Staphylococcus epidermidis in feedings and feces of preterm neonates

Staphylococcus epidermidis has emerged as the leading agent causing neonatal late-onset sepsis in preterm neonates; although the severity of the episodes caused by this species is often underestimated, it might exert relevant short- and long-term detrimental effects on neonatal outcomes. In this context, the objective of this study was to characterize a collection of S. epidermidis strains obtained from meconium and feces of preterm infants, and to assess the potential role of the enteral feeding tubes as potential reservoirs for this microorganism. A total of 26 preterm infants were enrolled in the study. Meconium and fecal samples were collected weekly during their first month of life (n = 92). Feeding samples were collected after their pass through the enteral feeding tubes (n = 84). S. epidermidis was present in the fecal samples of all the infants in, at least, one sampling time at concentrations ranging from 6.5 to 7.8 log10 CFU/g. Initially, 344 isolates were obtained and pulsed-field gel electrophoresis (PFGE) profiling allowed the reduction of the collection to 101 strains. Among them, multilocus sequence typing (MLST) profiling showed the presence of 32 different sequence types (ST). Globally, most of the STs to hospital-adapted high-risk clones and belonged to clonal complexes (CC) associated to the hospital environment, such as CC2. The virulence gene most commonly detected among the strains was altE. High resistance rates to macrolides and aminoglycosides were detected and 64% of the strains harboured the mecA gene, which was codified in SCCmec types. Our results indicates the existence of a complex and genetically diverse S. epidermidis population in the NICU environment. A better knowledge of S. epidermidis strains may help to devise strategies to avoid their conversion from symbiont to pathobiont microorganisms in the NICUs.


Introduction
Staphylococcus epidermidis is probably the most ubiquitous microorganism in human skin and mucosal surfaces [1,2]. This species seems to play a beneficial role in infants by inhibiting a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 virulent pathogens and educating the innate immune system [3]. Several studies have shown that S. epidermidis is the predominant bacteria in colostrum and milk from healthy women [4][5][6][7][8][9][10] and there is a mother-to-infant transmission through breastfeeding [11]; in fact, its presence is known to be a differential trait of the fecal microbiota of breast-fed infants when compared to that of formula-fed ones [10,[12][13][14][15][16], being already present in the first meconium obtained from both term and preterm breastfed neonates [17,18]; however, coagulase-negative staphylococci (CNS) have received a marginal attention regarding their role in the early colonization of the infant gut in contrast to lactobacilli, bifidobacteria and other gut-associated strict anaerobes.
The commensal or beneficial role of S. epidermidis in healthy hosts, including infants, may dramatically change in preterm neonates. Several factors in the neonatal intensive care unit (NICU), such as the widespread use of medical devices (catheters, enteral feeding tubes, mechanical ventilation. . .) where they can rapidly form thick biofilms, the selective pressure due to antibiotics or the immunocompromised status of the host, can explain the conversion of S. epidermidis from a symbiont member of the human microbiota to a very relevant opportunistic infectious agent [3]. Under such circumstances, this species has emerged as the predominant agent causing neonatal late-onset sepsis (LOS) in very low birth weight (VLBW) preterm neonates both in developed and developing countries [19][20][21][22][23]. Generally, LOS by CNS is associated to lower mortality rates when compared to other LOS agents [24]. However, beyond sepsis, S. epidermidis is associated with several neonatal morbidities, such as bronchopulmonary dysplasia, white matter injury, necrotizing enterocolitis and retinopathy of prematurity, which exert short-and long-term detrimental effects on neonatal outcomes [25,26].
In a previous work, we studied the microbiota of feces from preterm neonates by using both culture-dependent and -independent techniques [18]. The dominant species found were S. epidermidis, Staphylococcus aureus, Enterococcus faecalis, Enterococcus faecium, Serratia marcescens, Klebsiella pneumoniae and Escherichia coli. Subsequently, most of the dominant species were characterized and a high proportion of antibiotic-resistant high-risk clones was detected during the NICU admittance [27]. However, although S. epidermidis was the main species taken in account to the percentage of carrying infants, isolates from this species were initially excluded from the first analysis due to the high number of isolates and because they seemed to have a particularly wide genetic diversity. In this context, the objective of this study was to characterize a collection of S. epidermidis strains isolated from meconium and feces of preterm infants.

Patients and sampling
Twenty-six preterm infants (gestational age <32 weeks and/or weight <1,500 g) born between October 2009 and June 2010 in the University Hospital Doce de Octubre (Madrid, Spain) were included in the study. Preterm infants with malformations, metabolic diseases or severe conditions were excluded. The local ethic committee (CEIm Hospital Universitario 12 de Octubre) approved the study (reference 09/157), and parents signed informed consent. All infants were fed with human milk (pasteurized donor milk and/or their mother's own milk [MOM]) and, occasionally, with preterm formula. The fortifier FM 85 (Nestlé, Vevey, Switzerland) was used from approximately the fourth week of life. The high heterogeneity in feeding patterns prevented the formation of well-defined feeding groups.
First spontaneously evacuated meconium and fecal samples were collected weekly during the first month of life. Feeding samples were collected after their pass through the enteral feeding tubes from either the syringe or the extension set connected with the orogastric or nasogastric feeding tube. Extension sets were routinely replaced and discarded every 24 h, thus during this period different feed types could pass through the same tube. In contrast, the nasogastric feeding tubes were remained inserted for up to 4 days. Contamination of feeding tubes and the connector with the extension set was reviewed and confirmed in a previous work [28]. Preparation of the feedings previous to their pass through the enteral feeding system was as follows: (a) MOM was extracted using an electric pump and stored either refrigerated (5˚C) for maximum of 24 h, or frozen (-20˚C) up to 3 months; (b) donor milk was pasteurized (62.5˚C, 30 min) and stored frozen (-20˚C) after collection up to 3 months; and (c) commercial sterilized formula milk was ready-to-use in individual doses. All feedings were incubated for 10 to 15 min at 37˚C to 40˚C before their administration. After collection, the samples were immediately stored at −80˚C until analysis.

Culture-based analysis, bacterial identification and genotyping
Adequate dilutions of meconium, feces and feeding samples obtained during the NICU stay were spread onto Columbia Nalidixic Acid (CNA, BioMérieux) agar plates. Plates were aerobically incubated at 37˚C for up to 48 h. After bacterial counting, at least one representative of each colony morphology were analyzed by optical microscopy and those isolates compatible with staphylococcal morphology were identified by species-specific PCR based on the dnaJ genes with primers J-StGen (5'-TGGCCAAAAGAGACTATTATGA-3'), J-StEpi (5'-CCACC AAAGCCTTGACTT-3') [10]. Bacterial identification was also confirmed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry in a Vitek-MS instrument (BioMérieux).
In order to avoid repeated isolates, DNA from the S. epidermidis isolated colonies was extracted following the protocol of Ruiz-Barba et al. [29] and submitted to randomly amplified polymorphic DNA PCR (RAPD-PCR) using the primer OPL5 (5'-ACGCAGGCAC-3') as described by Rodas et al. [30]. At least a representative isolate from each RAPD profile per sample was selected to perform PFGE in a CHEF DR II apparatus (Bio-Rad, Birmingham, UK). SmaI-digested fragments were separated with the electrophoresis conditions of 5 to 15 s for 10 h and 15 to 60 s for 13 h. A dendrogram analysis of PFGE profiles was performed using the UPGMA method based on Dice similarity index and the INFO QUEST software (BioLine). Finally, the S. epidermidis MLST scheme was applied for PFGE-unrelated strains of S. epidermidis (http://www.mlst.net). Genetic diversity was represented by the Minimum Spanning Tree algorithm using the Phyloviz software (http://www.phyloviz.net).

Screening for potential virulence and antibiotic resistant determinants
Assessment of genes encoding virulence factors was carried out with the same set of strains that were submitted to MLST analysis. The tested genes included icaD, fbe and altE (which encode proteins involved in adhesion and biofilm production) [31] and mecA (resistance to methicillin) [32]. The staphylococcal cassette chromosome mec (SCCmec) was determined by using a typing procedure involving a PCR amplification (ccrB genes) followed by restriction fragment length polymorphism (RFLP) analysis using endonucleases HinfI and BsmI [32]. SCCmec elements are classified into types by a hierarchical system, considering types I, II and III have a hospital-acquired origin, whereas type IV is associated to the community.

Antimicrobial susceptibility testing
All the S. epidermidis strains submitted to MLST analysis were also tested for antibiotic susceptibility testing. Susceptibility to penicillin, oxacillin, gentamicin, tobramycin, levofloxacin, erythromycin, clindamycin, linezolid, daptomycin, teicoplanin, vancomycin, phosphomycin, fusidic acid, mupirocin, rifampicin and cotrimoxazol was assessed in a MicroScan WalkAway (Beckman Coulter) compact equipment. The breakpoints were selected according to the guidelines of the Clinical and Laboratory Standards Institute [33].

Statistical analysis
The statistical analysis was performed using R 2.15.3 (R-project, http://www.r-project.org). Data not normally distributed were represented as medians and interquartile ranges (Q1 and Q3); on the other hand, normally distributed data were expressed as means and 95% confidence interval (95% CI). The Kruskal-Wallis test was used to evaluate the differences between sampling times in non-normal data. Fisher's exact test was used to compare proportions. In all cases, P values of <0.05 were considered to be significant.

Characteristics of premature infants
Demographic and clinical data of the infants are provided in Table 1. Mean gestational age and birth weight were 27.7 weeks and 1,167 g, respectively. Most of them (92%) received antibiotic treatment during their hospital stay. Some of them (42%) received a combination of ampicillin and gentamicin as a prophylactic antibiotherapy while these and other antimicrobials (amphotericin B, amikacin, cloxacillin, erythromycin, fluconazole, fluorocytosine, meropenem, micafungin, teicoplanin, vancomycin) were occasionally used against a confirmed or suspected infection. Approximately half of the infants (54%) were born by Cesarean section and, consequently, their mothers received prophylactic or presurgical antibiotics per established clinical protocol. Respiratory disorders were common among our population since 21 of the newborns required continuous positive airway pressure (CPAP), 20 of them needed oxygen therapy and 16 of them required the use of mechanical ventilation, with median durations of 9, 32 and 8.5 days, respectively. Nearly all infants required parenteral nutrition at birth for a median duration of 7.5 days while all of them were fed with enteral feeding tubes for a mean period of 59 days. Median values for NICU and hospital stays reached 42 and 64 days, respectively. 27% of the infants suffered at least one sepsis episode during their NICU admittance. S. epidermidis was identified in the 71% of the sepsis episodes and affected to 19% of the infants, either alone or together with other microorganisms.

S. epidermidis from feces and feeding systems
The microbial culture of fecal samples yielded S. epidermidis colonies in all infants, although not in all samples. S. epidermidis was frequently detected in the meconium samples (38%), although its presence notably increased in feces collected at day 7 after birth (89%). Subsequently, its fecal presence decreased progressively until day 28, when only 27% of the infants harbored it (Table 2).
In the positive samples, the S. epidermidis counts varied between 6.5 and 7.8 log 10 CFU/g over the studied period. In order to evaluate the effect of antibiotherapy specific to Gram-positive bacteria on the S. epidermidis counts over time, samples were subdivided in those that were collected after an antibiotic treatment or not. In this case, S. epidermidis counts varied from 5.0 to 7.7 log 10 CFU/g (p.value 0.000) and 6.4 and 8.0 in samples collected when antibiotics were administered in the previous days or not respectively (Table 2). An increase in the concentration values from meconium to day 7 feces was observed; then, the concentrations decreased slowly and rose again at the end of the first month of life of the infants. When fecal bacterial counts were related to demographic and clinical data, S. epidermidis colonization tended to be higher in those infants with longer NICU stay, requiring more days of CPAP or mechanical ventilation, and born by Cesarean section, although the values did not reach a statistically-significant level. Interestingly, the S. epidermidis bacterial counts were slightly lower in those infants suffering sepsis caused by S. epidermidis, however is important to remark that these infants were treated with antibiotics (combinations of Teicoplanin with Vancomycin, Eritromycin, Gentamicin, Meropenem and/or Amikacin) at the moment of sample collection (Table 3).
In relation to the feeding samples after their pass through the enteral feeding system, S. epidermidis was present in 84% of the MOM samples while the detection frequency was significantly lower in donor milk and formula samples (42% and 14%, respectively). S. epidermidis counts were also considerably higher in MOM samples (Table 4).

Genetic diversity
Initially, 344 S. epidermidis isolates were obtained from 92 fecal and 84 feeding samples. After RAPD analysis, one representative of each RAPD profile per sample was selected for further analysis and, subsequently, 173 isolates were submitted to PFGE analysis. PFGE profiling allowed the selection of 101 strains, including one representative of each different PFGE profile per infant. This final collection was genetically characterized using the MLST scheme.
The 101 strains were distributed in 32 different sequence types (ST). The STs more frequently isolated from feces were ST69 (8 infants), ST425 (6 infants), ST59 (5 infants), and STs 2 and 51 (4 infants each), while those more widespread among the feeding samples were ST425 (9 samples), and STs 2, 32, 59 and 69 (5 samples each). Globally, the dominant STs in fecal samples were also majoritarian in the feeding ones. According to MLST profiling, the highest genetic diversity was related to MOM samples (30 different STs from 169 isolates), whereas in the fecal samples 12 different STs were identified from 103 isolates; in addition, up to 38% of the MOM samples contained 3 different ST strains while this percentage was lower (23%) in the fecal samples (Fig 1). MLST typing also revealed the transference or sharing of strains, such Table 4. S. epidermidis isolated from the feeding samples after their pass through the enteral feeding tubes. Counts expressed as log 10 CFU/g.  as ST59 in infants 1 and 3, ST35 in infants 2 and 11, ST69 in infants 5, 10, 13 and 14, ST32 in infants 4 and 14, or ST278 in infants 19 and 24, between fecal and feeding samples (Fig 1). Globally, 75% of STs detected in this study belonged to hospital-adapted high-risk clones although this percentage decreases to 53% when only milk samples are considered (Fig 1). Infants suffering a bacteremia episode caused by S. epidermidis were colonized by, at least, one strain of S. epidermidis belonged to high risk clones (Fig 1), with the exception of infant 24. From the 32 STs detected 13 belonged to clonal complexes (CC) associated to the hospital environment, as CC2, CC5 and CC6 (Fig 2). However, 7 STs had not been previously described; the new STs 589-591, 597 and 598 were assigned to five of them. The remaining two novel STs presented undescribed mutations on some of the studied alleles and their code assignment is still pending. The non-previously reported STs were mostly from MOM and only two of them (ST598 and STNew2), were also detected in fecal samples.

Presence of potential virulence genes and antimicrobial susceptibility
The results of the screening for potential virulence genes and antimicrobial susceptibility are presented in Fig 3. The genes most commonly detected in the isolated strains were altE (98%), involved in adhesion and biofilm production, and mecA (55%), implicated in methicillin resistance. Type III (28.4%) and IV (71.6%) SCCmec were identified in 16 and 29 isolates from fecal samples and 14 and 19 isolates from the enteral feeding system respectively.
Regarding antibiotic resistance, most of the S. epidermidis strains exhibited resistance to macrolides (70% to erythromycin and 56% to clindamycin), tetracycline (72%), methicillin (64%), aminoglycosides (62% for amikacin and tobramycin and 54% for gentamicin), and teicoplanin (22%) whereas almost all remained susceptible to vancomycin, tigecycline, daptomycin, quinolones and linezolid. A single isolate grouped on the ST2 lineage was resistant to all tested antibiotic including linezolid, daptomycin, fosfomycin, and cotrimoxazol. Globally, the strains with the highest virulence potential were those belonging to the STs 280, 2, 425, 190, 51 and 69, which were also resistant to oxacillin and clindamycin, and showed different resistance levels to gentamycin, tobramycin and mupirocin (Fig 3).

Discussion
In this study we use phenotyping and genotyping techniques to determine the diversity and virulence of S. epidermidis isolates obtained from fecal samples of preterm infants during their hospital stay and from the different types of feeds (MOM, donor milk and adapted formula) collected after their pass through the enteral feeding tubes.
S. epidermidis was relatively frequent in meconium samples (~40%), but the detection frequency notably increased (~90%) in the fecal samples collected 7 days after birth. A pioneering study using a culture-dependent approach revealed that S. epidermidis was often present in aseptically-collected meconium from term neonates and suggested that gut colonization may start in utero during late pregnancy under physiological conditions [10]. Although the issue is still controversial and the subject of an intense scientific debate [34,35], the results of subsequent studies that applied culture-dependent techniques to study meconium microbiota were similar, both in term [36][37][38][39] and preterm neonates [18]. The decrease in the fecal CNS populations after the first weeks of life has been previously reported [40][41][42][43][44].
In relation to the feeding samples, MOM contain their own microbiota while donor milk and preterm formula are usually sterile before their administration through the enteral feeding tubes. Studies carried in the last years have shown that human milk contains a sitespecific microbiota and represents a continuous supply of commensal and potentially probiotic bacteria to the infant gut [5]. Most culture-dependent and-independent studies on the human milk microbiota have found that CNS, and particularly S. epidermidis, is the dominant species in this biological fluid [5,[8][9][10][45][46][47][48]. Once in the infant gut, these bacteria drive the assembly of a healthy infant gut microbiota [49] and may play several functions, contributing to the infant metabolism, protection against infections, immunomodulation or neuromodulation [5].
The different milk types were subsequently administered through enteral feeding tubes which are well-suited devices for the growth and enrichment of high risk hospital-associated clones, which subsequently act as reservoirs of such clones [28,[50][51][52]. In preterm neonates, gastric pH values are closer to neutrality while the pylorus is more relaxed than in adults; this allows the entry of gut bacteria to the gastric compartment. Later, the widespread NICUs practice of aspirating and measuring the residual gastric content before administrating the next feeding leads to the contamination of the upper parts of the enteral feeding systems, including the external parts, with the high risk clones that characterize the preterm gut microbiota. Since feeding tubes are usually placed in the digestive tract of neonates for 24-72 h and are kept at a temperature optimal for the growth of many facultative anaerobic bacteria (CNS, enterobacteria. . .), such microorganisms have enough time to form thick mix biofilms in the internal surface of the tubes [28]. This may explain the presence of S. epidermidis in a percentage of donor milk and infant formula samples after their pass through such feeding systems; however, other sources, including healthcare workers and family members, can not be discarded and such possibility should be addressed in future studies.
The results of this study indicate that S. epidermidis population was genetically more diverse among MOM samples than among the fecal samples; this fact may reflect the coexistence of strains from both the human milk microbiota and the reservoir (enteral feeding tubes) when MOM passes through the enteral feeding system. Other contamination sources, such as endotracheal intubation or the hands of healthcare workers cannot be discarded. Later, the NICU environment may exert a role in selecting only those human milk strains which genome flexibility allow them to colonize the gastrointestinal tract of preterm neonates during their stay in the NICUs; as a result, the diversity of S. epidermidis strains is narrower in the feces of the infants. It must be highlighted that S. epidermidis is a bacterial species suited with an extraordinary genetic flexibility, and can employ a multitude of mechanisms to become adapted to the changing environment [3,53]. A study comparing 200 S. epidermidis strains isolated from women with lactational mastitis with 105 isolates from milk of healthy women showed that the number of strains that contained the biofilm-related icaD gene and that showed resistance to oxacillin, erythromycin, clindamycin and mupirocin was significantly higher among the strains isolated from the mastitis milk [31]. However, the same study also revealed that potentially pathogenic strains could also be isolated from some samples from healthy women. Another study confirmed that human milk from healthy women may be a reservoir of strains containing potential virulence factor and displaying (multi)resistance against clinically-relevant antibiotics [54]. In this context, most of the clones found in this work were precisely those enriched in virulence factors and antibiotic resistances (ST2, ST51 and ST425), showing higher adaptation to the NICU environment and, therefore, an enhanced capacity for persistence and dissemination. More specifically, a high proportion of the S. epidermidis strains exhibited resistance to ß-lactams, macrolides and aminoglycosides. In addition, almost all the methicillin-resistant isolates were also biofilm producers and predominantly carried cassettes type III and IV, which is in agreement with previous reports [3,20,[55][56][57]. Both antibiotic (multi)resistance and biofilm production capacity are common characteristics in strains (mainly CC2 ones) adapted to hospital environments [58,59].
As previously stated [3], S. epidermidis strains may play quite different roles (from "guarding angel to pathogenic devil") in the neonatal setting. A better knowledge of the predominant S. epidermidis clones (virulence factors, mechanism of biofilm formation in medical devices, mechanisms of immune-evasion, antibiotic resistances, specific sensitivity to phages, synergism or antagonism with other members of the preterm gut microbiota, CRISP systems. . .) that are particularly prevalent in NICUs is required in order to improve their control and, as a consequence, preterm health outcomes at the short and long terms. Microbiome-based strategies, including the restoration of the "safe" CNS population present in MOM when a neonate must be fed with pasteurized donor milk or formula have already been suggested [60]. CNS selected on the basis of a strain-by-strain rigorous safety assessment can be particularly useful to reduce the acquisition of undesired pathogens by infants, including preterm neonates, exposed to NICUs environments. It has been proposed that S. epidermidis and other CNS may have a probiotic function by preventing colonization of the host by more severe pathogens, such as S. aureus [2]. In fact, a future strategy to eradicate S. aureus from the mucosal surfaces has already been postulated based on their in vitro inhibition by S. epidermidis [61,62] A better knowledge of NICU-associated S. epidermidis strains may help to devise strategies to avoid their conversion from symbiont to pathobiont microorganisms.