GUT MICROBIOTA OF HEALTHY NEWBORNS: NEW DIAGNOSTIC TECHNOLOGIES — NEW OUTLOOK ON THE DEVELOPMENT PROCESS

Currently, there are no criteria allowing to adequately assess composition and volume of the newborns' gut microbiota, which prevents early detection of the pathological processes and appropriate intervention. This study aimed to apply the methods of culturomics, proteomics and molecular genetic technologies to investigate the development of gut microbiota in healthy newborns delivered in the city of Moscow both vaginally and through a cesarean section. We examined 66 children, 33 of them delivered vaginally and 33 by cesarean section. The luminal bacterial flora samples were collected on the 1st, 7th and 30th days of life. There were 136 species of microorganisms belonging to 40 genera identified. We established that cesarean section slows down normal development of the gut microflora: through the follow-up period (1 month of life), gut microbiocenosis in such children did not yield the results on par with those registered in children born vaginally. Bifidobacteria were significantly more common in the vaginal delivery group: 84% of 10 9 –10 12 CFU/g versus 33% of 10 5 –10 12 CFU/g in the cesarean section group. At the same time, the former group had significantly less clostridia (33.3% and 65.4%, respectively) and lactose-negative Escherichia coli strains (2.4 and 19.4%, respectively) than the latter group.

Against this background, the currently adopted gut microbiota standards (composition and volume) [1][2][3] require revision. To date, there are no clearly defined criteria for the newborns' gut microbiota assessment, the criteria that would allow timely interventions addressing the problems detected and early measures arresting pathological processes, such as necrotizing enterocolitis in premature babies. Very few researchers undertook statistical processing of the data describing composition and volume of the newborns' gut microbiota [4,5].
This study aimed to apply the methods of culturomics, proteomics and molecular genetic technologies to investigate the development of gut microbiota in healthy newborns delivered in the city of Moscow both vaginally and by a cesarean section.

METHODS
A cohort of healthy full-term newborns was selected for this prospective study. The cohort was randomized by blocks, which yielded two groups of children: 1) group I -33 children delivered vaginally; 2) group II -33 children born by caesarean section. The inclusion criteria were as follows: gestational age 38-40 weeks; Apgar score of 8-9 points (first and fifth minutes of life); body weight over 2800 g. Blood and urine of the newborns were tested for infectious pathologies or other factors that could affect development of the gut microbiota. All children were breastfed; feeding formulas were used in isolated cases only. No mother had a prolonged latency period; 3 women took antibiotics during pregnancy; 4 women had cervical incompetence (6%). In the surgical delivery cases the women received amoxicillin/clavulanic acid for prevention purposes before and during surgery.
The feces were collected from the newborns three times: on the 1st day, at the end of the 1st week and at the end of the 1st month of life. The samples of meconium obtained during the first bowel movement or feces collected from a sterile diaper and put into a sterile plastic container were delivered to the laboratory within 2 hours; plating was performed immediately upon delivery.
We studied the gut microbiota using an extended range of selective and non-selective culture media. The incubation was done in aerobic, microaerophilic and anaerobic conditions.
The microorganisms were identified with the help of a MALDI-TOF AutoFlex III time-of-flight mass spectrometer powered by the MALDI BioTyper v3.0 software (Bruker Daltoniks; Germany). At SCORE > 2.0, the culture was marked as identified on the species level (high probability of identification). At SCORE 1.7 to 2.0, the culture was considered identified at the genus level.
When the SCORE was below 2.0, i.e. the strain was hard to identify, we did the 16S rRNA gene sequencing. The Proba-CiTo (DNA-Technology; Russia) reagent set enabled extraction of the DNA. We used two DNA fragments for sequencing: an amplicon of ~440 bp (positions 339-785 of the 16S rRNA gene) and an amplicon of ~1340 bp (positions 42-1380 of the 16S rRNA gene). These were amplified using the DTprime detecting amplifier (DNA-Technology; Russia). For sequencing, we employed a 3130 Genetic Analyzer (Applied Biosystems; USA) and used the BigDye™ Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems; USA) following the protocol provided by the manufacturer. For species identification, the BLAST software package (National Center for Biotechnology Information; USA), freely available on the Internet, was used.
To determine differences in the frequency of occurrence of microorganisms depending on the mode of delivery, we applied the Fisher's exact test during statistical processing of the data. Median and interquartile range were used as descriptive statistics to characterize the degree of microbial contamination.

RESULTS
At the 1st time point we examined 66 children, 33 in each group. The meconium of every 5th child was sterile; in group II such a situation was observed twice as often as in group I (9 and 4 cases, respectively; p > 0.05). The non-sterile meconia featured diverse microflora (63 species; the average species diversity index was 3.8 species per child). The bacteria isolated most often (in 60% of the vaginally delivered babies, titer of 10 2 -10 3 CFU/g, and in 70% of the cesarean section babies, titer of 10 2 -10 4 CFU/g) were gram-positive facultative anaerobic microorganisms of 10 genera: Staphylococcus, Streptococcus, Enterococcus, Micrococcus, Gemella, Globicatella, Granulicatella, Rothia; Corynebacterium, Bacillus. It is noteworthy that in 12.2% of the group II children and in 6.1% of the group I children we detected Staphylococcus aureus (S. aureus). Escherichia coli, which normally makes up a significant part of the normal intestinal flora, was isolated thrice as often in the vaginally born babies (24.2% and 9.1%; p > 0.05). Same is true about other enterobacteria (27% and 15%, respectively; p > 0.05). Only one baby had Pseudomonas aeruginosa (group II).
Lactobacilli and bifidobacteria play the most important role in the development of the newborn's gut microbiota. These microorganisms were found as early as in the first day of life. The size of the lactobacilli colonies was significantly greater in group I (18% and 3%, respectively; p > 0.05); the lactobacilli forming those colonies were Lactobacillus crispatus and Lactobacillus jensenii, which are among the dominant species in the vaginal microbiota of healthy women. Bifidobacteria were also isolated more often in the sampled taken from the group I babies (9% and 3%, respectively; p > 0.05); the bacteria were Bifidobacterium bifidum, B. longum, B. adolescentis. Only 1 baby (group I) had bacteroids detected in the first day of life. Other anaerobes (veillonella, prevotella and campylobacter) were isolated almost exclusively in group II.
Through the first week of the study, some parents refused to continue participation and some children contracted an infectious disease. Therefore, at the 2nd time point we examined 52 babies, 25 from group I and 27 from group II.
Compared to the 1st time point, the average species diversity index increased almost 3-fold and reached 9.4 species per child. As in the first day of life, both groups had facultative anaerobic microorganisms (9 genera) detected most often. Their titer increased considerably: in the vaginal delivery group it was 10 7 -10 11 CFU/g, in the cesarean section group -10 7 -10 12 CFU/g; the frequency of occurrence was 100% in both groups. S. aureus became a more common bacterium: we isolated it in 44% of the group I babies and in 55.6% of the group II babies. The diversity of enterococci expanded from two to six species: Enterococcus faecalis, E. durans, E. faecium, E. gilvus, E. avium, E. gallinarum. Same as in the first day of life, E. faecalis dominated gut microbiota of the majority of participants, colonizing the intestines of 80% of the group I children and 88.9% of the group II children.
The range of gram-negative facultative anaerobic bacteria present extended from 6 to 15 species (Enterobactеriaceae). E. coli was still detected more often in the vaginally born babies (64% in group I and 44.4% in group II; p > 0.05). The frequency of isolation of other enterobacteria was approximately the same in both groups.
The composition of lactobacilli species extended with L. paracasei, L. curvatus, L. gasseri, L. fermentum, L. rhamnosus. It should be noted that the species of lactobacilli dominating vaginal microbiota of healthy women (L. crispatus, L. jensenii и L. gasseri) were only found in the group I children.
We registered statistically significant differences in the representation of bifidobacteria. In group I such were detected in 84% of samples with the titer of 10 9 -10 12 CFU/g, and in group II the figure was only 33%, with the titer of 10 5 -10 12  Thus, by the 7th day of life microbiota of the participants featured significantly more species of various genera, both facultatively anaerobic and obligate anaerobic. Bifidobacteria, lactobacilli and bacteroids were prevalent in group I.
At the 3rd time point we examined 50 children, 24 from group I and 26 from group II. The average species diversity index was 9.4 species per child. As in the earlier periods, grampositive facultative anaerobic bacteria (7 genera) were the most common. The frequency of detection of S. aureus remained largely unchained compared to the 7th day in group I (58.3%) and significantly increased in group II (76.9%). There were 2 new species of enterococci detected: E. casseliflavus and E. raffinosus.
As for the composition of gram-negative facultative anaerobic microflora, we isolated 11 Enterobacteriaceae species and 1 species of non-fermenting bacteria, Stenotrophomonas maltophilia. E. coli were still detected more often in the vaginally born babies (75% in group I and 61.5% in group II; p > 0.05). Samples from the group II children had other enterobacteria slightly more often (66.7% and 73%, respectively; p > 0.05).
We started seeing a statistically significant difference in the frequency of detection of clostridia (C. perfringens, C. butyricum, C. tertium, C. ramosum, C. paraputrificum, C. difficile): 33.3% of children in group I and 65.4% in group II (p < 0.05). There were twice as many species of clostridia in group II. One child from group II had C. difficile.
Candida yeast was found in both groups only on the 30th day of life; less than 10% of children had them.
Thus, regardless of the method of delivery, on the 30th day of life the gut microbiota featured multicomponent microbial associations of facultative and obligate anaerobes. At the same time, certain differences between obligate and transient microbiota components became obvious. The composition of lactobacilli was more diverse (almost twice as many species) in babies delivered vaginally, the species detected more often in their samples were those inhabiting vagina of healthy women. Bacteroids were predominantly isolated in the same group I children, while clostridia, with greater variety of species, were more peculiar to the C-section babies. The conditionally pathogenic facultative anaerobes (S. aureus and enterobacteria, with the exception of E. coli) more often colonized children by a cesarean section.
The hemolytic strains of Enterococcus sp. and E. coli occurred with the same frequency in both groups: Enterococcus sp. -8.3% in each group, and E. coli -9.5 and 9.7%, respectively. Lactose-negative strains of E. coli were isolated in the group II children 8 times more often than in group I, which is a significant difference (19.4  Identification of genera of 36 isolated bacteria (SCORE < 2.0) required 16S rRNA sequencing. In most cases, sequencing results confirmed findings of the MALDI-TOF MS study. Only for 5 isolates 16S rRNA gene sequencing was the only method allowing reliable identification. In some cases with closely related microorganisms it was not possible to identify the species neither with sequencing nor with MALDI-TOF MS. These species were: Bifidobacterium kashiwanohense/ Bifidobacterium pseudocatenulatum/Bifidobacterium catenulatum, Lactobacillus casei/Lactobacillus paracasei, Actinomyces naeslundii/Actinomyces viscosus, Actinomyces radingae/ Actinomyces ihumii. Eleven isolates produced discrepancies in identification: for 7 of them, it was impossible to find out the genera correctly with the help of MALDI-TOF MS, and 4 presented problems with the species identification. Median (CFU/g) 10 2 10 2 10 8 10 9 10 9 10 9 Interquartile range (CFU/g) 10 2 10 2 -10 3 10 8 -10 9 10 8 -10 10 10 8 -10 10 10 8 -10 10

DISCUSSION
Until recently, it was believed that the gastrointestinal tract of a newborn remains sterile for the first 10-20 h (aseptic phase) [7], and the bacteria first to colonize the intestines of a vaginally delivered child are the lactobacilli from the mother's vaginal microflora [8][9][10][11]. However, some experimental research efforts gave rise to another point of view [12]. According to the recently published data, a fetus receives foundation of the gut microflora through bacterial translocation from his mother during the second half of pregnancy; some researchers argue sterility around the fetus [13][14][15]. In our study, we found that the meconium collected during first defecation in the first hours of life was unsterile in 53 of 66 newborns (80%), which does not exclude intrauterine colonization.
The published works state that the first hours and the first day of life are the time of active colonization, when the colonies of Escherichia coli, enterococci develop rapidly; the period is known as the growing colonization stage and its course is independent from the degree of maturity, perinatal fetus development conditions and the type of feeding [16][17][18][19]. The enterobacteria titer during that period reaches 10 9 CFU/g of feces [20], and the anaerobes -bifidobacteria, lactobacilli, bacteroids -are usually absent [21]. According to the data we obtained, the first hours of life were the time when the intestines were mainly colonized by the gram-positive cocci (Staphylococcus spp, Streptococcus spp), and the cesarean section babies had larger colonies than the vaginally delivered children (frequency of occurrence 70 and 60.6%, titers up to 10 4 and 10 3 CFU/g, respectively). During this period, enterobacteria and obligate anaerobic microorganisms were discovered rarely and with low titers. At the same time, already in the first hours of life microbiota of the vaginally born children featured more bifidobacteria and lactobacilli than that of babies delivered by a Caesarean section: 9 and 3%, respectively, and 18 and 3%, respectively. Bacteroids were detected only in 3% of children born vaginally. A possible reason for the discrepancy between our data and the results demonstrated by other authors is the fact that we aimed to obtain the sample biomaterial directly during the first act of defecation of the newborn.
Further on, healthy babies have anaerobic and facultative anaerobic components of their microflora grow actively, the volumes reaching 10 5 -10 7 CFU/g [15]. By the 6th day, aerobic and anaerobic parts of the microflora become balanced, then bifidobacteria and lactobacilli begin to develop rapidly and by the 2nd month of life they reach 10 9 -10 10 CFU/g. According to other researchers, days 3-5 mark the beginning of the microflora transformation stage, which results in bifidobacterial flora replacing other microorganisms. This is when bifidobacteria become the core (resident) microflora of the intestines [22][23][24]. The bifidoflora grows dominant by the 5-20th day of life [25]. Our discoveries are different: the total number of microorganisms increases, reaching 10 10-12 CFU/g. This suggests that the facultative aerobic and obligate anaerobic components of the microbiota also achieve balance by the first week of life, but there are more various microorganisms colonizing the intestines. Another fact deserving a special note is that in the group II babies bifidobacteria did not colonize the gut actively while obligate anaerobes prevailed there (mainly veilonella, clostridia).
By the end of 1st month of life, both groups had the anaerobic part of microbiota taking leading positions; the dominant species were those of bifidoflora. Nevertheless, we detected bifidobacteria only in half of the babies delivered by the caesarean section, while in the vaginally born babies this indicator reached 83%.
The B. longum subsp. infantis, B. animalis subsp. lactis, B. breve, B. bifidum bacteria peculiar to the first year of life produce anti-inflammatory effect and contribute to the development of the Th1 immune response. In adults, the predominant strains are B. longum subsp. longum, B. adolescentis, B. pseudocatenulatum; they promote the Th2 immune response and prevail in the gut microbiota in case of obesity [24]. Besides, the genome of B. longum subsp. infantis has a cluster that encodes the synthesis of enzymes (sialidase, fucosidase, N-acetyl-β-hexosaminidase and β-galactosidase) capable of breaking down oligosaccharides to monosaccharides. [25]. Throughout life, bifidoflora remains prevailing and sterile, while other representatives of the obligate microflora can trigger a disease under certain conditions [13]. Our study revealed that it is not the most useful bifidobacteria that dominate the microbiota of the C-section newborns; in particular, they are not able to break down lactose (B. animalis, B. dentium, B. ruminantium).

CONCLUSIONS
Our results confirm that surgical delivery inhibits the normal process of gut microbiota development to a certain extent. In the future, we plan to continue accumulating information about the state of microbiota in this category of children and expand the scope of our research to include premature babies in order to determine the criteria reflecting normal gut microbiota of healthy full-term newborns. This study allowed establishing, and in the future -replenishing the collection of lactobacilli and bifidobacteria isolated from healthy full-term newborns with the aim to identify bacteria candidates for probiotic strains. Timely adequate assessment of the newborn's microbiota composition against the key biomarkers will allow targeted prevention of the immediate and long-term consequences of perinatal pathology in full-term and premature infants by prescribing an adequate composition of probiotics.