In this study, we evaluated gut microbiota evolution during the first month of life in CS-delivered and both HB and hospital-based VAG infants. Our results highlight the relevance of perinatal factors to the gut microbial colonization pattern, which affects the innate defensive mechanisms and functional maturation at the intestinal level. We observed how distinct microbial colonization patterns were associated with the mode of delivery and how HB altered the host gut barrier and innate immune response in an in vitro model.
CS procedures are associated with specific conditions, including the use of antibiotics, longer hospitalization, low neonatal–maternal contact and the delayed initiation of breastfeeding [35, 36], which may influence microbial colonization patterns [26]. A higher percentage of bottle or mixed feeding is commonly observed in CS-born infants [35], including the studied cohort. In our cohort, the HB infants showed a higher rate of exclusive breastfeeding than the hospital-born infants (95% in HB verses 67% and 61% in CS and VAG at hospital, respectively). Furthermore, the practices associated with hospital delivery, as antibiotic use, vaginal cleansing, controlled maternal food intake and mobility, oxytocin administration or anesthesia, may affect the maternal microbiota and, consequently, mother–infant microbial transmission. Researchers have described the importance of maternal–neonatal microbial transmission, especially certain specific strains from the maternal gut microbiota, at the onset of infant colonization [37]. These birth-related factors may affect the microbial establishment process, and these cannot be studied independently. Therefore, our results sum up the consequences of all the concomitant factors affecting the different groups of study.
Thus, the hospital environment, especially in the case of CSs, could be considered an intervention-heavy condition characterized by the high pressure of antibacterial therapy and instrumentalization. HB rates have increased in Europe over recent decades, ranging from 0.1% in Sweden to almost 20% in the Netherlands [38]. However, limited information concerning the possible benefits of HB is currently available due to the lack of adequate randomized clinical trials meaning that it is not possible to determine the risk of neonatal–maternal mortality and morbidity during HB [39]. Further, little is known about the impact of a non-hospital environment on infant gut colonization.
The mode and place (hospital versus home) of birth shaped the neonatal microbiota. The microbiota of hospital-delivered infants was enriched with gram-positive anaerobic cocci, including the Peptoniphilus and Finegoldia genera, while those of HB infants exhibit higher relative abundances of species from the Enterococcus and Bifidobacterium genera.
In all three groups, time was the main factor affecting the composition of the infant microbiota, with a significantly different pattern being observed at birth than at the subsequent time points. Primary events concerning gut microbial colonisation may impact the assembly and acquisition of the early microbiome, and they may have long-lasting consequences for gut ecology [29].
At delivery, we observed higher diversity indices in the neonatal microbiota than at the subsequent time points; however, the quantitative total bacteria count was lower at delivery. This is likely caused by bias in the sequencing data obtained from low biomass samples and by higher contact with environmental contaminants. The HB infants had lower microbial diversity but a higher total number of bacteria than the hospital-born infants, likely due to the latter having increased contact with bacterial environmental contaminants [40].
We found clearly identifiable differences in the global microbiota structure during time between the three groups, even at the phylum level, which is in agreement with previous studies [4, 34, 40]. Interestingly, the differences in the colonization patterns become more pronounced over time, at least during the first month of life. These results indicate the importance of birth-related events to neonatal colonization processes, which may affect the microbiome composition in later developmental stages of infants. Similar to prior studies, we found that CS-born infants showed higher relative abundances of Clostridium and Klebsiella and lower species from the Bifidobacterium genus [6]. We observed that both VAG and HB infants had higher quantitative Bifidobacterium species than CS-born neonates. The depletion of the Bifidobacterium species in the gut environment has been described in relation to immune-related diseases [41, 42], and members of this genus are commonly used as probiotics due to their capacity to modulate microbiota–immune system homeostasis [43].
Could these shifts in the microbiota composition alter the functional profile of the neonatal microbiota?
Descriptive and observational studies are relevant to understanding microbiome evolution during the neonatal period; however, mechanistic studies of the host–microbiome interplay during early life are still required. Thus, microbiota functional analysis has been proposed as a tool for clarifying the host–microbiome interactions [44]. Our results showed functional differences between the place and mode of birth in the infant microbiota at delivery, 7 and 31 days. Several amino acid (AA) biosynthesis routes were over-represented in the vaginal-delivered microbiota when compared with the CS-born infants, including tryptophan-related paths. It is known that AAs serve as regulators of several metabolic pathways in the host [45]. Specifically, tryptophan interacts with the immune system through microbial serotonin production [46], regulation by TLRs or interactions within the aryl hydrocarbon receptor–microbiota–immune system relationship [47, 48].
We also found that vaginal-delivered infants, especially HB babies, had a microbiota enriched with LPS biosynthesis-related functions. Similarly, Wampach et al. found differences in the earliest functional profile according to the delivery mode, including LPS biosynthesis routes being enriched in vaginal deliveries when compared with CS-born neonates [49], which may influence immune system maturation and neonatal health.
Do these differences in the microbiota influence the host immune system response?
Despite studies in animal models highlighting the possible effects of microbial colonization patterns on the host gut epithelium maturation and immune system response, little evidence is available from human studies. To the best of our knowledge, this is among the first studies to address this important issue. The intestinal epithelium is the gateway through which gut microbiome–host crosstalk effects intestinal functionality in the form of enterocytes maturation, mucus production or epithelial barrier development, which means it is a key anatomical location for host–microbiome interplay.
The samples from HB infants exhibited a higher immune stimulatory capacity than those from hospital-born infants (both VAG and CS), with an increased ability to induce the expression of immune system-related genes (TLR4 and IRAK mRNA) and cytokine responses, including IL6 and IL8, in the HT-29 and THP-1 models. In concordance with our results, Wampach et al. identified the higher immunostimulatory potential of the microbiota of vaginal-delivered infants when compared with CS-born infants, although they used LPS purified from infant fecal samples and primary human macrophages differentiated into dendritic cells [49]. Combellick et al. noted the higher expression of TLR4 and IL8 by the HT-29 cell line following exposure to sterile fecal supernatant from HB infants when compared with hospital-born infants. However, they also found higher levels of TGF-β, which is mainly an anti-inflammatory cytokine, in hospital-born neonates [34]. We observed that the THP-1 cell line response was more affected by the mode and place of delivery than the HT-29 cell line, which highlighted the importance of the epithelial integrity and the innate immune system on the in vitro assessment of the host-microbiome interplay.
Most prior studies with similar objectives involved acute exposure on unique cell lines. We hypothesized that acute exposure to microbial products could not accurately reflect the biological effect of microbial metabolites and so proposed long-term (7-days) in vitro exposure assays, including the crosstalk between different cell types, which enabled us to obtain personalized results for each participant, thereby translating the individual signatures to the in vitro system.
In our model, the HB fecal supernatants induced higher gut barrier integrity (Fig. 5) and functionality (IAP; Fig. 5) following a time-dependent response that highlighted the relevance of the dynamics of host–microbiome interplay. Interestingly, the impaired closure of gut mucosal membranes has been shown, alongside higher intestinal permeability, in preterm infants who received formula feeding rather than breastfeeding [50]. This increased permeability could be related to allergic diseases in non-breastfed children [51] and to other health disorders [52]. In addition, the HB samples induced the expression of anti-inflammatory molecules (e.g. IL10, TOLLIP) to a higher extent than the CS samples, indicating negative feedback on inflammatory signaling in the gut. Specifically, IL10 down-regulate the microbiota-activated mucosal inflammatory cytokines and reinforce the gut epithelium barrier and control gut permeability, which are both essential to maintaining intestinal homeostasis [53].
Higher mucus production was observed after cell-exposure to fecal supernatants obtained from hospital-born infants when compared with HB infants. Both, microbiota [54] and TLR expression [55] are involved in the regulation of mucus production. Skoczek et al. found increased mucus production to be mediated by TLR signaling following the microbial invasion of the epithelial layer [56]. Despite lower mucus production in the gut being associated with disease phenotypes (e.g. inflammatory bowel disease, higher susceptibility to bacterial infections) [57] in adults, little is known about the role of mucins in neonatal colonization processes. We hypothesize that a penetrable mucus layer in newborns would allow for microbial colonization and interaction with the epithelium during the immune-priming window.
Many researchers have discussed the possible relationship between CS and altered immune system development [26]. Generally, CS delivery is associated with the poor stimulation of the immune system [58–60]. Urbanization is also associated with an impaired immune system response [61]. Some researchers have proposed non-diverse environments in early life, including delivery, to play a role in this effect. Furthermore, recent evidence has shown that early exposure to rural areas or farm environments could affect microbial composition and diversity [62]. This may be reflected in the observed reduction in the atopy risk in adults [63]. Kirjavainen et al. recently described how a farm-like indoor microbiota could also decrease the asthma risk in a non-farm environment [64].
However, most of these results were derived from observational studies, as very few mechanistic analyses have been conducted to date. Our results suggest a possible link between CS and the delayed maturation of intestinal function and the innate immune system. Such a link could play a significant role in the diseases associated with intervention-based deliveries, including autoimmunity, allergy and other immune- and metabolic-related disorders.
In our study, CS-born infants showed higher BMI and W/L z-scores during the first 18 months of life. Other researchers have reported similar results, associating CS with the risk of overweight in children [65, 66]. This could indicate the relevance of priority events, including those that alter microbial colonization, in infant health, thereby supporting the early programming hypothesis [67, 68]. Researchers have described how antibiotic therapy during early life modulates weight gain in different ways depending on the antibiotic dose in both animal [69] and human epidemiological studies [66]. It has been suggested that high-dose antibiotics can cause important reductions in the microbiota population, which may be related to the weight loss observed in some studies [70]. However, lower antibiotic doses would cause microbiota composition shifts, more than population size variation, and trigger the weight gain shown in the above-mentioned studies. Thus, it remains to be discovered whether the proposed mechanism could be extended to other perinatal factors that also disrupt the microbiota composition and transmission.
The limitations of this study include the low number of participants and the possible confounding factors not included in the analysis (e.g. maternal diet, lifestyle or number of siblings, pets among others). Our microbiota analysis was based on the taxonomic profile obtained via 16S rRNA gene sequencing, which offers less resolution than complete shotgun metagenome sequencing. As we used sterile fecal supernatant, we observed the effects of soluble bacterial metabolites and also, of non-bacteria-related products, including growth factors or eukaryotic extracellular vesicles, which may have influenced the observed results. Yet, fecal supernatants contain a complex array of molecules representative of the in vivo condition, which retain inter-individual differences and features. The use of cell lines may hamper the translational results, although it offers a reproducible and economically viable strategy for further testing on more physiologically relevant models. Among the strengths of the study are the inclusion of three groups and the comparison of CS and vaginal delivery at both hospital and home, including 18 months of follow up. We performed the cellular exposure assays in cellular models with different degrees of complexity and different exposure times, including the epithelial barrier function and maturation as relevant targets, together with the innate immune response.