Intestinal dysbiosis in children with short bowel syndrome is associated with impaired outcome

Background

Pediatric intestinal failure (IF) has been defined as the inability of the gastrointestinal
tract to sustain adequate growth, hydration, and electrolyte homeostasis in children
without parenteral nutrition (PN). Short bowel syndrome (SBS) is the most common cause
of pediatric IF. The condition is caused by massive small bowel resections due to
necrotizing enterocolitis (NEC) or volvulus and congenital malformations such as gastroschisis
and jejunal atresia. Neonatal SBS is a disease with a high morbidity and mortality
1]. The medical management of SBS aims to establish full enteral/oral feedings and weaning
from PN.

The intestinal microbiota seems to be a major factor in determining the successful
clinical outcome in SBS defined as independence of PN treatment and intestinal adaptation.
Alterations in the microbiota can result in serious complications such as small bowel
bacterial overgrowth (SBBO) and intestinal mucosal inflammation that may prevent weaning
from PN by compromising intestinal absorptive functions 2],3]. Luminal/oral antibiotic long-term treatment has been recommended for SBBO in children
4],5]. Most cases of SBS occur in neonates during a period when the sterile intestines
normally are colonized by bacteria, reaching a microbial profile characteristic of
the adult gastrointestinal tract around 2 to 4 years of age 6]. A disruption in the balanced intestinal microbial community, that is, dysbiosis,
with an increased relative abundance of facultative anaerobic Enterobacteriaceae in the large bowel is seen in inflammatory bowel disease (IBD) in mouse models, in
humans with Crohn’s disease, and in NEC in preterm infants 7],8].

To date, there are no reported mappings of the intestinal microbiota in children with
SBS. Here, we present the first report on the microbial profile in children with SBS
by using 16S rRNA gene sequencing on the Illumina MiSeq platform.

Methods

Patients

This study was approved by the regional committee on medical research ethics in Uppsala
(Dnr2012/002). Informed written consent for sample collection and subsequent analyses
was obtained from the parents. Characteristics of the study group and corresponding
healthy siblings are demonstrated in Table 1. The study includes 11 children between the age of 1.5 to 7 years diagnosed with
IF/SBS in the neonatal period, of whom two are from a set of triplets (2A and 3A)
(Table 1). All children except one were prematurely born. Child 1A, 8A, and 13A underwent
bowel lengthening procedure with serial transverse enteroplasty (STEP) 9]. Five children were not weaned from PN at the time of the study. Seven healthy siblings
served as controls. Children on PN had an oral and/or enteral intake of lactose-free
hydrolysed protein formula and an age-appropriate intake of solid foods with reduction
in disaccharide content according to Table 1.

Table 1. Characteristics of the study group and corresponding healthy siblings

Data collection and statistical analysis

Fecal samples were collected and stored at ?80°C until analysis. DNA was extracted
from each fecal sample with ultra-clean fecal DNA isolation kit (MoBio, Naxo Ltd,
Tartu, Estonia) according to the manufacturer’s instructions.

Sequencing libraries were prepared by amplifying the V3-V4 region of the 16S rRNA
gene using the 341f-805r primers, described by Hugerth et al. 10]. After the initial amplification, a second PCR was performed to attach Illumina adapters
as well as barcodes that allows for multiplexing. Samples were sequenced using the
IlluminaMiSeq, producing in total 10,136,440 2?×?300 base pair reads with an average
of 307,165 reads per sample. Primer sequences were trimmed away, and the paired-end
reads produced by the sequencing instrument were merged using SeqPrep version 1.1
(https://github.com/jstjohn/SeqPrep) with default parameters and thereafter processed with the QIIME 1.8.0 pipeline (Quantitative
Insight into Microbial Ecology) 11]. Merged reads were randomly subsampled to an even depth of 151,610 reads per sample,
which was the minimum number of reads among the samples. Using the UCLUST 12] algorithm built into the QIIME pipeline, sequences were clustered at 97% identity
against the Greengenes reference database producing 4,216 OTUs (operational taxonomic
units) 11]. For each sample; number of non-singleton OTU as well as most dominant OTU, with
respective description, is presented in the supplemental data (Additional file 1). Shannon indexes for diversity were calculated for SBS children on and off PN and
tested for significance with Wilcoxon rank-sum test. Using the QIIME pipeline, unweighted
UniFrac distances were produced and used for investigation of beta diversity through
plotting PCA coordinates. Details on 16S rRNA gene primers, amplification conditions,
and sample barcodes are shown in supplemental data (Additional file 2).

Results

Figure 1 shows that Shannon diversity index is significantly reduced in children with SBS
still on PN compared to children weaned from PN. None of the children on PN had remaining
ICV.

Figure 1. Shannon diversity index in children with SBS still on PN compared to children weaned
from PN.

In children still on PN, four out of five (1A, 3A, 8A, and 9A) were examined for several
episodes of suspected SBBO, also at the time of fecal sampling (Table 1). They were treated with oral metronidazole, trimethoprim-sulfamethoxazole, gentamicin,
or amoxicillin-clavulanic acid. In these patients, Enterobacteriacae was the most relative abundant taxonomic family and totally dominated the microbial
community in these children (Figure 2). The remaining patient in this group (12A), still on PN and with a reduced Shannon
diversity index, showed a relative abundant dominance of Lactobacillaceae followed by Enterobacteriacae. Altogether, a high relative abundance of Enterobacteriacae was associated with SBS in 6 out of 11 patients (1A, 3A, 8A, 9A, 11A, and 12A). In
the remaining five SBS patients, all off PN (2A, 4A, 13A, 16A, and 18A), there was
a more diverse microbiota composition and a more uniform distribution of taxonomic
families. However, none except one (2A) of the SBS children reached Shannon diversity
indexes at the same levels as the controls (Table 1). In one of the children still on PN (1A), upper and lower endoscopy with biopsies
revealed macroscopic and histopathologic acute inflammation in the stomach, duodenum,
small bowel, and proximal colon. In child 3A, also on PN, upper endoscopy with biopsies
demonstrated small bowel villous atrophy.

Figure 2. Microbial communities in children with SBS on PN (1A, 3A, 8A, 9A, 12A), SBS children
weaned from PN (2A, 4A, 11A, 13A, 16A, 18A), and siblings (2C1, 2C2, 11C1, 11C2, 12C,
13C1, 13C2). The figure is showing the relative abundance of the 19 most common taxonomic
families that accounts for at least 84% of the abundance in all samples.

In Figure 3, the Shannon diversity indexes and, in Figure 4, the unweighted UniFrac distances in children with SBS on PN (1A, 3A, 8A, 9A, 12A),
SBS children weaned from PN (2A, 4A, 11A, 13A, 16A, 18A), and siblings (2C1, 2C2,
11C1, 11C2, 12C, 13C1, 13C2) are compared.

Figure 3. Shannon diversity index in children with SBS on PN (1A, 3A, 8A, 9A, 12A), SBS children
weaned from PN (2A, 4A, 11A, 13A, 16A, 18A), and siblings (2C1, 2C2, 11C1, 11C2, 12C,
13C1, 13C2).

Figure 4. PCoA plot describing unweighted UniFrac distance between samples. Pairwise distances
between all samples are projected onto a two-dimensional space where the axis PC1
describes the highest degree of variation. Samples that are clustered closely together
are thus considered to share a larger proportion of the phylogenetic tree in comparison
to samples that are more separated.

We had the unique opportunity to study triplets representing all three groups. Child
2A and child 3A were male triplets born at 23 weeks of gestation. Both boys suffered
from NEC in the neonatal period resulting in small bowel resections (Table 1). The third triplet boy (2C2) remained healthy. In child 2A, only 2 cm of the small
bowel was resected, however he developed IF after extensive NEC and became dependent
on PN. During PN treatment, he had no signs of SBBO. He was weaned to full oral feeding
and without antibiotics 3 months before the time of fecal sampling. His intestinal
bacterial diversity was similar to his healthy brother (SDI 4, 67 and 4, 97, respectively).

Child 12A was treated with antibiotics only during the first 2 weeks postnatally,
had no signs of SBBO, and weaning from PN advanced but slowly. Her fecal bacterial
diversity showed Lactobacillacae as the most relative abundant taxonomic family in accordance to our previous findings
(Figure 2) 13]. We could detect Clostridium difficile in two out of ten SBS patients (patient 2A and 11A) and in very low relative abundance
(data not shown).

Conclusions

The tendency for SBBO and bowel inflammation to delay or prevent weaning from PN in
these children with SBS seems to be related to microbial dysbiosis in the intestinal
tract. This finding is in accordance with a previous study demonstrating that PN administration
was independently associated with SBBO 14]. Influence of PN on the change in profile due to starvation of the microbiome is
unlikely since malabsorption of oral/enteral nutrition is the major problem in SBS.
In general, the observed changes in the microbiota in SBS children are most likely
both at cause and a consequence of the disease status of the child. The limitation
of the study is the small study group, and confounding factors that might influence
the results are age, intestinal length, and antibiotic treatment. However, the cohort
in the present study represents children with SBS in the clinical practice.

In our center, we treat SBBO with oral antibiotics as recommended by other centers
4]-6]. However, it is most likely that antibiotics will further contribute to dysbiosis
in these children. In children with SBS, normal colonization is disrupted due to early
and frequent antibiotic use and decreased bacterial diversity allows potential pathogenic
bacteria to expand. Antibiotics have been suggested to lower colonization resistance
against Enterobacteriacae such as Escherichia coli and Salmonella enterica, by increasing the inflammatory tone of the intestinal mucosa 15]. Most frequently, probiotics are used to modify the intestinal microbiota in SBS;
however, there are conflicting findings and reports of probiotic-associated septicemia
16].

The overall decreased bacterial diversity in our children with SBS is consistent with
intestinal dysbiosis in IBD patients, infants with NEC, and has also been described
in a piglet model of SBS 7],8],17],18]. In addition, children with recurrent C. difficile-associated diarrhea show a decreased fecal diversity with a reduction of Bacteriodetes and Firmicutes19]. In these children, fecal microbiota transplantation (FMT) has a success rate greater
than 90%. Such treatment has also been successfully used as a complement to treat
IBD 19],20]. Consequently, FMT could prove to be a treatment alternative in carefully selected
cases of SBS with dysbiosis. However, since children with SBS often are vulnerable
due to their initial health status, the difficulties and risks of FMT must be considered.
Although the incidence of severe side-effects is rare, one such risk is contracting
illness from the donor where asymptomatic microorganisms that cause no problems in
a healthy donor may cause a reaction in the recipient. In addition, mass arrival of
a new microbiota may also trigger autoimmune illness as well as bacteria and septic
shock. Extra care should also be taken with FMT if the patient has any sign of immunodeficiency
21].

This is the first report describing the intestinal microbial profile in children with
SBS using next-generation sequencing. We observed a pronounced microbial dysbiosis
in children with SBS still on PN compared to children weaned from PN with an increased
relative abundance of proteobacteria, most of whom has been long-term treated with
antibiotics. Our findings indicate that intestinal dysbiosis in children with SBS
is associated with impaired outcome with prolonged PN dependency. Future studies need
to find out new strategies to treat intestinal dysbiosis in these children.