Similar Fecal Microbiota Signatures in Patients With

Clinical Gastroenterology and Hepatology 2016;14:1602–1611

Similar Fecal Microbiota Signatures in Patients With Diarrhea-Predominant Irritable Bowel Syndrome and Patients With Depression Yixuan Liu,*,a Lu Zhang,*,a Xiaoqi Wang,‡,a Zhe Wang,‡ Jingjing Zhang,* Ronghuan Jiang,§ Xiangqun Wang,§ Kun Wang,* Zuojing Liu,* Zhiwei Xia,* Zhijie Xu,* Yong Nie,¶ Xianglin Lv,¶ Xiaolei Wu,¶ Huaiqiu Zhu,‡ and Liping Duan* *Department of Gastroenterology, Peking University Third Hospital, ‡Department of Biomedical Engineering, Center for Quantitative Biology, ¶Department of Energy and Resources Engineering, College of Engineering, Peking University, § Department of Psychiatry, Institute of Mental Health, Peking University, Beijing, China BACKGROUND & AIMS:

Patients with irritable bowel syndrome (IBS) often have psychiatric comorbidities. Alterations in the intestinal microbiota have been associated with IBS and depression, but it is not clear if there is a microbial relationship between these disorders. We studied the profiles of fecal microbiota samples from patients with IBS, depression, or comorbidities of IBS and depression; we determined the relationships among these profiles and clinical and pathophysiological features of these disorders.

METHODS:

We used 454 pyrosequencing to analyze fecal microbiota samples from 100 subjects (40 with diarrhea-predominant IBS [IBS-D], 15 with depression, 25 with comorbidities of IBS and depression, and 20 healthy individuals [controls]), recruited at Peking University. Abdominal and psychological symptoms were evaluated with validated questionnaires. Visceral sensitivity was evaluated using a barostat. Colonic mucosal inflammation was assayed by immunohistochemical analyses of sigmoid tissue biopsy specimens.

RESULTS:

Fecal microbiota signatures were similar between patients with IBS-D and depression in that they were less diverse than samples from controls and had similar abundances of alterations. They were characterized by high proportions of Bacteroides (type I), Prevotella (type II), or nondominant microbiota (type III). Most patients with IBS-D or depression had type I or type II profiles (IBS-D had 85% type I and type II profiles, depression had 80% type I and type II profiles). Colon tissues from patients with type I or type II profiles had higher levels of inflammatory markers than colon tissues from patients with type III profiles. The level of colon inflammation correlated with the severity of IBS symptoms.

CONCLUSIONS:

Patients with IBS-D and depression have similar alterations in fecal microbiota; these might be related to the pathogenesis of these disorders. We identified 3 microbial profiles in patients that could indicate different subtypes of IBS and depression or be used as diagnostic biomarkers.

Keywords: Microbe; Cytokine; Immune Response; Psychology.

rritable bowel syndrome (IBS) is the most prevalent functional gastrointestinal disorder, affecting 10% of the population.1 Comorbidity of IBS with psychological distress is common (range, 15.5%–39.5%).2,3 Different animal models developed to assess the pathophysiological mechanisms have found close correlations with dysregulation of the brain–gut axis,4 impaired mucosal barrier function,5 and alteration in gastrointestinal microbiota.6 Increasing evidence indicates variations in both mucosal and fecal microbiota in IBS.7–9 Consistent findings across studies appear to be a higher ratio of Firmicutes/Bacteroidetes,7,8 leading to lower diversity in

I

a

Authors share co-first authorship.

Abbreviations used in this paper: COMO, comorbidities of irritable bowel syndrome and depression; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, 4th edition; IBS, irritable bowel syndrome; MCP-1, monocyte chemotactic protein-1; MINI, Mini-International Neuropsychiatric Interview; MIP-1a, macrophage inflammatory protein-1a; PCA, principal component analysis; SDS, Self-Rating Depression Scale; SSS, Symptom Severity Scores. Most current article © 2016 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2016.05.033

Downloaded from ClinicalKey.com at Peking University Health Science Center November 17, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

November 2016

Shared Microbiota in IBS–D and Depression 1603

fecal microbiota of IBS patients.10 Ng et al8 reported that IBS patients had a high abundance of Bacteroides in mucosal microbiota, which was reduced after probiotic treatment. Another study showed that stool frequency was associated with the abundance of mucosa-associated Bifidobacteria and Lactobacilli.11 Differences in composition and diversity of mucosa and fecal microbiota might be substantial. Mucosal microbiota might play a more important role in the pathogenesis of IBS, whereas fecal microbiota is more convenient for biomarker identification.8,12,13 Patients with depression also reported microbiota dysbiosis.14 Microbiota might play an important role in supporting emotional well-being.15,16 It is speculated that microbiota dysbiosis–induced mucosal immunobarrier dysfunction may be an early step that contributes to the pathogenesis of depression and IBS.17,18 Immune cells such as mast cells, marcrophages, and CD3þ T cells, and their activators such as monocyte chemotactic protein-1 (MCP-1) and macrophage inflammatory protein-1a (MIP-1a),19–21 were activated by lipopolysaccharide or short-chain fatty acids, which are associated with dysbiosis,22–24 hinting that some intermediaries participated in these early steps. The aims of the present study were to identify the microbiota profiles of patients with IBS-D, depression, and their comorbidity, and to clarify the microbial relationship between IBS and depression. We further assessed possible correlations between microbiota variation and intestinal immunity, and associations between intestinal immune and clinical symptoms of patients.

Materials and Methods Subject Recruitment IBS-D patients (assessed according to the Rome III criteria) were recruited sequentially from the Outpatient Department of Gastroenterology of Peking University Third Hospital, depression patients (assessed according to the Mini-International Neuropsychiatric Interview [MINI] Diagnostic and Statistical Manual of Mental Disorders, 4th edition [DSM-IV]) were recruited from the Outpatient Department of the Institute of Mental Health of Peking University. Patients who met both Rome III and MINI DSM-IV criteria were recruited as having a comorbidity of the 2 disorders. Healthy volunteers were enrolled as controls. All subjects were evaluated by both a gastroenterologist and a psychologist. All subjects underwent colonoscopy or received a barium enema to rule out organic colonic diseases. Patients who had taken antibiotics, probiotics/prebiotics, and psychotropic medications during the previous 4 weeks were excluded. Other exclusive criteria included a history of systemic or gastrointestinal tract diseases, such as diabetes mellitus and inflammatory bowel

disease; current infectious diseases of the respiratory, digestive, or urinary system; and a history of abdominal surgery. Patients also were excluded if they had other types of psychological disorders. Healthy volunteers met the same criteria. Diet was not controlled.

Study Design and Procedures Diet information was collected by gastroenterologist interview from all participants. Symptom severity was evaluated by a validated questionnaire including IBS Symptom Severity Scores (SSS) and Self-Rating Depression Scale (SDS). Fecal samples were analyzed using the Roche 454 GS FLXþ Titanium platform (Roche 454 Life Sciences, Branford, CT) according to standard protocols.25 Visceral sensitivity was measured using a barostat. Host expression of inflammatory chemokines (mast cells, CD3þ T cells, MCP-1, and MIP-1a) were analyzed on sigmoid tissue biopsy specimens. All subjects provided written informed consent, and the study protocol was approved by the Ethics Committee of Peking University Health Science Center (no. 2013-112) (for technical details see the Supplementary Materials and Methods section).

Statistical Analysis Data were expressed as mean standard deviation/ standard error. The difference in bacterial composition between each of the patient groups and healthy controls was examined by the Wilcoxon rank-sum test, and the q value was used for correction (for details see the Supplementary Materials and Methods section). Comparisons of parametric data (age, body mass index, abdominal pain score, abdominal pain-onset frequency, abdominal bloating score, total IBS-SSS score, SDS score, MINI item number, and thresholds of visceral sensitivity, MCP-1, MIP-1a, mast cell, and CD3þ T-cell levels) between groups were performed by 1-way analysis of variance when more than 2 groups were compared and by the Student t test when 2 groups were compared. Nonparametric data (sex) were compared by the chisquare test. Correlations between inflammatory factors and abdominal symptoms were tested by Pearson correlation analysis. Analysis was performed with SPSS 18.0 (SPSS, Inc, Chicago, IL). A P value less than .05 was considered statistically significant.

Results Descriptive Analysis of Clinical Manifestations One hundred subjects were enrolled: 40 IBS-D patients, 15 depression patients, 25 comorbidity patients, and 20 healthy controls. The descriptive data of these 4 groups are summarized in Supplementary Table 1 and Figure 1. Age, body mass index, illness duration, and constitution of diets (Supplementary


1604 Liu et al

Clinical Gastroenterology and Hepatology Vol. 14, No. 11

Figure 1. Descriptive analysis of symptoms and barostat test data. (A) Abdominal pain and bloating scores were higher in patients with IBS-D and COMO than controls (mean SE, *P < .05). The number of bowel movements was significantly greater in patients with IBS-D and COMO than controls (mean SE, *P < .05). (B) Thresholds for urgency and maximum tolerance were lower in IBS-D patients than controls (*P < .05). Depression and COMO patients were between IBS-D and controls. (C) SDS scores were higher in IBS-D, depression, and COMO patents than in controls (mean SE, *P < .05), and even higher in depression and COMO patients than in IBS-D patients (mean SE, #P < .05). (D) The number of MINI secondary items was lower in COMO patients than depression patients. CON, control.

Figure 1) did not differ among the 4 groups. Abdominal symptoms, including abdominal pain, bloating, and maximum number of bowel movements, were significantly more severe in patients with IBS-D (P < .001, P < .001, P < .001, respectively) and comorbidities of IBS-D and depression (COMO) (P < .001, P < .001, P < .001, respectively) than in controls (Figure 1A). Among 50 subjects (7 controls, 20 IBS-D patients, 6 depression patients, and 17 COMO patients) who underwent rectal barostat examination, the sensory thresholds for urgency and maximum tolerance were lowest in IBS-D patients and significantly decreased compared with controls (P < .05 and P < .05, respectively) (Figure 1B). SDS scores were significantly higher in patients with IBS-D (P < .001), depression (P < .001), and COMO (P < .001) compared with controls (Figure 1C). On further comparison of the number of MINI secondary items between depression and COMO patients, depression patients had more items than COMO patients (Figure 1D).

Fecal Microbiome Features of DiarrheaPredominant Irritable Bowel Syndrome, Depression, and Comorbidities of Irritable Bowel Syndrome and Depression Patients, and Healthy Controls Phylum-level analysis showed markedly diminished distribution, with enrichment in Bacteroidetes phylum, and reduction of Firmicutes phylum in IBS-D, depression, and COMO (Figure 2A). At the genus level, a drastic increase in abundance of Bacteroides (P < .001, P < .001, and P < .001, respectively), Prevotella (P < .001, P < .001, and P < .001, respectively), Paraprevotella (P < .001, P < .001, and P < .001, respectively), and a decrease in abundance in Lachnospiracea incertae sedis (P < .001, P < .001, and P < .001, respectively), Coprococcus (P < .001, P < .001, and P < .001, respectively), Clostridium XI (P < .001, P < .001, and P < .001, respectively) were observed in IBS-D, depression, and COMO patients (Table 1). We used


November 2016


box plots to illustrate the biodiversity of each group. The median and average values of the Shannon index imply the microbial diversity. The diversity of bacterial communities was significantly lower in patients with IBS-D, depression, and COMO than in controls (P < .001, P < .001, P < .01, respectively) (Figure 2B). IBS-D and depression patients had similar microbial variations both in composition and diversity, whereas COMO patients showed less change than IBS-D and depression patients (Figure 2B and C). This feature also was observed on heat map analysis, an obvious reduction of several less-abundant genus groups were found in IBS-D and depression compared with controls (Figure 2D).

Characterization of Types Distinguished by Levels of Bacteroides and Prevotella Principal component analysis (PCA) and partitioning around medoids identified 3 distinct types designated as type I (containing 41 samples), type II (containing 17 samples), and type III (containing 42 samples), across all 100 samples (Figure 3A). The detailed analysis methods were introduced in the Supplementary Materials and Methods section. The heat map analysis showed high similarity among samples in the same type, and apparent dissimilarities between samples from different types (Figure 3B). The difference in the average Shannon index reflected statistically lower diversity of bacterial communities in type I and type II than in type III (P < .01, P < .01, and P < .01, respectively) (Figure 3C). The bacterial taxa composition of the 100 samples (Figure 3D) further showed the detailed differences among types. Microbial communities of samples in types I and II were dominated by a high level of 2 genera: Bacteroides (type I) and Prevotella (type II), respectively (Supplementary Figure 2). Samples in type III showed relatively balanced microbiota composition accompanied with higher diversity.

Differential Microbial Types in Diarrhea-Predominant Irritable Bowel Syndrome, Depression, and Comorbidities of Irritable Bowel Syndrome and Depression Patients PCA analysis was applied in patients with IBS-D, depression, and COMO separately to observe their differences from controls. Sixty percent of IBS-D patients were assigned to type I and 25% were assigned to type II, whereas 95% of control samples were classified as type III and only 5% were classified as type I (Figures 4A and 3D). Depression patients showed a similar distribution as IBS-D patients, with 60% of subjects fitted to type I and 20% fitted to type II (Figures 4B and 3D). In comparing COMO patients with controls, the first 2 principal components (58.6%) represented less than 70% of total

features, so the primary 3 principal components (70.13%) were applied. When investigated on the plane of PC2 (principal components 2) and PC3 (principal components 3), COMO patients and controls could be distinguished (Figure 4C). The results of PCA analysis indicated that the dominant genus in COMO patients was similar to that in controls. However, the nondominant bacteria showed different compositions (Supplementary Table 2).

Colonic Mucosal Expression of Monocyte Chemotactic Protein-1, Macrophage Inflammatory Protein-1 a and Immune Cells Associated With Fecal Microbial Types The expression of MCP-1, MIP-1a, mast cell tryptase, and CD3þ T cells in sigmoid mucosa was assayed immunohistochemically in 24 samples of type I, 14 samples of type II, and 23 samples of type III (including 33 IBS-D patients, 3 depression patients, 18 COMO patients, and 7 controls) (see Supplementary Table 3 for more detail). The histology graphs (Figure 5) representing different types are all from IBS-D patients. As shown in Figure 5A and B, significantly higher expression of MCP-1 (P < .001, P ¼ .001) and MIP-1a (P < .001, P ¼ .002) and more mast cells (P < .001, P < .001) and CD3þ T cells (P < .05, P < .05) were observed in subjects with microbial types I and II than in subjects with microbial type III. Upon performing further correlation analyses to evaluate the relationships between MCP-1 and MIP-1a overexpression and clinical symptoms in IBS-D and depression patients, the level of MCP-1 expression was found to correlate significantly with the scores of abdominal pain (P ¼ .014), abdominal pain frequency (P ¼ .001), and IBS-SSS (P ¼ .001) (Figure 5C).

Discussion An increasing number of studies have reported the dysbiosis in either IBS or depression. However, few human studies have analyzed the microbial association between IBS and depression, as well as the microbial composition in comorbidity of IBS and depression. In the present study, we found similar fecal microbiota signatures in IBS-D and depression patients and identified microbial types in IBS-D and depression, which correlated with clinical and pathophysiological parameters.

Microbiota-Related Types Different types were identified in patients with IBS-D, depression, COMO, and healthy controls. Types I and II were characterized by high levels of Bacteroides/ Prevotella, and type III showed a relatively balanced microbiota composition without any dominant genera. Arumugam et al26 analyzed the gut microbiome in 98 healthy volunteers across different populations and identified 2 clusters that they defined as enterotypes distinguished by levels of Bacteroides and Prevotella. In


1606 Liu et al


our study, types I and II also were characterized by Bacteroides and Prevotella, respectively, but at much higher levels than that reported by Arumugam et al26

(Bacteroides: 65% vs 46%, Prevotella: 70% vs 47%, respectively). Moreover, high levels of colonic mucosa inflammation were associated with types I and II. These


November 2016


Table 1. Bacterial Taxa Differ Between IBS-D Patients, Depression Patients, COMO Patients, and Healthy Controls Mean taxa number Phylum Actinobacteria Bacteroidetes

Firmicutes

Fusobacteria Proteobacteria

Genus Bifidobacterium Collinsella Prevotella Bacteroides Barnesiella Paraprevotella Odoribacter Butyricimonas Alistipes Parabacteroides Faecalibacterium Lachnospiracea_incertae_sedis Blautia Coprococcus Ruminococcus Mitsuokella Megamonas Clostridium XI Dialister Oscillibacter Clostridium sensu stricto Clostridium IV Catenibacterium Clostridium XlVb Roseburia Acetivibrio Veillonella Phascolarctobacterium Exiguobacterium Fusobacterium Escherichia/Shigella Parasutterella Gemmiger Comamonas Sutterella Haemophilus Vampirovibrio Acinetobacter

Control 25.90 4.85 164.20 2048.00 180.45 7.90 45.80 45.95 181.40 145.95 710.30 578.15 114.20 41.35 142.20 17.45 82.30 46.25 24.30 93.60 18.30 20.65 0.65 17.40 120.65 8.45 3.40 40.45 0.00 52.55 131.45 100.60 12.85 6.70 158.45 6.00 4.60 0.00

IBS-D

Depression a

6.90 3.33 1284.05a 3442.68a 90.45a 89.40a 19.08a 26.63a 168.48a 159.13 318.83a 358.10a 57.98a 8.48a 97.03a 0.15a 51.38a 21.45a 9.43a 72.65a 5.98a 10.65b 8.25b 9.75b 138.55 4.55 0.23 73.15 0.00 8.43a 4.63a 66.70a 4.80b 1.78b 208.66 8.68 7.78 0.48

b

18.47 3.93 1142.13a 4003.87a 24.60a 39.00a 13.80a 10.27a 112.67a 70.00a 558.47a 427.67a 49.47a 15.93a 36.13a 1.93a 61.27a 8.60a 51.73b 52.80a 4.13a 0.80a 0.27 12.33 69.27a 0.93a 16.33b 63.47 0.00 9.47a 22.73a 137.60 2.40a 1.73b 75.73a 26.47a 0.00b 0.07

COMO 19.92a 23.56b 970.40a 4318.04a 76.60a 29.40b 25.00a 26.96a 156.64a 158.44a 895.44a 400.28a 216.92 27.96a 150.64a 9.88a 24.64a 15.40a 15.84a 163.76 6.64a 11.88a 5.20 6.84a 181.44 7.64 3.56 151.24a 6.84b 335.88a 21.60a 166.48 14.44 4.60 38.76a 7.04 0.36b 100.48a

NOTE. Genus groups differed significantly between the healthy controls and the IBS-D, depression, and COMO patients. a P < .01. b P < .05.

data indicate that types I and II, with high levels of Bacteroides or Prevotella in the microbiota composition, might be associated with a high risk of IBS-D, depression, and COMO. Type I and Type II with high levels of Bacteroides or Prevotella may be as disease microbial signatures and worthy to further identify diagnostic biomarkers via other methods such as metagenomic sequencing.

Similarity of Microbial Signature Between Depression and Diarrhea-Predominant Irritable Bowel Syndrome Besides the dysbiosis in IBS, dysbacteriosis recently was reported in depression patients.14 In the present study, depression patients showed similar variation in microbiota composition to IBS-D patients. The

= Figure 2. Comparison of fecal microbiomes in IBS-D patients, depression patients, COMO patients, and controls. (A) Phylum distributions of microbiota in IBS-D patients, depression patients, COMO patients, and controls are shown by bar chart (mean SE). (B) Shannon diversity of bacterial communities was significantly lower in IBS-D patients, depression patients, and COMO patients than in controls (P < .001, P < .001, P < .01, respectively). (C) Family distributions of the microbial community in IBS-D patients, depression patients, COMO patients, and controls. The average proportion of microbial composition on family level are shown by a pie chart. (D) Heat map analysis on the genus level shows the general genus distributions of 100 samples. Downloaded from ClinicalKey.com at Peking University Health Science Center November 17, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

1608 Liu et al


Figure 3. Characterization of types distinguished by levels of Bacteroides and Prevotella. (A) Typing of 100 samples. The geometric centers of the 3 types are indicated by circles. PC, principal components. (B) The heat map analysis showed high similarity among samples in the same type and apparent dissimilarities between samples from different types. The scale bar represents the distance between samples. Blue color and number 0 means the 2 samples are exactly the same. The color turning to yellow represents dissimilarities between the given 2 samples. (C) Microbiome diversity in 3 types. (D) Microbiota communities on the genus level. The rows represent the 3 types, and columns represent the control, IBS-D, depression, and COMO patients.

proportion of IBS-D and depression patients classified into type I and type II also were similar (85% vs 80%). It already is well known that the reciprocal communication between the enteric nervous system and the central nervous system allows the brain to influence the motor, sensory, and secretory functions of the gastrointestinal

tract, as well as allowing the gastrointestinal tract to modulate brain function.27,28 According to our data, IBS-D and depression might be involved in common dysbacteriosis-related pathogenesis, which disturbed communication between the enteric nervous system and central nervous system by inducing intestinal


November 2016


Figure 4. PCA analysis showed that the microbiota communities of IBSD, depression, and COMO patients differed from controls. (A–C) Phylogenetic difference between patients and controls (IBS-D, depression, and COMO vs controls, respectively). Results were visualized from PCA and clustering based on genus composition. PC, principal components.

inflammation. These findings may offer insight into understanding the observed high comorbidity rate of IBS and depression. Abundance comparison between IBS-D and depression at the genus level (Table 1) also showed similar alterations, except for some genera that only changed in

depression but not in IBS-D patients (Parabacteroides [P < .01], Roseburia [P < .01], Verllonella [P < .05], sutterella [P < .01], and Haemophilus [P < .01]). Those bacterial subgroups might be responsible for the different clinical manifestation in IBS-D and depression patients. Among them, Parabacteroides and Roseburia

Figure 5. Association between colonic mucosa inflammation and microbiota signatures (types). (A) MCP-1, MIP-1a, and mast cell tryptase expression and CD3þ T cells in colonic mucosa of patients in type I, type II, and type III observed by immunohistochemistry (magnification, 200). (B) Expression of MCP-1 and MIP-1a and number of mast cells and CD3þ T cells were significantly higher in patients assigned to types I and II than in patients assigned to type III (mean SD, *P < .05). (C) Strong correlation between MCP-1 expression and clinical symptoms. Expression of MCP-1 correlated significantly with abdominal pain (P ¼ .014), frequency of abdominal pain (P ¼ .001), and IBS-SSS scores (P ¼ .001).


1610 Liu et al


were found to be decreased in IBD, and Roseburia mainly produces butyric acid, which is always used as an inflammation inhibitor.29,30 Tagami et al31 reported a fulminant case of Guillain–Barré syndrome with depression after Haemophilus infection, suggesting the close connection between Haemophilus and depression consistent with Haemophilus. Actually, our following study based on metagenomic sequencing found some differences in molecular mechanisms between IBS-D and depression (data not shown). These differential bacterial groups and their effects on further mechanisms may provide evidence for explaining the different clinical symptoms between IBS-D and depression.

Microbial and Clinical Variation in the Comorbidity of Diarrhea-Predominant Irritable Bowel Syndrome and Depression A microbiota population analysis in COMO patients showed less variation in both microbial abundance and diversity than IBS-D or depression. There also were fewer COMO patients than IBS-D and depression in types I and II, which were linked to a high risk of diseases. Although COMO patients had higher self-rated abdominal bloating and psychological SDS scores (Figure 1A and C) than IBS-D and depression patients, further analysis showed that they had fewer MINI secondary items than depression patients (Figure 1D) and less visceral sensitivity than IBS-D patients (Figure 1B), indicating that their pathophysiological alteration is not as severe as their complaint. Our results are consistent with those of Lackner et al,32 who identified a stronger correlation between psychosocial factors and self-ratings of health, among which depression had the strongest correlation with self-ratings of health. Thus, we speculate that depression made COMO patients tend to complain more about their discomfort and seek more medical care, even though they had less severe dysbiosis and pathophysiological disorders. This important finding will help to develop more appropriate methods of managing patients with comorbidities of IBS and depression. Microbiota is linked to abdominal symptoms in IBS as well as abnormal behavior and psychiatric disorders. The present study deeply analyzed the microbial association between IBS-D and depression and showed shared microbiota signatures between them. These data indicate that IBS-D and depression both are involved in microbiota–brain–gut interaction disorders, which may offer insight into understanding their observed high comorbidity rate. The defined microbial types associated with clinical and pathophysiological parameters may help to identify new subtypes and effective diagnostic biomarkers in IBS-D and depression patients. However, further investigations such as metagenomics and metabonomics are required to characterize the function and molecular mechanisms of the variable bacteria in

IBS-D and depression and help to develop new diagnostic biomarkers and treatment strategies.

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2016.05.033.

References 1. Spiller R, Aziz Q, Creed F, et al. Guideline on the irritable bower syndrome: mechanisms and practical management. Gut 2007; 56:1770–1798. 2. Bengtson MB, Aamodt G, Vatn MH, et al. Co-occurrence of IBS and symptoms of anxiety or depression, among Norwegian twins, is influenced by both heredity and intrauterine growth. BMC Gastroenterol 2015;15:9. 3. Modabbernia MJ, Mansour-Ghanaei F, Imani A, et al. Anxietydepressive disorders among irritable bowel syndrome patients in Guilan, Iran. BMC Res Notes 2012;5:112. 4. Mayer EA, Tillisch K. The brain-gut axis in abdominal pain syndromes. Annu Rev Med 2011;62:381–396. 5. Barreau F, Salvador-Cartier C, Houdeau E, et al. Long-term alterations of colonic nerve-mast cell interactions induced by neonatal maternal deprivation in rats. Gut 2008;57:582–590. 6. Xu D, Gao J, Gillilland M, et al. Rifaximin alters intestinal bacteria and prevents stress-induced gut inflammation and visceral hyperalgesia in rats. Gastroenterology 2014;146:484–496. 7. Jeffery IB, O’Toole PW, Öhman L, et al. An irritable bowel syndrome subtype defined by species-specific alterations in fecal microbiota. Gut 2012;61:997–1006. 8. Ng SC, Lam EF, Lam TT, et al. Effect of probiotic bacteria on the intestinal microbiota in irritable bowel syndrome. J Gastroenterol Hepatol 2013;28:1624–1631. -Stojanovic M, Biagi E, Heilig HG, et al. Global and deep 9. Rajilic molecular analysis of microbiota signatures in fecal samples from patients with irritable bowel syndrome. Gastroenterology 2011;141:1792–1801. 10. Codling C, O’Mahony L, Shanahan F, et al. A molecular analysis of fecal and mucosal bacterial communities in irritable bowel syndrome. Dig Dis Sci 2010;55:392–397. 11. Parkes GC, Rayment NB, Hudspith BN, et al. Distinct microbial populations exist in the mucosa-associated microbiota of sub-groups of irritable bowel syndrome. Neurogastroenterol Motil 2012;24:31–39. 12. Ouwehand AC, Salminen S, Arvola T, et al. Microbiota composition of the intestinal mucosa: association with fecal microbiota? Microbiol Immunol 2004;48:497–500. 13. Durban A, Abellan JJ, Jimenez-Hernandez N, et al. Assessing gut microbial diversity from feces and rectal mucosa. Microb Ecol 2011;61:123–133. 14. Naseribafrouei A, Hestad K, Avershina E, et al. Correlation between the human fecal microbiota and depression. Neurogastroenterol Motil 2014;26:1155–1162. 15. Diaz Heijtz R, Wang S, Anuar F, et al. Normal gut microbiota modulates brain development and behavior. Proc Natl Acad Sci U S A 2011;108:3047–3052. 16. Ogbonnaya ES, Clarke G, Shanahan F, et al. Adult hippocampal neurogenesis is regulated by the microbiome. Biol Psychiatry 2015;78:e7–e9.


November 2016


17. Mayer EA, Savidge T, Shulman RJ, et al. Brain gut microbiome interactions and functional bowel disorders. Gastroenterology 2014;146:1500–1512. 18. Cryan JF, Dinan TG. More than a gut feeling: the microbiota regulates neurodevelopment and behavior. Neuropsychopharmacology 2015;40:241–242. 19. Murphy PM, Baggiolini M, Charo IF, et al. International Union of Pharmacology. XXII. Nomenclature for chemokine receptors. Pharmacol Rev 2000;52:145–176. 20. Alemán JO, Eusebi LH, Ricciardiello L, et al. Mechanisms of obesity-induced gastrointestinal neoplasia. Gastroenterology 2014;146:357–373. 21. Olefsky JM, Glass CK. Macrophages, inflammation, and insulin resistance. Annu Rev Physiol 2010;72:219–246. 22. Souza DG, Vieira AT, Soares AC, et al. The essential role of the intestinal microbiota in facilitating acute inflammatory responses. J Immunol 2004;173:4137–4146. 23. Ritter U, Moll H. Monocyte chemotactic protein-1 stimulates the killing of leishmania major by human monocytes, acts synergistically with IFN-gamma and is antagonized by IL-4. Eur J Immunol 2000;30:3111–3120. 24. van Zuiden M, Heijnen CJ, van de Schoot R, et al. Cytokine production by leukocytes of military personnel with depressive symptoms after deployment to a combat-zone: a prospective, longitudinal study. PLoS One 2011;6:e29142. 25. Margulies M, Egholm M, Altman WE, et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature 2005; 437:376–380. 26. Arumugam M, Raes J, Pelletier E, et al. Enterotypes of the human gut microbiome. Nature 2011;473:174–180. 27. Rhee SH, Pothoulakis C, Mayer EA. Principles and clinical implications of the brain-gut-enteric microbiota axis. Nat Rev Gastroenterol Hepatol 2009;6:306–314.

28. Mayer EA. Gut feelings: the emerging biology of gut-brain communication. Nat Rev Neurosci 2011;12:453–466. 29. Zitomersky NL, Atkinson BJ, Franklin SW, et al. Characterization of adherent bacteroidales from intestinal biopsies of children and young adults with inflammatory bowel disease. PLoS One 2013;8:e63686. 30. Chen L, Wang W, Zhou R, et al. Characteristics of fecal and mucosa-associated microbiota in Chinese patients with inflammatory bowel disease. Medicine (Baltimore) 2014; 93:e51. 31. Tagami S, Susuki K, Takeda M, et al. Fulminant case of GuillainBarré syndrome with poor recovery and depression following Haemophilus influenzae infection. Psychiatry Clin Neurosci 2008;62:486. 32. Lackner JM, Gudleski GD, Thakur ER, et al. The impact of physical complaints, social environment, and psychological functioning on IBS patients’ health perceptions: looking beyond GI symptom severity. Am J Gastroenterol 2014; 109:224–233.

Reprint requests Address requests for reprints to: Liping Duan, MD, Department of Gastroenterology, Peking University Third Hospital, No. 49 North Garden Road, Haidian District, Beijing, 100191 China. e-mail: [email protected]; fax: (86) 10-82801250; or Huaiqiu Zhu, PhD, Department of Biomedical Engineering, Peking University, No. 5 Yiheyuan Road, Haidian District, Beijing, 100871 China. e-mail: [email protected]; fax: (86) 10-62767261. Conflicts of interest The authors disclose no conflicts. Funding The studies were supported by the National “Twelfth Five-Year” Plan for Science and Technology of China (2012BAI06B02) and the National Natural Science Foundation of China (91231119).


1611.e1 Liu et al


Supplementary Materials and Methods Identification of Depression With the Mini-International Neuropsychiatric Interview The MINI was developed by Sheehan et al1 to meet the need for a brief, reliable, and valid structured diagnostic interview for each major Axis I psychiatric disorder in the DSM-IV and the International Classification of Diseases, 10th revision. The sensitivity and specificity for the Chinese version of the depression MINI is 92.2% and 86.0%, respectively.2

Laboratories, Inc, Carlsbad, CA). The procedure includes the following main steps: (1) 1 g fecal sample was homogenized by vortexing for 30 to 40 seconds together with 2.5 mL MO BIO lysis buffer in a 50-mL Falcon tube and then centrifuged for 5 minutes at 1500 relative centrifugal force; (2) 750 mL supernatant was removed to a 2 mL MO BIO Bead tube and homogenized further for 10 minutes at high speed on a vortex adaptor (Vortex-Genie 2; MO BIO Laboratories, Inc); and (3) series regent was added subsequently for elimination impurities such as proteins and lipids. DNA finally was dissolved in purified water at a volume of 100 mL and stored at -80 C until sequencing.

454 Pyrosequencing

Rectal Barostat The subjects were asked to self-administer an enema with 10 mL glycerin at least 30 minutes before the function test for the preparation of the rectum and then to lie down in a bed in a right lateral position. A computer-controlled barostat device (Distender Series II; G&J Electronics, Ontario, Canada) was applied to assess visceral sensitivity of the subjects. A double-lumen catheter attached to a noncompliant polyethylene barostat balloon, which was 10-cm long and had a maximum volume of 600 mL, at the distal end was inserted into the rectum and positioned 5-cm proximal to the anal verge, the proximal end of the catheter was connected to the barostat device. The barostat balloon was inflated with 200 mL air for 1 minute to ensure that the bag was placed in the right position and to let the subjects accommodate to the distension, then the balloon was deflated completely and the subjects were allowed a 5-minute rest. Subsequently, a series of isobaric distensions were performed with stepwise increments starting at 0 mm Hg and increasing 2 mm Hg until 60 mm Hg or when the subjects felt an urgent need to defecate. Each distension lasted for 60 seconds and every 2 distensions were separated by a 30-second interval. The subjects were instructed to report their feelings at each distension, 4 sensory thresholds were recorded—that is, first sense to distension, first sense to defecate, urge to defecate, and the maximum tolerance pressure. The first sense to distension was defined as the minimal pressure to make the subjects conscious of the distensions, the first sense to defecate was defined as the minimal pressure to make the subjects want to defecate, urge to defecate was defined as the pressure to make the subjects feel an urgent need to defecate, the maximum tolerance was defined as the pressure that subjects could not tolerate any longer. For the subjects who never felt an urgent need to defecate, a pressure of 60 mm Hg was regarded as the maximum tolerance.

DNA Extraction Stool samples were stored at -80 C as soon as they were collected. Fecal microbial DNA was isolated according to the procedure recommended by the Human Microbiome Project3 by using the PowerSoil DNA Isolation Kit (MO BIO

The V1 to V3 region of the bacteria 16S ribosomal RNA gene were amplified by polymerase chain reaction (95 C for 2 min, followed by 25 cycles at 95 C for 30 s, 55 C for 30 s, and 72 C for 30 s, and a final extension at 72 C for 5 min) using primers 27F 5’-AGAGTTTGATCCTGGCTCAG-3’, 533R 5’-TTACCGCGGCTGCTGGCAC-3’, and MID tags (for distinguish the direction of the sequence) 5’-ATTACCGCGGCTGCTGGCA-3’. Polymerase chain reactions were performed in a 20-mL mixture containing 4 mL of 5 FastPfu Buffer, 2 mL of 2.5 mmol/L deoxynucleoside triphosphates, 0.8 mL of each primer (5 mmol/L), 0.4 mL of FastPfu Polymerase, and 10 ng of template DNA (TransStart Fastpfu DNA Polymerase kit, TransGen, Beijing,China). After purification using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA) and quantification using QuantiFluor-ST (Promega, Madison, WI), a mixture of amplicon was used for pyrosequencing on a Roche 454 GS FLXþ Titanium platform (Roche 454 Life Sciences) according to standard protocols.

Immunohistochemistry Immunohistochemistry staining was performed using the ZSGB-BIO ALK system (ZK-9600; Origene; and ZSGB-BIO; MO BIO, Beijing, China) on formalin-fixed, paraffin-embedded biopsy specimens with primary antibodies against MCP-1 (ab9669, 1:400; Abcam, Cambridge, MA), MIP-1a (sc-33203, 1:400; Santa Cruz Biotechnology, Dallas, TX), mast cell tryptase (ab2378, 1:200; Abcam), and CD3þ T cells (ab699, 1:50; Abcam). As a secondary antibody and for visualization, a peroxidase/3,3’-diaminobenzidine-positive was used according to the manufacturer’s protocol (ZSGB-BIO ALK Detection System Peroxidase 3,3’-diaminobenzidine-positive rabbit/mouse; PV-6000-D; Origene; and ZSGB-BIO; MO BIO). The expression levels were evaluated based on staining intensity and the percentage of positive cells. MCP-1 and MIP-1a were scored using the immunoreactive score4,5: immunoreactive score ¼ staining intensity percentage of positive cells. Staining intensity was scored as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. The percentage of positive cells was scored as follows: 0, 0% to


November 2016

Shared Microbiota in IBS–D and Depression 1611.e2

1%; 1, 2% to 10%; 2, 11% to 50%; 3, 51% to 80%; and 4, more than 80% positive cells. Thus, a score from 0 to 12 points is obtained. MCP-1 and MIP-1a overexpression was defined as an immunoreactive score of 6 or greater.

Operational Taxonomic Unit Calculation The rarefaction curve was plotted by means of several modules of Mothur.6 First, a mixture of the entire 16S ribosomal RNA sequences with a total number of 953,571 from the 100 samples were aligned by default parameters of the align.seqs function, and the dist.seqs function was used to calculate the distance matrix via the furthest neighbor approach. Then the cluster function assigned the sequences into operational taxonomic units. Finally, the make.shared and the rarefaction.shared functions transformed the operational taxonomic unit information into rarefaction data by sampling with identity 100%, 97%, and 95%. The rarefaction curve consequently was plotted with abscissa and ordinate to be the number of samples and observed operational taxonomic units, respectively.

Phylogenetic Annotation and Species Abundance Measurement To assign the 16S ribosomal RNA sequences at each phylotype level accurately, all reference bacterial 16S ribosomal RNA sequences from Ribosomal database project (RDP) resources were downloaded and constructed into the basic local alignment search tool (BLAST) database. Then the blastall function in BLAST software was used to perform slow and accurate alignment with an e-value of 10-5 or less and the highest score. After that, the occurrence at the level of genus, family, order, class, and phylum of each sample was calculated. Then a distance matrix was created first through the genus frequency matrix using Phylip. The Euclidean distance was used to calculate the sample difference, as follows: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uX u n 2 distðXi1 ; Xi2 Þ ¼ jXi1 Xi2 j ¼ t xi1 j xi2 j j¼1

In which, Xi ¼ ðxi1 ; xi2 /; xin Þ represents the normalized frequency distribution of genus abundance among the whole n detected genera in the i-th sample. Then heat maps were drawn using the pheatmap, which is a command line, and lattice library in R software. To estimate the biodiversity of microorganisms in each sample, the Shannon index of each sample was calculated.

these samples cannot be clustered simply through the Euclidean or maximum distance. Assuming that each genus frequency in 1 sample represents the 1 feature vector of the sample, all of these samples were distributed in high dimensional space. To cluster and separate samples correctly and show the best result in the plane, PCA was used to transform multivariate into few main components through projecting samples onto 2- or 3-dimensional space with the demand that samples can be best separated. Usually, the first 2 or 3 principal components must represent 70% of all the sample features at least. Assuming M represents the number of samples and N represents the number of genera, the variance matrix S is formed as follows: S ¼

M T 1 X XðkÞ X XðkÞ X M 1 k¼1

In which, XðkÞ ¼ ðxk1 ; xk2 ; /; xkN ÞT ðk ¼ 1; 2; /; MÞ X ¼

M T 1 X XðkÞ ¼ x1 ; x2 ; /; xp M k¼1

Then the eigenvalues of S were calculated. By adding the proportion of the first n eigenvalue until they represented more than 70% of sample features, samples were separated on the n-dimension space, and the n eigenvalues is the spatial coordinates of the sample. After PCA, partitioning around medoids was used to cluster these samples according to the n eigenvalues. The optimal cluster numbers were estimated by the Calinski–Harabasz index. The eigenvectors corresponding to the n eigenvalues show the spatial coordinates of each sample.

References 1. Sheehan DV, Lecrubier Y, Sheehan KH, et al. The Mini International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 1998;59:22–33. 2. Si TM, Shu L, Dang WM, et al. Evaluation of the reliability and validity of Chinese version of the Mini-International Neuropsychiatric Interview in patients with mental disorders [Chinese]. Chin Ment Health J 2009;23:493–497. 3. McInnes P, Cutting M. Manual of procedures for human microbiome project: core microbiome sampling protocol A HMP protocol # 07-001, Version 12.0 2010. Available at: http:// hmpdacc.org/resources/tools_protocols.php. 4. Remmele W, Stegner HE. Recommendation for uniform definition of an immunoreactive score (irs) for immunohistochemical estrogen receptor detection (er-ica) in breast cancer tissue. Der Pathologe 1987;8:138–140. 5. Lindner JL, Loibl S, Denkert C, et al. Expression of secreted protein acidic and rich in cysteine (SPARC) in breast cancer and response to neoadjuvant chemotherapy. Ann Oncol 2015;26:95–100.

Sample Cluster PCA and partitioning around medoids were used to classify samples. The count matrix of samples was normalized to frequency matrix. We have shown that all of

6. Schloss PD, Westcott SL, Ryabin T, et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 2009;75:7537–7541.


1611.e3 Liu et al


100%

Balanced High fat/protein

80%

Vegetarian

60% 40% 20% 0% Control

IBS-D

Depression

COMO

Supplementary Figure 1. The distribution of 3 diet types. Constituent ratio of 3 different diet types of subjects in each group. There was no significant difference of the constituent ratio of any diet type among groups. DEP, depression.


November 2016

Shared Microbiota in IBS–D and Depression 1611.e4

Supplementary Figure 2. Dominant genera abundance in types I and II. Type I is identified by Bacteroides dominance (average, 63.05%); and type II is marked by Prevotella dominance (average, 70.89%).

Supplementary Table 1. Descriptive Data of the Subjects Enrolled in the Study

Number Age SD, y Sex, male/female BMI Illness duration, mo

Control

IBS-D

Depression

COMO

20 43.9 11.2 7/13 24.6 2.2 NA

40 38.5 13.6 28/12 22.6 3.1 40.5 38.1

15 44.8 14.9 4/11 22.0 3.2 30 36.5

25 39.0 13.9 14/11 22.0 3.8 34.5 54.1

NOTE. Descriptive data of patients with IBS-D, depression, COMO, and controls. BMI, body mass index; NA, data not available.


1611.e5 Liu et al


Supplementary Table 2. Comparison of Bacteria Groups on Genus Level Between COMO and Health Controls (COMO vs Control) Phylum

Genus

Bacteroidetes

Paraprevotella Bacteroides Lachnospiracea_incertae_sedis Coprococcus Turicibacter Veillonella Flavonifractor Victivallis Parasutterella

Firmicutes

Lentisphaerae Proteobacteria

Health 10 1846.54 368.54 35.15 2.85 3.15 1.92 5.92 77.08

24.62 1401.67 322.40 64.42 3.63 7.87 3.28 10.56 105.16

COMO

P value

.097 .091 .037 .045 .075 .079 .089 .010 .086

13.5 5152.56 316.31 32.18 0.875 95.625 5.56 0.0625 67.0625

23.18 4013.98 319.37 97.93 1.93 371.85 7.87 0.25 156.78

Supplementary Table 3. Number of Subjects Counted in the Comparison of the Expression of Inflammatory Factors

Control IBS-D Depression COMO

Type I

Type II

Type III

0 20 1 3

8 1 5

7 5 1 10
