Dynamics of the preterm gut microbiome in health and disease

Abstract

Advances in metagenomics have allowed a detailed study of the gut microbiome, and its role in human health and disease. Infants born prematurely possess a fragile gut microbial ecosystem that is vulnerable to perturbation. Alterations in the developing gut microbiome in preterm infants are linked to life-threatening diseases such as necrotizing enterocolitis (NEC) and late-onset sepsis; and may impact future risk of asthma, atopy, obesity, and psychosocial disease. In this mini-review, we summarize recent literature on the origins and patterns of development of the preterm gut microbiome in the perinatal period. The host-microbiome-environmental factors that portend development of dysbiotic intestinal microbial patterns associated with NEC and sepsis are reviewed. Strategies to manipulate the microbiome and mitigate dysbiosis, including the use of probiotics and prebiotics will also be discussed. Finally, we explore the challenges and future directions of gut microbiome research in preterm infants.

INTRODUCTION

The human gastrointestinal (GI) tract harbors a complex and diverse microbial community which is dominated by bacteria, but also includes viruses, archaea, fungi, and other eukaryotes. This complex ecosystem with 10¹²⁻¹⁴ cells from 100 to 1,000 species is considered to be the most dense microbial habitat on earth (1). Compared with ∼23,000 genes of human genome, the gut microbiome has been referred to as a “hidden” metabolic organ and encodes over 3 million genes producing thousands of metabolites (2). Briefly, 90% of the GI microbiota is composed of bacteria from two major phyla namely Bacteroidetes and Firmicutes. Other phyla consistently found in human gut include Proteobacteria, Actinobacteria, Fusobacteria, and Verrucomicrobia (3). At lower taxonomic levels however, individuals differ considerably in composition of fecal microbiota and each individual may have his or her own distinctive pattern of microbial profile (4).

Studies suggest that the early pattern of infant gut microbial colonization is critical for proper development of the human GI tract. The neonatal gut microbiota plays an essential role in the acquisition of postnatal intestinal endotoxin tolerance (5) and specifically regulates maturation of regulatory T-cells (CD4+, Foxp3+), natural killer cell, and gamma delta T-cells (6). Dysbiosis or imbalance in gut microbial communities during this critical period is linked to several diseases including inflammatory, metabolic, neurologic, cardiovascular, and GI illnesses (7). Compared with term neonates, infants born prematurely are at greater risk for disruptions to the gut microbiota. Herein, we briefly review commonly used metagenomic methods for microbiome analysis, compare the challenges of preterm versus term gut microbiome, delineate antecedents and consequences of preterm gut dysbiosis, and discuss advances in microbiome modulation therapy in preterm infants. We focus predominantly on the bacterial communities as most research has been done in this area.

METHODS OF MICROBIAL PROFILING WITH 16S AND SHOTGUN SEQUENCING

Currently available high-throughput sequencing technologies used to profile the genomic composition of a microbial community in a culture-independent manner can be subdivided into two different approaches, namely 16S rRNA gene sequencing and shotgun metagenomics. In 16S rRNA sequencing, the 16S rRNA gene is used as a genetic fingerprint to decipher bacterial phylogeny and taxonomy. A component of the 30S subunit of prokaryotic ribosome, the 16S rRNA gene is 1,500 bp long and comprises nine variable regions (V1–V9) interspersed among the conserved regions of its sequence. These hypervariable regions are unique to each bacterial taxon, whereas the conserved regions allow for the development of universal primers that bind to known sequences shared among most bacteria. During data analysis, amplified 16S rRNA sequences are assigned to operational taxonomic units (OTUs) and OTU clustering methods using a moderately stringent sequence identity threshold allowing for basic taxonomic assignments (family level or higher) (8). Amplicon sequencing targeting the 16S rRNA gene is limited by differences arising from the different variable region chosen for PCR amplification, sequencing errors, poor discriminatory power for some genera, and severely limited resolution at the species level due to short-read amplicon sequencing, potential for contamination, and sequencing platform used (9–11).

Unlike 16S sequencing, which only targets 16S rRNA sequences, shotgun metagenomics generates whole genome sequencing data that can be compared with existing reference databases for accurate taxonomic identification. By sequencing all genetic information within a sample, shotgun metagenomic studies can extract both taxonomic and functional information from complex microbial communities and therefore guide phenotypic studies to understand their potential roles in health and disease. Although shotgun metagenomics can be used to extract species- and even strain-level information through computational approaches, it can still be challenging to differentiate between closely related or coexisting bacterial taxa (12). This has mandated the need for extensive and well-characterized collections of reference genomes including those from the Human Microbiome Project and the Human Gastrointestinal Bacteria Genome Collection (13, 14).

CHALLENGES OF THE PRETERM INFANT THAT SHAPE THE GUT MICROBIOME IN COMPARISON WITH THE TERM INFANT

Preterm infants face several unique environmental and host conditions that negatively impact the development of their gut microbiome. Even before birth, ∼25%–30% of preterm infants are exposed to microbes in the context of preterm premature rupture of membranes and intra-amniotic infection (15). Cesarean section, with subsequent gut colonization by skin microbiota rather than vaginal and rectal microbiota from vaginal delivery, is also more common in preterm than term infants (16). After delivery, most preterm infants need stabilization and subsequent care in the neonatal intensive care unit which, while necessary, is also invasive and results in exposure to the hospital microbial environment. Consequently, as risk of serious nosocomial infection is high, preterm infants are often exposed to powerful broad-spectrum antibiotics throughout their hospital stay. Preterm nutrition and feeding practices—such as delayed introduction of enteral feedings, frequent withholding of feeds, use of acid-suppressive medications, and provision of exclusive human milk diet or formula—are also major modifiers of gut microbiota composition (17). Together, these environmental factors interact with the intrinsically immature GI tract and immune system of the preterm infant in a highly complex but poorly defined process that significantly impact the developing preterm gut microbiome (18, 19).

In general, the intestinal microbiota in the preterm infant differs from the term infant in having delayed colonization, fewer bacterial species, and less diversity and abundance (18, 20, 21). There is also a predilection to being colonized by potentially pathogenic facultative anaerobes (e.g., Enterobacter, Escherichia, and Klebsiella) and decreased levels of commensal strictly anaerobic organisms (e.g., Bifidobacterium, Bacteroides, and Clostridium) (Table 1) (27). A shared patterned progression of microbial communities from Bacilli to Gammaproteobacteria to Clostridia has also been described in preterm infants (28). The pace of this patterned progression is predominantly influenced by gestational age, whereas the mode of delivery, antibiotic exposure, and type of feeding appear to have less impact (28).

Table 1.

Major differences in bacterial gut colonization of preterm infants compared with term infants

Bacterial Groups	Important Characteristics and Function
Decreased in preterm infants
Bacteroidetes (e.g., Bacteroides fragilis)	Anti-inflammatory function through surface component polysaccharide A (22)
Actinobacteria (e.g., Bifidobacterium infantis, Bifidobacterium breve)	Anti-inflammatory function through secretion of short chain fatty acids (23), and acceleration of maturation of intestinal innate immune response genes (24)
Increased in preterm infants
Firmicutes (e.g., Enterococcus faecalis, Staphylococcus)	Gram-positive cocci that can act as opportunistic pathogens, especially in patients with prolonged hospitalization or has received multiple antibiotic therapy (25)
Proteobacteria (e.g., Escherichia coli, Klebsiella pneumoniae)	Gram-negative bacteria commonly associated with opportunistic nosocomial infections; have an outer membrane that contains the endotoxin lipopolysaccharide (26)

Whether the differences in gut microbiota of preterm versus term infant constitute true dysbiosis or merely represent the common gut microbiome signature in very-low-birth weight infants is debatable. Current studies do not clarify whether a distinct pattern of dysbiosis that portends disease susceptibility exists, apart from a modest increase in Enterobacteriaceae in infants who develop necrotizing enterocolitis (NEC), as further discussed in consequences of dysbiosis in preterm infants.

ANTECEDENTS OF DYSBIOSIS IN PRETERM INFANTS

Prenatal Factors

The development of intestinal microbiome in the preterm neonate and the proclivity to a dysbiotic signature are influenced by several prenatal factors (Fig. 1). Maternal gestational diet appears to alter the composition of the first stool after birth (meconium), with depletion of Bacteroides observed with high-fat diet (29). Other studies have described an interactive effect between maternal diet and mode of delivery on the neonatal gut microbiome, with fruit intake increasing intestinal Streptococcus/Clostridium in infants delivered vaginally, whereas increased dairy intake was associated with higher abundance of Clostridium among infants delivered by cesarean section (30). In pregnancies complicated by intrauterine amniotic infection or chorioamnionitis, a higher relative abundance of Mycoplasmataceae and phylum Bacteroidetes has been identified in stool of infants postnatally (31). Although previous studies had suggested the presence of microbial communities in the placenta and amniotic fluid of healthy pregnancies potentially colonizing the gut prenatally, more vigorous recent studies have disproved the existence of a distinct placental microbiome (32, 33).

Figure 1.

Summary of antecedents and consequences of gut dysbiosis in preterm infants. Several prenatal, birth, and postnatal factors negatively impact the developing gut microbiome of the preterm infant. The resulting dysbiotic gut microbial communities during this critical period of development have been linked to both short-term and long-term morbidities. *Highlights factors considered to have relatively greater impact than others. Figure 1 created using www.biorender.com.

Intrapartum antibiotics (IAP) administered to prevent Group B streptococcal transmission has also been shown to have significant early effects on the composition of gut microbiome in neonates that persist through the first year of life (34–36). Nogacka et al. (36) showed a persistent increase in Proteobacteria and Firmicutes with decreases in Acinetobacter and Bacteroides in association with higher occurrence of some bacterial genes that code for β-lactamase resistance. Others have shown that the use of IAP was associated with decreases in Bifidobacteria spp. at 7 days of life, whereas Lactobacillus spp. remains unchanged (35). Furthermore, the effects of IAP on the infant microbiome may be selective with penicillin suppressing Bacteroides spp., and cephalosporin delaying the increase in Bifidobacteria (34). These studies suggest that maternal gestational diet, chorioamnionitis, and IAP influence the composition of the neonatal microbiome and these effects can last well beyond the neonatal period.

Postnatal Factors

Several postnatal factors influence the development of gut microbiome and can program intestinal dysbiosis (Fig. 1). The mode of birth, i.e., vaginal versus cesarean, has been shown to influence the neonatal microbiome (16, 37, 38). Infants born by cesarean section have less intestinal microbial diversity, decreased colonization with Bifidobacteria, Bacteroides, and Lactobacilli, and tended to have increased skin microbiota such as Streptococcus and Staphylococcus in the first weeks after birth (16, 38). Interestingly, differences related to mode of delivery are less evident with maturity, likely because other factors like postnatal diet impact the gut microbiota (37).

Early exposure to parenteral antibiotics is also a risk factor for gut dysbiosis (39, 40). Fouhy et al. (41) demonstrated that early (<48 h) antibiotic treatment with ampicillin and gentamicin was associated with increases in Proteobacteria and decreased abundance of Actinobacteria, Bifidobacteria, and Lactobacillus. Gibson et al. (42) demonstrated that broad spectrum antibiotics such as meropenem, cephalosporins, and ticarcillin-clavulanate decreased species richness whereas vancomycin and gentamicin had nonuniform results. They also noted emergence of antibiotics resistant genes that was antibiotic- and species-specific. Zwittink et al. (43) noted that the use of short- or long-term amoxicillin/ceftazidime treatment was associated with significant higher abundance of Enterococcus during the first two postnatal weeks at the expense of Bifidobacterium and Streptococcus.

Nutrition is also a strong modifier gut microbiome. Preterm infants fed with formula milk exhibit a less diverse gut microbiome with less enrichment of Clostridiales and Bifidobacteria compared with preterm infants fed mothers’ own milk (44). In summary, major postnatal factors linked to a dysbiotic gut signature in preterm infants include mode of delivery, exposure to antibiotics and other medications, and type of milk feeding.

CONSEQUENCES OF DYSBIOSIS IN PRETERM INFANTS

Necrotizing Enterocolitis

NEC is the most widely recognized consequence of gut dysbiosis in preterm infants (Fig. 1). The role of gut bacteria in NEC is demonstrated from mouse models showing absence of disease in a germ-free environment, direct studies of human stools showing dysbiotic signatures, and demonstration of bacterial invasion by fluorescent techniques in surgically resected human tissues (45–47). Clusters of NEC outbreaks related to specific organisms that have varied by institution have also been reported, providing additional evidence of the role of microbes in NEC pathogenesis. With advances in metagenomics, several studies have demonstrated that preterm infants who develop NEC have altered gut microbiota signatures compared with preterm infants who do not develop NEC, though these signatures have varied between studies (48).

In a small study, Heida et al. (49) identified enrichment of Clostridrium perfringens in meconium of preterm infants who subsequently developed NEC, suggesting that a pathogenic microbial signature for NEC may be present in the very first day of life. However, meconial or very early stool signatures for NEC have not been demonstrated in subsequent studies (21, 50). In one of the largest longitudinal studies in preterm infants, Warner et al. (50) studied >2,400 stool samples from 122 infants born with a birthweight <1,500 g to demonstrate that NEC was associated with Gammaproteobacteria abundance, and relative depletion of Negativicutes, an anaerobe. Interestingly, these patterns emerged typically after the first 4 weeks of life. Other interesting observations were that antibiotic use increased proportions of Bacilli, whereas vaginal birth was associated with decreased proportions of Bacilli, and the volume of human milk feeds did not impact bacterial community architecture.

Similar to the landmark study by Warner et al. (50), several investigators have reported an enrichment of potentially pathogenic Gram-negative bacteria in the intestine in addition to a relative decrease in the abundance of Firmicutes and Bacteroides before the onset of NEC (21, 50–53). Furthermore, although preterm infants in general have low gut microbial diversity compared with term infants, several studies have noted further reduced levels of diversity in preterm infants who develop NEC compared with preterm infants without NEC (21, 48, 54, 55). A meta-analysis of published studies confirmed a relative abundance of Gammaproteobacteria, with decreased relative proportions of Firmicutes and Bacteroides preceding NEC, but did not confirm reduced microbial diversity before NEC onset (21). Gammaproteobacteria, which include several well-known pathogenic species that have the endotoxin lipopolysaccharide, can induce excessive intestinal and systemic inflammation through stimulation of Toll-like receptors (45); whereas Firmicutes and Bacteroides can enhance gut development by facilitating intestinal epithelial cell differentiation and preserving mucin and tight junction integrity (56).

Although these studies have primarily focused on bacterial signatures, a potential role for viruses such as norovirus, rotavirus, and cytomegalovirus (CMV) in NEC causation is also recognized (48, 57). For example, case reports (58) have described infants with CMV-associated NEC, and retrospective studies using surgical specimens (59, 60) have detected CMV in intestinal tissue of preterm infants with NEC or spontaneous intestinal perforation. Infants with CMV-associated NEC tend to have lower gestational age and thrombocytopenia than non-CMV associated cases. The relative impact of viruses on NEC remains poorly characterized as reported rates of detection of CMV and other viruses can vary depending on the sample (blood, stool, and intestinal tissue), technique (immunohistochemistry, polymerase chain reaction, enzyme-linked immunosorbent assay), and sample size of NEC cohort (61–63). The potential role of fungi and the mycobiome in NEC also remain largely unexplored (64).

Although metagenomic studies provide a snapshot of microbial patterns associated with NEC, more rigorous studies are needed to elucidate their mechanistic role and establish their contribution in NEC. More recently, multi-omic studies combining metagenomics, metatranscriptomics, and metabolomics are beginning to shed new light on bacterial replication rates, quorum sensing patterns, and metabolic profiles associated with intestinal dysbiosis preceding NEC (21, 52, 55). For example, Morrow et al. (51) evaluated the urinary metabolome and gut microbiome of 11 preterm infants with NEC and 21 matched controls. Using this two complementary “-omic” approaches, the authors found that the urinary metabolite alanine was directly correlated with Firmicutes and indirectly correlated with Proteobacteria in the gut, suggesting that specific intestinal bacteria can influence production and utilization of specific metabolites which may be useful as surrogate biomarkers for early diagnosis of NEC. Although still at its infancy, these novel multi-omic approaches are anticipated to further advance our understanding and provide a full mechanistic picture of NEC pathogenesis (65).

Sepsis and Other Diseases

The consequences of dysbiosis in the preterm infants extend beyond the predisposition to develop NEC (Fig. 1). Stewart et al. (40) noted that the bacteria isolated from a blood culture during sepsis corresponded to the dominant bacterial genera in the gut microbiome, and that Bifidobacteria abundance was observed in infants who did not develop sepsis or NEC. Other studies have also noted decreased diversity and lower abundance of Bifidobacteria among infants who developed sepsis (66). Potential mechanisms by which Bifidobacteria can protect against sepsis include reduced bacterial translocation through enhanced intestinal barrier function and increased production of beneficial short-chain fatty acids like acetate (67, 68). Sepsis related to antibiotic-resistant strains of bacteria originating from the gut is also increased in preterm infants treated with broad spectrum antibiotics that increase dysbiosis (40, 42, 50). Although not the focus here, dysbiosis in preterm infants has also been linked to bronchopulmonary dysplasia (a chronic debilitating lung condition of infants) and growth failure (69, 70).

Potential Long-Term Consequences

Epidemiological studies suggest a link between early disturbances to the gut microbiome from antibiotics or cesarean delivery with later childhood diseases including allergy (71–73), obesity (74, 75), and attention-deficit hyperactivity disorder (76). A recent population-based cohort study using the Rochester Epidemiological Project found similar associations between early-life antibiotic exposure and increased risk for several childhood immunological, metabolic, and neurobehavioral health conditions (77). Although these studies did not differentiate term from preterm infants, the aforementioned stressors that preterm infants experience in the hospital environment puts them at greatest risk for early gut microbial dysbiosis (78). Additional studies in mice provide supporting evidence for the hypothesis that early-life perturbations during key developmental periods can have long-term consequences. For example, in a mouse model of early gut microbial disruption with low-dose penicillin, long-lasting metabolic effects of increased fat mass were observed even with limited antibiotic administration where microbial communities recovered after cessation of antibiotics (79).

MICROBIOME MODULATION THERAPY

With accumulating evidence of how gut microbiome perturbations contribute to preterm diseases, novel strategies that reverse or alleviate dysbiosis are being explored.

Probiotics

Probiotics are defined as living microorganisms that confer health benefits when administered in adequate amounts. In addition to favorable alteration of gut microbiome, probiotics are now recognized to confer benefit through other mechanisms such as colonization resistance against pathogenic bacteria, enhanced gut barrier function, and immunomodulation (80). Numerous clinical trials have demonstrated the safety and efficacy of prophylactic administration of probiotics in preterm infants for the prevention of NEC. A recent network meta-analysis (63 trials, 15,712 preterm infants) identified that a multispecies combination that contains Lactobacillus and Bifidobacterium provided superior benefits for NEC prevention compared with single species or other multispecies probiotic combinations for reducing mortality and NEC in preterm infants (81).

Prebiotics

Prebiotics are nutrient substrates that selectively enrich beneficial microbiota in the gut. The most important source of prebiotics for preterm infants is human milk oligosaccharides (HMO)—complex sugars not otherwise digestible by the human gut but utilized by specific microbes such as Bifidobacteria and Bacteroides to promote their growth and activity. In addition to stimulating probiotic growth, prebiotics also help improve intestinal mobility and gut barrier function. Artificial or nonhuman milk oligosaccharides designed to function in a similar manner to HMOs have also been tested in small clinical trials with promising results (82). Other active areas of investigation include synbiotics which are combination products that contain both prebiotics and probiotics (83).

Postbiotics

Postbiotics are metabolic byproducts of probiotic bacteria that can exert positive biological activity in the host. In a recent study, Zheng et al. (23) demonstrated that short-chain fatty acids, such as butyrate and acetate produced by Bifidobacterium infantis from metabolism of complex carbohydrates expressed in breastmilk, have anti-inflammatory effects in immature enterocytes. Other studies have shown how butyrate can also enhance intestinal barrier function and regulate mucosal immunity (84). By using metabolites instead of bacteria, postbiotics potentially provide an effective, simpler, and safer alternative to ingestion of live microorganisms. A related strategy is the use of para-probiotics, which are inactivated microbial cells or cell fractions typically produced by heat killing of probiotics (85).

Fecal Microbiota Transplantation

In fecal microbiota transplantation (FMT), fecal material from a healthy donor is transplanted to the recipient’s GI tract to restore microbial homeostasis. Compared with probiotics, FMT allows more robust and longer-lasting community of diverse microbes with a single dose. Safety concerns include inadvertent introduction or promotion of pathogenic species and transference of antibiotic resistance. FMT in preterm infants remains largely unexplored except in experimental animal research (86). A similar thematic approach is vaginal seeding, wherein bacteria from the mother’s vaginal tract is transferred to the face and body of infants born via cesarean section (87).

CONCLUSIONS

The concept of the holobiont, consisting of the human host and the resident microbial communities, is important in considering perturbations of intestinal microbial assembly in preterm infants (7, 88). The development of a symbiotic intestinal ecosystem is a central event for successful adaptation to postnatal life. Maternal disease, intrauterine infection, and perinatal use of antibiotics can disrupt establishment of coevolved microbial communities, which are integral for the optimal function and development of the intestine, the immune system, the brain, and other physiological processes (7, 88). Restoration to a normal microbial assembly is further impeded in the preterm neonate fed formula milk, exposed to long-term antibiotics and deviant microbial dispersal patterns in a pathogen-rich milieu, and microbial community priority effects. These adverse influences result in the establishment of a dysbiotic microbial signature characterized by a relative abundance in Gammaproteobacteria, decreased diversity, and a modest decrease in Bifidobacteria. Several studies link these dysbiotic signatures to increased vulnerability to life-threatening neonatal illness such as NEC and sepsis, and increasingly to childhood asthma, allergic disease, disorders of mood, and obesity (7, 88). Efforts at restoring microbial community structure using probiotics have shown promise in alleviating NEC and sepsis risk. Other approaches based on prebiotics or enteral immune therapy, however, warrant careful trials. Understanding the role of host genetics in programming the neonatal gut microbial assembly, combining metabolomic and transcriptomic signatures concurrent with dysbiotic microbiota signatures, and personalized approaches in reconstituting the deviant microbial assembly in preterm neonates to prevent disease and ensure healthy outcomes remain highly significant areas for translational research (55, 89, 90).

GRANTS

This study was supported by institutional funds at Children’s Mercy Hospital (to V. Sampath and A. Cuna), R01DK117296 (to V. Sampath), and K08DK125735 (to A. Cuna).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.C. and V.S. conceived and designed research; A.C., M.J.M., and V.S. prepared figures; A.C., M.J.M., S.U., and V.S. drafted manuscript; A.C., M.J.M., I.A., S.U., and V.S. edited and revised manuscript; A.C., M.J.M., I.A., S.U., and V.S. approved final version of manuscript.