months) and after (n ϭ 210; age 14 to 19 months) the study and were stored at Ϫ80°C until DNA extraction and microbiota profiling as described previously (12). The fecal microbiotas were profiled in 255 and 201 samples before and after the study, respectively, and the treatment had no impact on the fecal microbiota (12). The SKOT cohort studies monitored children during the first 3 years of life with the overall aim to investigate relationships between early diet, growth development, and later disease risks. In SKOT I (18), infants from a random sample of mothers were recruited (n ϭ 311), whereas in SKOT II (19), infants of obese mothers (body mass index [BMI] of Ͼ30 kg/m 2 ) were recruited (n ϭ 184). Inclusion criteria for both cohorts were single birth, full-term delivery, age of 9 months Ϯ 2 weeks at the first visit, and absence of chronic illness. Participants in both cohorts were examined at 9 (Ϯ2 weeks), 18 (Ϯ4 weeks), and 36 (Ϯ12 weeks) months of age, and fecal samples were collected and stored at Ϫ80°C until DNA extraction. The fecal microbiota data from the first two sampling points in subsets of 114 (SKOT I) and 113 (SKOT II) infants have been published previously (11). In all study populations, fecal samples were freshly delivered on the morning of the visit or had been stored in the participant's home, either in the freezer (Ϫ18°C) or in the fridge (4°C) for maximally 24 h before storage at Ϫ80°C. Information on gender, gestational age, Caesarean section, breastfeeding (months of breastfeeding only and the frequency of breastfeeding at sampling times), presence of older siblings and furred pets was collected from parental interviews at the sampling times in all three study populations. Anthropometrics were obtained at examinations and Z-scores were calculated as described previously (11,17). Food records were obtained and analyzed in the SKOT studies as described previously (11,19).
DNA extraction and microbiota profiling. The procedures for DNA extraction, PCR amplification, and sequencing of the V3 region of the 16S rRNA gene have been described previously (11,12) , and samples were pooled to obtain equimolar libraries, which were sequenced using the Ion OneTouch and Ion PGM (personal genome machine) platform with an Ion 318 Chip kit. Sequence data were analyzed as described previously for SKOT data (11) and ProbiComp data (12), and the levels of all sequences classified as Faecalibacterium by the RDP classifier, with a confidence threshold of 0.5 (20) against the Greengenes database v. 13.8 (21) were pooled within each sample and normalized to the total reads in that sample.
Statistics. Mann-Whitney, Kruskal-Wallis, Spearman's rank correlations, Fisher exact test, and chi-square tests were performed with the GraphPad Prims software (v. 7.0; GraphPad Software, Inc., La Jolla, CA). When indicated, P values were corrected for multiple testing using the false-discovery rate (FDR) (22). Partial Spearman's rank correlation analysis adjusted for age or dietary parameters was performed in R (version 3.1.0; R Core Team. 2014. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria) using the ppcor package.
Accession number(s). Sequencing data are deposited at NCBI Sequence Read Archive with the accession number SRP100762 under BioProject PRJNA360073 for ProbiComp study and accession number SRP052851 under BioProject PRJNA273694 for SKOT cohorts.

ACKNOWLEDGMENTS
We thank the infants and their families participating in the SKOT and ProbiComp studies for their contributions. We thank Vivian Anker, Birgitte Hermansen, Julie Brønholm, and Inge Rasmussen for technical assistance.
The ProbiComp study was funded by Innovation Fund Denmark (grant 0603-00579b). SKOT studies were supported by the Danish Directorate for Food, Fisheries and Agribusiness (grant 3304-FSE-06-0503). The microbiota analysis was supported by the Gut, Grain and Greens (3G) Center, financed by the Danish Council for Strategic Research (grant 11-116163/0603-00487B).
We declare that we have no conflict of interest.