Crohn’s disease (CD) is an inflammatory disease of the gastrointestinal tract, with a rising incidence worldwide, particularly in children. CD is thought to arise due to an immune response to environmental factors. The role of bacteria in CD has recently been highlighted, and here, we examine the prevalence of two bacterial species, enterotoxigenic Bacteroides fragilis (ETBF) and Fusobacterium nucleatum, implicated in gastrointestinal pathologies, in a pediatric CD cohort. Stool samples from 30 children with treatment-naïve CD and 30 age- and sex-matched controls were collected, and DNA was extracted. Quantitative PCR was used to determine the levels of ETBF and F. nucleatum in stool samples. Bacterial positivity and relative abundance were assessed between cases and controls and in relation to disease severity. No associations were found between colonization with ETBF and CD, or between colonization with either ETBF or F. nucleatum and disease severity or presence of C. concisus. However, a strong association was observed between positivity for F. nucleatum in the stool samples and the occurrence of CD in patients (25/30) as compared to controls (8/30) (P=0.003). F. nucleatum is more prevalent in the stool samples of pediatric CD patients, compared to healthy controls, and may have potential use as a biomarker of pediatric CD.
1. Background
Inflammatory bowel diseases (IBDs) are characterized by chronic inflammation in the gastrointestinal tract (GIT), and the two most commonly described forms of IBDs are Crohn’s disease (CD) and ulcerative colitis (UC). CD is characterized by inflammatory lesions, which may be transmural, in any location in the GIT, with granulomas being a common feature. Although the etiology of CD is, as yet, unknown, there is evidence that dysregulated immune response to environmental triggers may lead to CD onset in genetically predisposed individuals.
CD affects around 3.2/1000 people in Europe and North America, and the incidence is rising, particularly in children. While previously the incidence of CD in children resident in nonindustrialized countries has been low, current evidence would suggest that, with industrialization, the incidence of CD is increasing [1].
CD is associated with a slightly lower life expectancy and considerable morbidity. Patients with long-standing CD located in the colon have a similar increased risk of colorectal cancer development as do those with UC [2, 3], and as a result, there has been intense interest in identifying potential etiological agents of the disease. The contribution of microbial dysbiosis to inflammation in the gut has recently been described, and several studies have identified changes in bacterial communities and particular species that are associated with CD. Decreased abundance of beneficial bacterial species from the Lachnospiraceae and Ruminococcaceae families and increased abundance of potential pathobionts, such as Campylobacter spp. and Fusobacterium nucleatum, have been identified using metagenomic and targeted studies of CD patients [4–7]. Both C. concisus and F. nucleatum are oral pathogens with epithelial cell adherence and invasive potential [4, 8–10]; certain strains of C. concisus produce exotoxin 9 and/or zonula occludens toxin (ZOT) [11]. Enterotoxigenic Bacteroides fragilis (ETBF) produces a zinc metalloprotease toxin called the Bacteroides fragilis toxin (Bft) that has been shown to induce colonic inflammation in mice [12, 13] and is associated with diarrheal disease and the development of colorectal cancer [14–16].
In this study, we investigate the abundance of ETBF and F. nucleatum in a cohort of pediatric Crohn’s disease patients and age-matched controls. We also examined the association of these bacteria with the presence of previously determined C. concisus and its toxins, as well as with patient characteristics.
2. Materials and Methods2.1. Study Subjects
Thirty children with Crohn’s disease, diagnosed at Sydney Children’s Hospital (Randwick, Australia), were recruited, in addition to 30 healthy controls. Stool samples were collected from the patients prior to treatment, and exclusion criteria included antibiotic or anti-inflammatory treatment in the four weeks prior to sample collection. This patient cohort has previously been described [6], and the levels of C. concisus and ZOT have previously been reported [17] in this cohort. Informed consent was obtained from all participants (or their parent/guardian), and the study was approved by the Research Ethics Committee of the University of New South Wales and the South East Sydney Area Health Service-Eastern Section, Sydney (Ethics nos. 03/163, 03/165, and 06/164).
2.2. Reference Strains
ETBF strain VPI 13784 [18] was generously supplied by Professor Cynthia Sears, Baltimore, USA, and was used as a reference strain in this study. This was cultured anaerobically on sheep blood agar (Fort Richard Laboratories). F. nucleatum was sourced from the American Type Culture Collection (ATCC) strain 25586. DNA was extracted from colonies recovered from the plates using the DNeasy Blood & Tissue Mini Kit (Qiagen, Hilden, Germany), as per the manufacturer’s instructions. DNA extraction included digestion with Proteinase K for 3 hours at 56°C. The samples were stored at −20°C.
2.3. Primer Design
TaqMan primer/probe sets were designed to amplify regions of the Bft gene of ETBF and the nusG gene of F. nucleatum. Primer and probe sequences are shown in Table 1.
Primers and probes used for quantitative PCR.
Gene
Sequence 5′ → 3′
Reference
F. nucleatum nusG
[19]
Forward primer
CAACCATTACTTTAACTCTACCATGTTCA
Reverse primer
GTTGACTTTACAGAAGGAGATTATGTAAAAATC
Probe
TCAGCAACTTGTCCTTCTTGATCTTTAAATGAACC
Bft
[20]
Forward primer
GGATAAGCGTACTAAAATACAGCTGGAT
Reverse primer
CTGCGAACTCATCTCCCAGTATAAA
Probe
CAGACGGACATTCTC
2.4. Quantitation of Genomic DNA in Bacterial Reference Strain Samples
DNA was extracted from fecal samples from CD cases and controls using the QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany), as per the manufacturer’s instructions. Levels of each bacterium were measured from genomic DNA samples using TaqMan probes on a LightCycler® 480 thermocycler (Roche Diagnostics, Indianapolis, IN, USA), as detailed in the Methods section of the study of Purcell et al. [22]. In brief, each reaction consisted of 25–35 ng of genomic DNA, 5 µl of TaqMan Fast Advanced Master Mix (Applied Biosystems), and 0.5 µl of TaqMan primer/probe (Thermo Fisher Scientific) in a 10 µl reaction. Thermal cycling conditions were as follows: 1 cycle at 95°C for 10 min, followed by 50 cycles at 95°C for 10 sec and 60°C for 30 sec. All reactions were performed in triplicate.
2.6. Statistical Analysis
Fisher’s exact tests, with 95% confidence intervals, were used to test for associations between Crohn’s disease and the presence or absence of bacterial species. Pediatric CD activity index (PCDAI) values in the CD cohort were compared between patients tested negative and positive for each bacterial transcript, using t-tests, with Satterthwaite’s adjustment for unequal variances and linear models, including covariates for age and gender. P values for association of CD and PCDAI values were adjusted for multiple testing using the Benjamini and Yekutieli method [23]. All analyses were carried out using R version 3.2.1 (Vienna, Austria).
3. Results
This study used a cohort of 30 newly diagnosed pediatric CD patients, of whom 20 were male; the patients ranged in age from 3 to 17.1 years (median = 11.4 years). Thirty healthy controls were also recruited, half of whom were male. The disease severity in patients was measured using the pediatric Crohn’s disease activity index (PCDAI) and ranged from 7.5–70. Thirteen patients had mild disease (PCDAI < 30), while the remaining 17 patients had moderate-to-severe CD (PCDAI ≥ 30). All but one patient had evidence of colonic involvement, and as a result, we were unable to analyze bacterial abundance by disease location. Previous analysis of C. concisus and its toxins, exotoxin 9 and ZOT, showed that while the numbers of children tested positive for C. concisus were similar between cases and controls, the relative abundance of the bacteria and exotoxin 9 was higher in CD patients [6].
3.1. Association of Disease Severity with Presence or Absence of qPCR Products
Firstly, the distribution of PCDAI values was analyzed among the disease cohort using the Shapiro–Wilk test; there was no evidence that PCDAI was not normally distributed among CD patients in the cohort (P=0.646; Figure 1). No evidence of an association between CD severity, as measured by the PCDAI, and the presence or relative abundance of any bacterial transcripts or toxins was observed (all adjusted P values > 0.1; Table 2). This is similar to findings reported with C. concisus.
Histogram showing no evidence that disease severity, as measured by the pediatric Crohn’s disease activity index (PCDAI), is not normally distributed.
Association of PCDAI and transcripts in the gut.
Negative mean
Positive mean
Negative SE
Positive SE
P value
P value (adj.)
Cc
27.500
35.292
3.416
3.375
0.124
0.345
ZOT
29.375
35.318
4.324
3.505
0.301
0.749
ETBF
33.685
34.167
3.126
3.333
0.919
0.319
Fn
35.500
33.380
7.045
3.140
0.793
0.768
Cc, Campylobacter concisus; ZOT, zonula occludens toxin; ETBF, enterotoxigenic Bacteroides fragilis; Fn, Fusobacterium nucleatum; SE, standard error; adj., adjusted for age and gender.
3.2. Association of CD and Bacterial Transcripts in Fecal Samples
Quantitative PCR revealed that two patients and three controls were positive for ETBF, while 25 patients with CD were positive for F. nucleatum, compared with only eight controls. Analysis of associations between bacterial transcript positivity and the presence of CD, using Fisher’s exact test, found a strong association between F. nucleatum positivity in stool samples and the occurrence of CD. The P value was 0.00002, and when adjusted for age and gender, it was 0.0003. There was no significant difference or trend in the relative abundance of F. nucleatum between cases and controls.
4. Discussion
In our study, we have investigated the prevalence and abundance of specific bacterial species in a cohort of pediatric Crohn’s disease patients and matched controls. The study was conducted on fecal stool samples from newly diagnosed patients prior to treatment, and information on disease severity and the abundance of C. concisus and its toxins was known prior to the study being carried out. Microbial dysbiosis, involving a reduction in fecal bacterial diversity, has been widely reported in gastrointestinal diseases, including CD, ulcerative colitis, and colorectal cancer (CRC) [24–27], and there has been much focus on identifying particular bacterial species that may play a role in disease pathogenesis, as therapeutic targets and biomarkers of disease.
Bacteroides fragilis is a common gut commensal in adults, although the rate of carriage is not well characterized in children. Enterotoxigenic strains of B. fragilis (ETBF) secrete a proinflammatory metalloprotease toxin and have been associated with diarrheal disease in both children and adults. ETBF has also been proposed as a potential initiator of CRC via E-cadherin cleavage and NF-kB activation, as demonstrated in in vitro and animal studies and increased carriage of the bacterium in stool samples of CRC patients compared to controls [16, 28]. In contrast, the evidence linking ETBF to IBD is limited, although a recent publication has demonstrated increased carriage of ETBF in patients with UC, compared to controls [29]. In this study, we found no association between the presence or abundance of toxigenic strains of B. fragilis in the stool samples and the occurrence of CD in patients as compared to healthy controls.
Intestinal colonization with oral bacteria has recently been implicated in inflammatory bowel diseases [5], with toxin-producing Campylobacter species being associated with pediatric CD as compared to healthy controls in this cohort and others [6, 17, 30].
Fusobacterium nucleatum is another prevalent oral bacterium found in extraoral diseases and has the ability to adhere to and invade epithelial cells via its surface adhesin molecule, FadA. Strains of F. nucleatum isolated from IBD have been shown to be significantly more invasive than strains from healthy tissue or controls [4], and colonization of epithelial cells by F. nucleatum is associated with an inflammatory response [31–33]. Although causality has yet to be established, increased abundance of this bacterium has frequently been associated with CRC [19, 34, 35] and IBD, including CD, a known risk factor for CRC. Two separate microbiome studies of pediatric CD (n=19 patients in both studies) found enrichment of Fusobacterium sp. in fecal samples of patients compared to healthy controls. The study by Shaw et al. found that Fusobacterium sp. were more abundant in cases compared to controls and between nonresponders and responders to immunomodulatory treatment [36], while the study by Kaakoush et al. [17] found Fusobacterium sp. in 8/19 CD patients compared to 1/19 controls. Although these studies used a 16S rRNA sequencing approach that often does not allow for discrimination to the species level, the findings are similar to those of our direct bacterial species testing using qPCR in this study, where we found a strong association of F. nucleatum with CD as compared to control samples. F. nucleatum was detected in 83.3% of CD cases compared to 26.7% of controls, with similar abundance of the bacterium in samples from the two groups.
Testing for F. nucleatum in fecal samples, in combination with fecal immunochemical testing (FIT), improves detection of adenomas and CRC compared to FIT alone [37, 38], and our findings indicate that this oral pathogen may also prove to be a useful tool in noninvasive fecal testing for CD. A study by Dharmani et al. has reported that different strains of F. nucleatum vary in relation to their inflammatory and invasive potential [10]. This strain variance should be taken into account in future clinical studies.
Future directions would include validation of different strains of F. nucleatum as markers of disease in a larger independent cohort and testing the utility of F. nucleatum as a marker of disease activity. F. nucleatum has recently been linked to chemoresistance in colorectal cancer [39], and its influence on response to therapy may also have relevance in the treatment of CD. The strong association of the bacteria with CD patients compared to controls also provides direction for future functional studies on the pathogenesis and progression of CD using in vitro and in vivo models.
5. Conclusions
F. nucleatum is more prevalent in the stool samples of pediatric Crohn’s disease patients, compared to healthy controls, and may have potential use as a biomarker of pediatric CD.
Data Availability
The qPCR data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare no conflicts of interest.
NgS. C.ShiH. Y.HamidiN.Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies2017390101142769277810.1016/S0140-6736(17)32448-02-s2.0-85031499214SacharD. B.Cancer in Crohn’s disease: dispelling the myths199435111507150810.1136/gut.35.11.15072-s2.0-0027961105von RoonA. C.ReeseG.TeareJ.ConstantinidesV.DarziA. W.TekkisP. P.The risk of cancer in patients with Crohn’s disease200750683985510.1007/s10350-006-0848-z2-s2.0-34249873237StraussJ.KaplanG. G.BeckP. L.Invasive potential of gut mucosa-derived Fusobacterium nucleatum positively correlates with IBD status of the host20111791971197810.1002/ibd.216062-s2.0-79959474060GeversD.KugathasanS.DensonL. A.The treatment-naive microbiome in new-onset Crohn’s disease201415338239210.1016/j.chom.2014.02.0052-s2.0-84896092821KaakoushN. O.Castano-RodriguezN.DayA. S.LembergD. A.LeachS. T.MitchellH. M.Campylobacter concisus and exotoxin 9 levels in paediatric patients with Crohn’s disease and their association with the intestinal microbiota20146319910510.1099/jmm.0.067231-02-s2.0-84890629207KaakoushN. O.DayA. S.HuinaoK. D.Microbial dysbiosis in pediatric patients with Crohn’s disease201250103258326610.1128/jcm.01396-122-s2.0-84866463340ManS. M.KaakoushN. O.LeachS. T.Host attachment, invasion, and stimulation of proinflammatory cytokines by Campylobacter concisus and other non-Campylobacter jejuni Campylobacter species2010202121855186510.1086/6573162-s2.0-78650386765ManS. M.KaakoushN. O.MitchellH. M.The role of bacteria and pattern-recognition receptors in Crohn’s disease20118315216810.1038/nrgastro.2011.32-s2.0-79952445703DharmaniP.StraussJ.AmbroseC.Allen-VercoeE.ChadeeK.Fusobacterium nucleatum infection of colonic cells stimulates MUC2 mucin and tumor necrosis factor alpha20117972597260710.1128/iai.05118-112-s2.0-79959438740KaakoushN. O.ManS. M.LambS.The secretome of Campylobacter concisus201027771606161710.1111/j.1742-4658.2010.07587.x2-s2.0-77949579886RabizadehS.RheeK. J.WuS.Enterotoxigenic Bacteroides fragilis: a potential instigator of colitis200713121475148310.1002/ibd.202652-s2.0-38049023064RheeK. J.WuS.WuX.Induction of persistent colitis by a human commensal, enterotoxigenic Bacteroides fragilis, in wild-type C57BL/6 mice20097741708171810.1128/iai.00814-082-s2.0-63149120571SearsC. L.The toxins of Bacteroides fragilis200139111737174610.1016/s0041-0101(01)00160-x2-s2.0-0034862657BoleijA.HechenbleiknerE. M.GoodwinA. C.The Bacteroides fragilis toxin gene is prevalent in the colon mucosa of colorectal cancer patients201560220821510.1093/cid/ciu7872-s2.0-84924971514KeenanJ. I.AitchisonA.PurcellR. V.GreenleesR.PearsonJ. F.FrizelleF. A.Screening for enterotoxigenic Bacteroides fragilis in stool samples201640505310.1016/j.anaerobe.2016.05.0042-s2.0-84989900446KaakoushN. O.Castano-RodriguezN.DayA. S.LembergD. A.LeachS. T.MitchellH. M.Faecal levels of zonula occludens toxin in paediatric patients with Crohn’s disease and their association with the intestinal microbiota201564330330610.1099/jmm.0.0000052-s2.0-84923701028MoncriefJ. S.DuncanA. J.WrightR. L.BarrosoL. A.WilkinsT. D.Molecular characterization of the fragilysin pathogenicity islet of enterotoxigenic Bacteroides fragilis19986617351739FlanaganL.SchmidJ.EbertM.Fusobacterium nucleatum associates with stages of colorectal neoplasia development, colorectal cancer and disease outcome20143381381139010.1007/s10096-014-2081-32-s2.0-84903817580ChenL. A.Van MeerbekeS.AlbesianoE.Fecal detection of enterotoxigenic Bacteroides fragilis20153491871187710.1007/s10096-015-2425-72-s2.0-84940203176PurcellR. V.PearsonJ.FrizelleF. A.KeenanJ. I.Comparison of standard, quantitative and digital PCR in the detection of enterotoxigenic Bacteroides fragilis2016613455410.1038/srep345542-s2.0-84989954888PurcellR. V.PearsonJ.AitchisonA.DixonL.FrizelleF. A.KeenanJ. I.Colonization with enterotoxigenic Bacteroides fragilis is associated with early-stage colorectal neoplasia2017122e017160210.1371/journal.pone.01716022-s2.0-85011319812BenjaminiY.DraiD.ElmerG.KafkafiN.GolaniI.Controlling the false discovery rate in behavior genetics research20011251-227928410.1016/s0166-4328(01)00297-22-s2.0-0035829361HansenR.RussellR. K.ReiffC.Microbiota of de-novo pediatric IBD: increased Faecalibacterium prausnitzii and reduced bacterial diversity in Crohn’s but not in ulcerative colitis2012107121913192210.1038/ajg.2012.3352-s2.0-84870824919WalkerA. W.SandersonJ. D.ChurcherC.High-throughput clone library analysis of the mucosa-associated microbiota reveals dysbiosis and differences between inflamed and non-inflamed regions of the intestine in inflammatory bowel disease2011111710.1186/1471-2180-11-72-s2.0-78650973052AhnJ.SinhaR.PeiZ.Human gut microbiome and risk for colorectal cancer2013105241907191110.1093/jnci/djt3002-s2.0-84891473723HuipengW.LifengG.ChuangG.JiayingZ.YuankunC.The differences in colonic mucosal microbiota between normal individual and colon cancer patients by polymerase chain reaction-denaturing gradient gel electrophoresis201448213814410.1097/mcg.0b013e3182a267192-s2.0-84892941261ToprakN. U.YagciA.GulluogluB. M.A possible role of Bacteroides fragilis enterotoxin in the aetiology of colorectal cancer200612878278610.1111/j.1469-0691.2006.01494.x2-s2.0-33745593712ZamaniS.Hesam ShariatiS.ZaliM. R.“Detection of enterotoxigenic Bacteroides fragilis in patients with ulcerative colitis2017915310.1186/s13099-017-0202-02-s2.0-85029513085KugathasanS.DensonL. A.WaltersT. D.Prediction of complicated disease course for children newly diagnosed with Crohn’s disease: a multicentre inception cohort study2017389100801710171810.1016/S0140-6736(17)30317-32-s2.0-85014105950McCoyA. N.Araujo-PerezF.Azcarate-PerilA.YehJ. J.SandlerR. S.KekuT. O.Fusobacterium is associated with colorectal adenomas201381e5365310.1371/journal.pone.00536532-s2.0-84872387623HanY. W.ShiW.HuangG. T.Interactions between periodontal bacteria and human oral epithelial cells: Fusobacterium nucleatum adheres to and invades epithelial cells20006863140314610.1128/iai.68.6.3140-3146.20002-s2.0-0034079238ParkS. R.KimD. J.HanS. H.Diverse toll-like receptors mediate cytokine production by Fusobacterium nucleatum and Aggregatibacter actinomycetemcomitans in macrophages20148251914192010.1128/iai.01226-132-s2.0-84898868970CastellarinM.WarrenR. L.FreemanJ. D.Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma201222229930610.1101/gr.126516.1112-s2.0-84856534859KosticA. D.GeversD.PedamalluC. S.Genomic analysis identifies association of Fusobacterium with colorectal carcinoma201222229229810.1101/gr.126573.1112-s2.0-84863022950ShawK. A.BerthaM.HofmeklerT.Dysbiosis, inflammation, and response to treatment: a longitudinal study of pediatric subjects with newly diagnosed inflammatory bowel disease2016817510.1186/s13073-016-0331-y2-s2.0-84979498467WongS. H.KwongT. N.ChowT. C.Quantitation of faecal Fusobacterium improves faecal immunochemical test in detecting advanced colorectal neoplasia20166681441144810.1136/gutjnl-2016-3127662-s2.0-84994759860BaxterN. T.ZackularJ. P.ChenG. Y.SchlossP. D.Structure of the gut microbiome following colonization with human feces determines colonic tumor burden2014212010.1186/2049-2618-2-202-s2.0-84924040920YuT.GuoF.YuY.Fusobacterium nucleatum promotes chemoresistance to colorectal cancer by modulating autophagy201717054856310.1016/j.cell.2017.07.0082-s2.0-85026371130