Prendi

YaJan

Gast

Texa

Veterinary Microbiology 174 (2014) 463–473

A R

Artic

Rece

Rece

Acce

Keyw

Clos

Dys

Dog

Ente

Mic

Path

perf

SCFA

*

447

http

037

evalence of Clostridium perfringens, Clostridium perfringensterotoxin and dysbiosis in fecal samples of dogs witharrhea

sushi Minamoto, Naila Dhanani, Melissa E. Markel, Jorg M. Steiner, S. Suchodolski *

rointestinal Laboratory, Department of Small Animal Clinical Sciences, College of Veterinary Medicine and Biomedical Science,

s A&M University, College Station, TX, USA

T I C L E I N F O

le history:

ived 16 June 2014

ived in revised form 7 October 2014

pted 9 October 2014

ords:

tridium perfringens

biosis

rotoxin

robiota

ogen

A B S T R A C T

Clostridium perfringens has been suspected as an enteropathogen in dogs. However, its

exact role in gastrointestinal (GI) disorders in dogs remains unknown. Recent studies

suggest the importance of an altered intestinal microbiota in the activation of virulence

factors of enteropathogens. The aim of this study was to evaluate the relationship between

diarrhea, dysbiosis, and the presence of C. perfringens and its enterotoxin (CPE). Fecal

samples were collected prospectively from 95 healthy control dogs and 104 dogs with GI

disease and assessed for bacterial abundances and the presence of CPE using quantitative

PCR and ELISA, respectively. C. perfringens was detected in all dogs. Potentially

enterotoxigenic C. perfringens were detected in 33.7% (32/95) of healthy control dogs

and 48.1% (50/104) diseased dogs, respectively. CPE was detected by ELISA in 1.0% (1/95) of

control dogs and 16.3% (17/104) of diseased dogs. Abundances of Fusobacteria,

Ruminococcaceae, Blautia, and Faecalibacterium were significantly decreased in diseased

dogs, while abundances of Bifidobacterium, Lactobacillus, and Escherichia coli were

significantly increased compared to control dogs. The microbial dysbiosis was indepen-

dent of the presence of the enterotoxigenic C. perfringens or CPE. In conclusion, the

presence of CPE as well as fecal dysbiosis was associated with GI disease. However, the

presence of C. perfringens was not indicative of GI disease in all cases of diarrhea, and the

observed increased abundance of enterotoxigenic C. perfringens may be part of intestinal

dysbiosis occurring in GI disease. The significance of an intestinal dysbiosis in dogs with GI

disease deserves further attention.

Published by Elsevier B.V.

Abbreviations: RNA, 16S rRNA 16S ribosomal; AHD, acute hemorrhagic diarrhea; CPE, Clostridium perfringens enterotoxin; cpe gene, Clostridium

ringens enterotoxin gene; GI, gastrointestinal; qPCR, quantitative real-time polymerase chain reaction; RPLA, reverse passive latex agglutination assay;

, short-chain fatty acid.

Corresponding author at: Department of Small Animal Clinical Sciences, College of Veterinary Medicine and Biomedical Science, Texas A&M University,

4 TAMU, College Station, TX 77843-4474, USA. Tel.: +1 979 458 0933; fax: +1 979 458 4015.

E-mail address: jsuchodolski@cvm.tamu.edu (J.S. Suchodolski).

Contents lists available at ScienceDirect

Veterinary Microbiology

jo u rn al ho m epag e: ww w.els evier .c o m/lo cat e/vetmic

://dx.doi.org/10.1016/j.vetmic.2014.10.005

8-1135/Published by Elsevier B.V.

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473464

1. Introduction

Clostridium perfringens is a commensal in the caninegastrointestinal (GI) tract (Marks et al., 2011) andresponsible for a spectrum of diseases. C. perfringens

enterotoxin (CPE) is thought to be an important virulencefactor in dogs with C. perfringens-associated diarrhea(Weese et al., 2001). CPE is encoded by the enterotoxingene (cpe gene) and synthesized during sporulation. Afterrelease from sporulating cell, CPE induces toxicity byinteracting with intestinal tight junctions, formingtransmembrane pores on the cytoplasmic membraneand leading to altered epithelial permeability (Mcclane,1992). In humans, CPE is responsible for several GIdiseases including C. perfringens type A food poisoning(Lahti et al., 2008), antibiotic-associated diarrhea (Abra-hao et al., 2001), and nosocomial diarrheal disease(Watanabe et al., 2008). Detection of CPE in feces fromhuman patients with diarrhea is a criterion for diagnosisof CPE-associated diarrhea (Miyamoto et al., 2012).Several studies have evaluated the role of C. perfringens

in dogs, and have reported C. perfringens as a potentialcause of nosocomial diarrhea (Kruth et al., 1989) andacute hemorrhagic diarrhea (AHD) (Cave et al., 2002;Unterer et al., 2014). Clinical signs are usually mild andself-limiting. Therefore, further diagnostic modalities,such as an endoscopy and abdominal exploration, arerarely conducted. Recently, evaluation of histopatholog-ical changes and the presence of bacteria in duodenalbiopsies from dogs with AHD revealed an increasedabundance of mucosa-adherent C. perfringens in 6/9 ofdogs (Unterer et al., 2014).

However, despite these studies, the exact role of C.

perfringens in canine GI disease remains unknown becausethis organism is detected at similar isolation rates inhealthy and diarrheic dogs. Furthermore, these studieswere conducted in a different country (i.e., Canada)(Goldstein et al., 2012; Weese et al., 2001), used samplesfrom shelter animals or referral hospitals (Cave et al., 2002;Marks et al., 1999; Tupler et al., 2012), used a differentELISA assay (Kruth et al., 1989), or used different detectionmethods (reverse passive latex agglutination assay[RPLAA]) (Marks et al., 1999). There is currently also nogold standard for the diagnosis of C. perfringens-associateddiarrhea in dogs. Therefore, confirmatory interpretation ofdata from different diagnostic assays (i.e., detection of thecpe gene by PCR and CPE by ELISA in fecal samples fromdogs) is recommended (Marks et al., 2011).

Recently, molecular studies have evaluated the diversebacterial communities in the canine GI tract, and haveshown the presence of intestinal dysbiosis, defined asaltered GI microbial communities, in dogs with GI diseases.Commonly observed alterations are decreases in Rumino-coccaceae, Faecalibacterium, Turicibacter, and Bacteroi-detes, with concurrent increases in Proteobacteria,especially Escherichia coli (Suchodolski et al., 2008,2012a,b,c; Xenoulis et al., 2008). Of particular interest isthat, while it has been shown that GI inflammation inducesGI dysbiosis (Craven et al., 2012), it is also thought thatprolonged dysbiosis may aggravate intestinal inflamma-tion (Duboc et al., 2013). Therefore, dysbiosis plays a

pivotal role in the pathogenesis of GI disease (Hall, 2011;Packey and Sartor, 2009; Round and Mazmanian, 2009).

Limited information is available regarding the relation-ship between diarrhea, dysbiosis, and/or the presence of anenteric pathogen and its virulence factors. This is ofimportance, as recent studies suggest that activation ofvirulence factors of Clostridium difficile and Salmonella isassociated with dysbiosis and concurrent changes inmetabolite profiles such as altered bile acid and short-chain fatty acid (SCFA) concentrations (Bearson et al.,2013; Weingarden et al., 2014). Therefore, this study firstaimed to investigate the prevalence of C. perfringens andCPE in healthy dogs and dogs with clinical signs of GIdisease. The second aim was to quantify the abundance ofC. perfringens and enterotoxigenic C. perfringens bydetecting C. perfringens 16S ribosomal RNA (16S rRNA)gene and C. perfringens enterotoxin gene (cpe gene),respectively. The third aim was to evaluate the relation-ships between diarrhea, dysbiosis, and the presence of C.

perfringens and its enterotoxin (CPE).

2. Materials and methods

2.1. Fecal samples

Fecal samples from healthy control dogs and dogs withclinical signs of GI disease were collected prospectivelyfrom April 2010 to June 2012. The protocol for samplecollection was approved by the Clinical Research ReviewCommittee of the College of Veterinary Medicine, TexasA&M University (CRRC#09-06).

2.1.1. Healthy control dogs

A total of 95 privately owned dogs without clinical signsof GI disease (i.e., vomiting, diarrhea, anorexia, weight loss,etc.) within the past 3 months of sample collection wereenrolled. Dogs that received antibiotics within the past3 months were excluded. Fresh fecal samples werecollected at home or at public dog parks, and transportedon ice to the Gastrointestinal Laboratory.

2.1.2. Diseased dogs

A total of 104 left-over fecal samples from submissionsto the Gastrointestinal Laboratory at Texas A&M Universitywere utilized for this study. These submissions were fromdogs with clinical signs of GI disease (i.e., vomiting,diarrhea, anorexia, weight loss, etc.) based on the clinicalhistory and were submitted for enteric pathogen testingand/or fecal biomarker testing. Only the first submissionsample was utilized for this study when multiple samplesfrom same dog were submitted. The time of fecal samplecollection after onset of diarrhea varied between samplesdepending on the time the dog was presented to theveterinarian.

2.2. Clinical history of dogs

Questionnaires were sent to veterinarians who submit-ted fecal samples and the owners of healthy control dogs.The questionnaire was composed of three major parts:signalment of dogs (breed, age, sex, body weight, and body

consamisti(mediseclasnonnatwe

2.3.

shiconfromfor

posfor

andaftetemunpcon

2.4.

weextCorinshomCA)�8

2.5.

per

eva(Ta16S10

Pro2.2(30finacon

Tab

Olig

Ta

C.

C.

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473 465

dition score), health status of dogs at time of fecalple collection (presence of GI signs and its character-

cs, duration of GI signs), and medical history of dogsdication [use of antibiotics, probiotics, etc.], concurrentases). Dogs with clinical signs of GI disease weresified based on the type of diarrhea (acute, chronic, or-diarrhea). Diarrhea was characterized as acute in

ure if present for <3 weeks and chronic if present for �3eks.

Evaluation of the stability of CPE in fecal samples

To test the stability of the CPE in fecal samples duringpping to the laboratory, the stability at different storageditions was evaluated. Thirteen leftover fecal samples

dogs with clinical signs of GI disease were screenedCPE. Of these samples, 8 fecal samples were initiallyitive for CPE and 5 fecal samples were initially negativeCPE. These samples were subdivided into 8 aliquots, then evaluated at 4 time points (day 0, 2, 5, and 10)r having been stored at 3 storage conditions (roomperature, 4 8C, and �20 8C). Each aliquot was storedrocessed in microcentrifuge tubes at each storagedition until processing for the ELISA.

DNA extraction

DNA was extracted from an aliquot of 100 mg (wetight) of each fecal sample using a commercial DNAraction kit (ZR Fecal DNA KitTM; Zymo Researchporation, Irvine, CA) following the manufacturer’structions. The bead-beating step was performed on a

ogenizer (FastPrep-24; MP Biomedicals, Santa Ana, for 60 s at a speed of 4 m/s. Fecal DNA was stored at

0 8C until analysis.

Quantitative PCR (qPCR) assay

The abundances of C. perfringens 16S rRNA gene and C.

fringens enterotoxin gene (cpe gene) in feces wereluated by qPCR assays using published oligonucleotidesble 1). Quantitative PCR for detection of C. perfringens

rRNA gene was conducted using a total volume ofmL, with the mastermix containing 5 mL of SsoFastTM

bes supermix (Biorad Laboratories, Hercules, CA), mL of water, 0.3 mL of each forward and reverse primer0 nM final concentration), 0.2 mL of the probe (200 nMl concentration), and 2 mL of DNA. The qPCR cyclingditions were: an initial incubation at 94 8C for 10 min,

45 cycles of denaturation at 94 8C for 10 s, annealing at58 8C for 20 s, and extension at 70 8C for 10 s. The qPCR forthe detection of the cpe gene in feces was conducted usinga total volume of 10 mL, with the mastermix containing5 mL of SsoFastTM Probes supermix, 2.35 mL of water,0.25 mL of each primer (final concentration: 250 nM),0.15 mL of the probe (150 nM final concentration), and 2 mLof DNA. The qPCR cycling conditions were: an initialincubation at 95 8C for 2 min, 40 cycles of denaturation at95 8C for 5 s, and annealing for 10 s at 55 8C.

To assess the abundances of bacterial groups, whichpreviously have been shown to be altered in canine GIdiseases (Rossi et al., 2014; Suchodolski et al., 2008,2012a,b,c; Xenoulis et al., 2008), qPCR assays wereperformed for total bacteria, Fusobacteria, Ruminococca-ceae, Bifidobacterium spp., Blautia spp., Faecalibacterium

spp., Lactobacillus spp., and E. coli. The assay conditions, theoligonucleotide sequences of primers and probes, andrespective annealing temperatures were described previ-ously (Suchodolski et al., 2012c). A commercial real-timePCR thermal cycler (CFX384 TouchTM Real-Time PCRDetection System; Biorad Laboratories, Hercules, CA)was used for all qPCR assays and all samples were runin duplicate.

2.6. ELISA for CPE

C. perfringens enterotoxin (CPE) was detected using acommercially available ELISA kit (C. perfringens Enterotox-in Test; TechLab, Blacksburg, VA). The test was performedaccording to the manufacturer’s instructions. Briefly, fecalsamples (an amount equal to 3 mm of formed feces or50 mL of liquid feces) were emulsified in 200 mL of diluentand vortexed for 10 s. One hundred microliter of thediluted sample was then transferred to the microassay wellcontaining the detecting polyclonal antibody against thetoxin. The ELISA reaction was evaluated spectrophotomet-rically using a commercial multi-mode microplate reader(Synergy 2 Multi-Mode Microplate Reader; BioTek,Winooski, VT) at 450 nm wavelength. Samples with opticaldensity (OD)450� 0.120 were considered positive, andsamples with OD450< 0.120 were considered negative.

2.7. Statistical analysis

Datasets for healthy control dogs and dogs with clinicalsigns of GI disease were tested for normality using aShapiro–Wilk test, and then compared using a Wilcoxonrank-sum test (for 2 groups) or Kruskal–Wallis tests with

le 1

onucleotides used for the detection of the C. perfringens 16S rRNA gene and the C. perfringens enterotoxin gene in this study.

rget qPCR primers/probes Sequence (50–30) Annealing

temperature (8C)

Reference

perfringens 16S

rRNA gene

CPerf165F CGCATAACGTTGAAAGATGG 58 Wise and Siragusa (2005)

CPerf269R CCTTGGTAGGCCGTTACCC

CPerf187F (probe) FAM-TCATCATTCAACCAAAGGAGCAATCC-TAMURA

perfringens

enterotoxin gene

cpe F AACTATAGGAGAACAAAATACAATAG 55 Gurjar et al. (2008)

cpe R TGCATAAACCTTATAATATACATATTC

cpe Pr (probe) FAM-TCTGTATCTACAACTGCTGGTCCA-TAMURA

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473466

Dunn’s post-tests (>2 groups) wherever appropriate. Ap < 0.05 was considered significant. For the evaluation ofGI microbiota, all data were adjusted for multiplecomparisons using a Bonferroni correction and an adjustedp < 0.05 was considered significant. A Spearman’s rankcorrelation coefficient was used to evaluate the correlationbetween the abundance of C. perfringens 16S rRNA geneand the cpe gene. All statistical analyses were conductedusing a statistical software package (JMP1 Pro version 10,SAS Institute Inc, Cary, NC).

3. Results

3.1. Clinical history of dogs

A total of 95 fecal samples from healthy control dogsand 104 fecal samples from dogs with clinical signs of GIdisease were utilized for this study. The median age ofhealthy dogs and dogs with clinical signs of GI disease was3 years (range: 0.6–12 years) and 5 years (range: 0.4–15years), respectively (p = 0.006; Fig. 1). Of the healthycontrol dogs, 38 were male (3 intact, 35 castrated) and57 female (4 intact, 53 spayed). Of the dogs with clinicalsigns of GI disease, 54 were male (11 intact, 43 castrated)and 50 female (13 intact, 37 spayed). There was nosignificant difference between the genders between thetwo groups (p = 0.117). The median body weight of healthycontrol dogs and dogs with clinical signs of GI disease were23.9 kg (range: 3.0–83.2 kg) and 23.2 kg (range: 1.5–106.0 kg), respectively (p = 0.856). A large variety of breedswere represented in this study population. The healthycontrol group consisted of dogs of 39 breeds and the 3 mostcommon breeds were mixed breed (19/95 [20.0%]),Labrador Retriever [18/95 (18.9%)], and Australian Shep-herd (6/95 [6.3%]). The diseased group consisted of dogs of51 breeds and the 3 most common breeds were LabradorRetriever (17/104 [16.3%]), mixed breed (15/104 [14.4%]),and German Shepherd dogs (7/104 [6.7%]). None of thehealthy control dogs received antibiotics for at least3 months before sample collection. Of the dogs withclinical signs of GI disease, 38 dogs received antibiotics atthe time of sample collection, while 58 dogs did not receiveantibiotics at the time of sample collection, and 8 dogs hadan unknown history of antibiotic administration.

3.2. Evaluation of the stability of CPE in fecal samples

A total of 104 fecal aliquots made from 8 CPE positiveand 5 CPE negative fecal samples were analyzed toevaluate the stability of CPE over 10 days for the 3 storageconditions. All fecal aliquots were consistent with theinitial result, regardless of storage condition or time(Table 2).

3.3. Prevalence of C. perfringens (C. perfringens 16S rRNA

gene)

The C. perfringens 16S rRNA gene was detected in allsamples from either healthy control dogs or dogs withclinical signs of GI disease. The abundance (i.e., amount ofDNA) of the C. perfringens 16S rRNA gene was significantlyhigher in dogs with clinical signs of GI disease (p < 0.001)than in healthy control dogs. A subset analysis, in whichdogs with clinical signs of GI disease were divided intothree groups based on the type of diarrhea, revealed that asignificantly higher abundance of the C. perfringens 16SrRNA gene was observed in dogs with acute and chronicdiarrhea compared to the healthy control dogs(p = 0.003 and 0.010, respectively; Fig. 2A). No significantdifference was observed between control dogs and thosedogs with clinical signs of GI disease but without diarrhea,and between dogs with acute and chronic diarrhea(Fig. 2A).

3.4. Prevalence of potentially enterotoxigenic C. perfringens

(C. perfringens enterotoxin gene; cpe gene)

The prevalence of the cpe gene was 32/95 (33.7%) inhealthy dogs and 50/104 (48.1%) in dogs with clinical signsof GI disease, and was significantly different between thegroups (p = 0.044; Table 3). The abundance (i.e., amount ofDNA) of the cpe gene was significantly higher in dogs withclinical signs of GI disease than in healthy dogs (p = 0.021).A subset analysis, in which dogs with clinical signs of GIdisease were divided into 3 groups based on the type ofdiarrhea, revealed that a significantly higher abundance ofcpe gene was observed in dogs with acute diarrheacompared to healthy control dogs (p = 0.002), but nosignificant differences were observed among other groups(i.e., between healthy control, chronic diarrhea, and non-diarrheic groups; Fig. 2B). Of the samples positive for thecpe gene, 15/82 samples (18.3%; healthy control dog, n = 1;dogs with acute diarrhea, n = 4; chronic diarrhea, n = 7;non-diarrhea, n = 4) were positive for the enterotoxin(CPE), and 67/82 (81.7%) samples were negative for CPE. Asignificant positive correlation was observed between theabundances of the C. perfringens 16S rRNA gene and the cpe

gene (r = 0.428, p < 0.001; Fig. 2C).

3.5. Prevalence of C. perfringens enterotoxin (CPE)

The prevalence of CPE was significantly different(p < 0.001) between healthy control dogs and dogs withclinical signs of GI disease (1/95 [1.0%] and 17/104 [16.3%],respectively; Table 3). Fecal samples from dogs that werepositive for CPE had significantly higher abundances of

Fig. 1. Distribution of ages in healthy control dogs and dogs with signs of

GI disease. The lines represent the medians of both groups.

C. p

tho

3.6.

3.6.

wit

cocliba

dogof

Tab

Stab

Sa

1

2

3

4

5

6

7

8

9

10

11

12

13

Each

enzy

Fig.

corr

the

(p <

Tab

The

cp

cp

CP

CP

cpe-

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473 467

erfringens 16S rRNA gene and cpe gene compared tose dogs that were CPE negative (both p < 0.001; Fig. 3).

Evaluation of the fecal microbiota

1. Comparison between healthy control dogs and dogs

h clinical signs of GI disease

The abundances of Fusobacteria (p < 0.001), Rumino-caceae (p < 0.001), Blautia spp. (p < 0.001), and Faeca-

cterium spp. (p < 0.001) were significantly decreased ins with clinical signs of GI disease, while the abundancesBifidobacterium spp. (p < 0.001), Lactobacillus spp.

(p < 0.001), and E. coli (p < 0.001) were significantlyincreased compared to healthy control dogs (Fig. 4). Nosignificant difference was observed in the abundance oftotal bacteria between healthy control dogs and dogs withclinical signs of GI disease (p = 0.888; Fig. 4). A subsetanalysis, in which dogs with clinical signs of GI diseasewere divided into 3 groups based on the type of diarrhea(i.e., either acute, chronic, or non-diarrhea), revealed thatdogs with chronic diarrhea had significantly decreasedabundances of Fusobacteria (p = 0.002), Ruminococcaceae(p < 0.001), Blautia spp. (p < 0.001), and Faecalibacterium

spp. (p < 0.001) and significantly increased abundances of

le 2

ility of C. perfringens enterotoxin at various storage conditions.

mple Initial result Time point and storage condition

Day 0 Day 2 Day 5 Day 10

RT 4 8C RT 4 8C RT 4 8C �20 8C

Positive 0.765 0.631 0.770 0.477 0.388 1.241 0.574 0.589

Positive 0.181 0.495 0.338 0.408 0.215 0.269 0.393 0.274

Positive 0.258 0.602 0.371 0.413 0.297 0.466 0.525 0.464

Positive 0.609 0.646 0.820 0.802 0.248 0.531 0.882 0.563

Positive 0.556 0.756 0.574 0.621 0.519 0.492 0.611 0.694

Positive 1.956 1.371 1.501 1.968 1.934 1.077 1.722 1.840

Positive 0.843 0.947 0.864 0.549 0.510 0.492 1.097 0.749

Positive 0.220 0.275 0.426 0.457 0.263 0.530 0.224 0.394

Negative 0.045 0.048 0.042 0.041 0.059 0.069 0.051 0.048

Negative 0.048 0.099 0.046 0.045 0.047 0.061 0.055 0.060

Negative 0.045 0.049 0.069 0.045 0.051 0.046 0.048 0.081

Negative 0.046 0.041 0.049 0.063 0.053 0.044 0.065 0.050

Negative 0.050 0.053 0.082 0.041 0.063 0.049 0.057 0.048

value represents OD450 value. Sample with OD450� 0.120 was considered positive, and sample with OD450< 0.120 was considered negative. ELISA,

me-linked immunoabsorbent assay; RT, room temperature.

2. Abundances of the C. perfringens 16S rRNA gene (A) and the cpe gene (B) in healthy control dogs and dogs with clinical signs of GI disease, and

elation of the C. perfringens 16S rRNA gene and the cpe gene (C). The bottom and top of the box represent the 25th and 75th percentiles, and the line of

box represents the medians. Whiskers represent the 10th and the 90th percentile. Columns not sharing a common superscript are significantly different

0.05). ACT, dogs with acute diarrhea; CHR, dogs with chronic diarrhea; NON, dogs with clinical signs of GI disease but without diarrhea.

le 3

prevalences of enterotoxigenic C. perfringens and C. perfringens enterotoxin in feces.

Acute diarrhea

(n = 22)

Chronic diarrhea

(n = 58)

Non-diarrheic

(n = 24)

Total GI disease

(n = 104)

Healthy

(n = 95)

p-value (healthy

vs. GI disease)

e-gene positive 16 24 10 50 (48.1%) 32 (33.7%) 0.044

e-gene negative 6 34 14 54 (51.9%) 63 (66.3%)

E positive 4 8 5 17 (16.3%) 1 (1.0%) <0.001

E negative 18 50 19 87 (83.7%) 94 (99.0%)

gene, C. perfringens enterotoxin gene; CPE, C. perfringens enterotoxin; Non-diarrheic, dogs with clinical signs of GI disease but without diarrhea.

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473468

Bifidobacterium spp. (p = 0.001), Lactobacillus spp.(p < 0.001), and E. coli (p < 0.001) compared to healthycontrol dogs (Fig. 5). Dogs with acute diarrhea and dogswith clinical signs of GI disease but without diarrhea hadsignificantly decreased abundances of Ruminococcaceae(both p < 0.001), Blautia spp. (both p < 0.001), andFaecalibacterium spp. (both p < 0.001) and significantlyincreased abundances of E. coli (p < 0.001, p = 0.027,respectively) compared to healthy control dogs (Fig. 5).

3.6.2. Relationship between intestinal dysbiosis and the

presence of the cpe gene

Based on the qPCR results, both healthy control dogsand dogs with clinical signs of GI disease were classifiedinto two groups (cpe gene positive or negative; total

4 groups; Fig. 6). There were no significant differences inthe abundance of any bacterial group between dogs thatwere cpe gene positive and cpe gene negative within eitherthe healthy control group or the disease group (Fig. 6). Theresults indicated that the differences in bacterial groupsbetween diseased and healthy control dogs were generallyindependent of the presence of the cpe gene.

3.6.3. Relation between dysbiosis and the presence of CPE

According to the ELISA results, healthy control dogs anddogs with clinical signs of GI disease were classified intotwo groups (CPE positive or negative: total 3 groups). Thehealthy control dog that was positive for CPE was excludedfrom this analysis because only one sample was present inthis group. There were no significant differences in the

Fig. 3. Abundance of the C. perfringens 16S rRNA gene (A) and the cpe gene (B) in samples either positive or negative for CPE. The bottom and top of the box

represent the 25th and 75th percentiles, and the line of the box represents the median. Whiskers represent the 10th and the 90th percentile. *Significantly

different (p < 0.05) compared to the CPE negative samples. CPE neg, C. perfringens enterotoxin assay negative; CPE pos, C. perfringens enterotoxin assay

positive.

Fig. 4. Abundances of bacterial groups in healthy control dogs and dogs with clinical signs of GI disease. The bottom and top of the box represent the 25th

and the 75th percentiles, and the line of the box represents the median. Whiskers represent the 10th and the 90th percentile. *Significantly different

(adjusted p < 0.05) compared to healthy control dogs.

Fig. 5. Abundances of bacterial groups in healthy control dogs and dogs with clinical signs of GI disease. The bottom and top of the box represent the 25th

and the 75th percentiles, and the line of the box represents the median. Whiskers represent the 10th and the 90th percentile. Columns not sharing a

common superscript are significantly different (adjusted p < 0.05). ACT, dogs with acute diarrhea; CHR, dogs with chronic diarrhea; NON, dogs with clinical

signs of GI disease but without diarrhea.

Fig. 6. Relationship between the presence of the cpe gene and the abundance of bacterial groups. The bottom and top of the box represent the 25th and the

75th percentiles, and the line of the box represents the median. Whiskers represent the 10th and the 90th percentile. Columns not sharing a common

superscript are significantly different (adjusted p < 0.05). H_neg, healthy control dogs that were cpe gene negative; H_pos, healthy control dogs that were

cpe gene positive; D_neg, dogs with clinical signs of GI disease that were cpe gene negative; D_pos, dogs with clinical signs of GI disease that were cpe gene

positive.

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473 469

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473470

abundance of any bacterial groups between dogs withclinical signs of GI disease that were CPE positive and CPEnegative, indicating that differences in bacterial groupsbetween diseased and healthy control dogs were generallyindependent of the presence of CPE (Fig. 7).

3.7. Effects of antibiotics

To evaluate if the administration of antibiotics had aconfounding effect on the prevalence of enterotoxigenicC. perfringens, the prevalence of CPE, and the abundancesof bacterial groups, samples from dogs with signs of GIdisease were analyzed based on the history of antibioticadministration. No significant differences were observed

in the prevalence of enterotoxigenic C. perfringens and CPEbetween dogs with signs of GI disease that were receivingantibiotics and those that were not receiving antibiotics(p = 0.946 and 0.419, respectively). No significant differ-ences were also observed when dogs were divided into3 groups based on the type of diarrhea, and comparedbased on the history of antibiotics within each group (allp > 0.100). With regard to the effects on the abundance ofbacterial groups, no significant differences were observedin the abundances of C. perfringens 16S rRNA gene, cpe

gene, and any other bacterial group analyzed betweendogs with signs of GI disease that were receivingantibiotics and those that were not receiving antibiotics(Table 4).

Fig. 7. Relationship between the presence of CPE and the abundance of bacterial groups. The bottom and top of the box represent the 25th and the 75th

percentiles, and the line of the box represents the median. Whiskers represent the 10th and the 90th percentile. Columns not sharing a common superscript

are significantly different (adjusted p < 0.05). H_neg, healthy control dogs that were CPE negative; D_neg, dogs with signs of GI disease that were CPE

negative; D_pos, dogs with signs of GI disease that were CPE positive.

Table 4

Abundance of bacterial groups in dogs with signs of GI disease that did or did not receive antibiotics at the time of sample collection.

Target organism Dogs WITHOUT antibiotics (log

DNA)

Dogs WITH antibiotics

(log DNA)

Adjusted p-value

Median Range Median Range

C. perfringens 5.3 (1.2–7.2) 5.0 (1.0–9.6) 1.0

Enterotoxigenic C. perfringens 2.9 (0.0–8.4) 0.0 (0.0–7.8) 1.0

Fusobacteria 7.5 (5.2–9.7) 7.3 (5.7–9.5) 1.0

Ruminococcaceae 8.0 (5.2–8.6) 7.7 (6.0–9.0) 1.0

Bifidobacterium spp. 5.6 (3.7–7.5) 5.6 (3.2–8.1) 1.0

Blautia spp. 9.9 (6.5–10.7) 9.5 (6.9–10.8) 1.0

Faecalibacterium spp. 4.5 (2.3–6.8) 4.8 (1.4–7.8) 1.0

Lactobacillus spp. 6.3 (3.7–9.3) 6.4 (3.7–8.9) 1.0

E. coli 7.3 (3.3–8.4) 7.3 (3.3–8.5) 1.0

Total bacteria 10.7 (8.3–11.6) 10.9 (9.2–11.3) 1.0

4. D

det24

witdiff(Maof sourevaundandongchaass

basper

33.of GculGol200betin nandthediseof

preit udegpreall

statbeechrstubettheevawitprecondetentC. p

genind

abuandconrRNdogper

dolcorabuthe

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473 471

iscussion

A previous study reported the effects of storage on theection of CPE by RPLA in canine fecal samples stored forh at 25 8C and 4 8C. Although there were some samplesh a discordant RPLA result, overall no significanterences among storage conditions were reportedrks et al., 2000). To exclude the possibility that lacktability of CPE during long-term storage confounded

results, the stability of CPE as assessed by ELISA wasluated. The results indicate that CPE is stable in feceser the tested conditions. However, we have not tested, therefore, cannot exclude the possibility that theoing production of enterotoxin due to environmentalnges and storage may yield a false positive result asessed by ELISA.In this study, C. perfringens was detected in all dogsed on the 16S rRNA gene. Potentially, enterotoxigenic C.

fringens (i.e., harboring the cpe gene) were detected in7% of control dogs and 48.1% of dogs with clinical signs

I disease. These results are consistent with previoustivation and/or PCR-based studies (Cave et al., 2002;dstein et al., 2012; Marks et al., 2002; Weese et al.,1). In previous reports, prevalence of CPE rangedween 5% (Weese et al., 2001) and 16% (Cave et al., 2002)on-diarrheic dogs, and between 14% (Cave et al., 2002)

41% in diarrheic dogs (Kruth et al., 1989). In our study, prevalence of CPE in dogs with clinical signs of GIase was similar with 16.3%. In contrast, the prevalence

CPE in control dogs (1.0%) was slightly lower thanviously reported. The results of our stability study makenlikely that this lower prevalence was due to sampleradation. Another possible explanation for lowervalence may be due to geographical differences becausesamples from control dogs were collected only in onee (i.e., Texas). In humans, the important role of CPE hasn highlighted in acute diarrhea, but the role of CPE inonic diarrhea has been rarely reported. In our currentdy, the prevalence of CPE was not significantly differentween acute diarrhea and chronic diarrhea. Therefore, role of CPE in chronic diarrhea warrants furtherluation. To conclude the first part of this study, dogsh clinical signs of GI disease had a significantly highervalence of enterotoxigenic C. perfringens compared totrol dogs. However, its virulence factor CPE wasected in only 18.3% of dogs that were positive forerotoxigenic strains. This indicates that the detection oferfringens by PCR for 16S rRNA gene and enterotoxi-ic C. perfringens by PCR for the cpe gene is not alwaysicative of the presence of CPE in dogs with diarrhea.The quantitative analysis revealed significantly higherndances of both genes (16S rRNA gene for C. perfringens

cpe gene) in dogs with acute diarrhea compared to thetrol group. A recent study using sequencing of 16SA gene showed changes in the fecal microbiome ins with acute diarrhea, and significant increases of C.

fringens in dogs with hemorrhagic diarrhea (Sucho-ski et al., 2012c). In the current study, we evaluated therelation between these genes, and found that thendance of the cpe gene was positively correlated with

abundance of the C. perfringens 16S rRNA gene. These

findings may suggest that increases in populations ofenterotoxigenic C. perfringens are associated with acutediarrhea, but also that such an increase may be simply dueto the increased overall population of C. perfringens withconcurrent reduction of commensal microbiota (i.e.,intestinal dysbiosis) as indicated by results of qPCR assays.Our results are also consistent with previous studiesshowing significant changes in abundances of severalbacterial groups in dogs with GI diseases (Suchodolskiet al., 2008, 2012a,b,c; Xenoulis et al., 2008). Evaluatingchanges in the intestinal microbiota in GI disease areimportant for better understanding of disease pathogene-sis. It has been well documented that the intestinalmicrobiota plays a crucial role in intestinal health, and thatreduction of normal protective microbiota may confersusceptibility to intestinal inflammation. Prolonged imbal-ances of the GI microbiota may result in a dysregulation ofimmune responses and reduced activity against infection(Round and Mazmanian, 2009). Furthermore, recent datasuggest that the composition of intestinal microbiota andits associated metabolite profile is an important factor foractivation of virulence genes of some enteropathogens. Forexample, a study using a mouse model showed thatspecific patterns in intestinal microbial communities had adirect effect on the pathogenicity of Salmonella (Bearsonet al., 2013). The exact mechanisms behind this interplaybetween commensal bacteria and virulence factors ofenteropathogens have not been well elucidated. Recentreports have associated intestinal dysbiosis and changes inbacterial metabolisms (e.g., altered SCFA and bile acidprofiles) with the activation of toxin production in patientswith C. difficile infection (Antharam et al., 2013). There isevidence that alterations in intestinal bile acid composi-tion enhance germination of C. difficile, increasing suscep-tibility to infection (Theriot et al., 2014; Weingarden et al.,2014). This suggests an important cross-talk betweenenteropathogens and commensals.

We observed an association between the abundance ofenterotoxigenic C. perfringens and fecal dysbiosis, as theabundances of C. perfringens 16S rRNA gene and cpe genewere higher in dogs with diarrhea and these dogs had adysbiosis manifested as reduction of several commensalbacterial groups. Furthermore, there was a strong associa-tion between the presence of CPE and GI disease. This maysuggest that the increased abundance of C. perfringens andenterotoxigenic C. perfringens may be part of dysbiosis andmay not necessarily play a primary pathological role indiarrhea. However, it may also be possible that initialdysbiosis due to various causes and subsequent changes inbacterial metabolite profiles within the intestinal lumenmay trigger the production of CPE from enterotoxigenic C.

perfringens, as has been demonstrated with C.

difficile. Further studies evaluating the activity of tran-scription of CPE are warranted. Clearly, the relationshipsbetween enteropathogens, fecal altered microbial com-munities, and altered bacterial metabolite profiles deservefurther studies in dogs with GI disease.

There are several observations in this study thatdeserve discussion. There were six samples from non-diarrheic dogs that were positive for CPE (five dogs withclinical signs of GI disease but without diarrhea and one

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473472

healthy control dog). Possible explanations for this are thatthere may be a threshold amount of enterotoxin requiredto cause diarrhea in dogs. For example, in food poisoning inhumans due to enterotoxin produced by C. perfringens typeA, a certain bacterial load is needed to induce disease(Brynestad and Granum, 2002). Other virulence factors ofC. perfringens may play a role in canine diarrhea. Finally, itis important to note that the assay used here haslimitations. The ELISA has not been properly validatedfor canine fecal samples. Therefore, potentially falsepositive results due to unspecific binding, and potentiallyfalse negative results due to low binding affinity of theantibodies utilized toward canine isolates of C. perfringens

should be considered. Of all samples that were positive forthe cpe gene, 67/82 (81.7%) samples were negative for CPE.This may suggest that abundance of CPE was below thedetection limit of the ELISA assay or enterotoxigenic C.

perfringens presented in feces, but enterotoxin was notproduced due to the lack of spore formation. The latterexplanation could be validated by quantifying spores.However, previous studies showed no statistical correla-tion between the presence of CPE and spore counts in feces(Marks et al., 1999; Weese et al., 2001). Therefore, furtherstudies investigating the CPE expression by evaluatingRNA levels may provide the activity of CPE production.Finally, of samples negative for the cpe gene, 3/117 (2.5%)were positive for CPE. This suggests unspecific cross-reactivity in the ELISA assay. Alternatively, this findingmay have been due to a low sensitivity of the qPCR assay.

There are some limitations to this study. There was asignificant age difference between control dogs and dogswith clinical signs of GI disease. However, as Fig. 1 shows,the age range of control group did overlap with that of thedisease group, thus, the statistically significant difference inage may not have been a major bias in this study. Asmentioned previously, the CPE ELISA assay has not beenvalidated for use in dogs, and its sensitivity and specificityare unknown. In the current study, we evaluated only one ofthe C. perfringens virulence factors and its encoding gene.Other virulence factors such as C. perfringens b2 toxin genehave previously been detected in diarrheic dogs (Goldsteinet al., 2012; Thiede et al., 2001), and a dog with fatal acutehemorrhagic gastroenteritis (Schlegel et al., 2012). Recentstudies have identified three types of cpe loci organizationsin C. perfringens type Aisolate, and have suggested that therewere differences in the pathogenesis between these types(Miyamoto et al., 2012). Therefore, further evaluation ofthese genes may provide more information about theclinical significance of the isolated strains. The time of fecalsample collection after onset of diarrhea varied with eachsample in our study. This might confound the prevalence ofCPE because detection of CPE is much more reliable whenfeces are collected early after the onset of diarrhea inhumans. Lastly, we did not evaluate the presence of otherpotential enteric pathogens such as Campylobacter andSalmonella. Therefore, it is unknown that the cause ofclinical signs was purely due to pathological effect of CPE.

In conclusion, increased abundances of C. perfringens

and enterotoxigenic C. perfringens were observed in dogswith clinical signs of GI disease, and these were most

presence of CPE was associated with GI disease, thepresence of this organism, either non- or enterotoxigenicstrain, was not indicative of GI disease in all cases ofdiarrhea. On the other hand, dysbiosis was significantlyassociated with GI disease. Therefore, the increasedabundance of C. perfringens and enterotoxigenic C.

perfringens may be part of intestinal dysbiosis. It remainsunknown whether dysbiosis is a cause or result of GIdisease, and how dysbiosis affects enterotoxin productionof C. perfringens. However, regardless of the initial cause ofdysbiosis, an abnormal microbiota may exacerbate GIdisease or may lead to metabolic changes in the intestinallumen that may favor the activation of bacterial virulencegenes in dogs with GI disease. These findings enlighten theimportance of balanced microbial communities for GIhealth, and further evaluation of intestinal dysbiosis indogs with GI disease is warranted.

References

Abrahao, C., Carman, R.J., Hahn, H., Liesenfeld, O., 2001. Similar frequencyof detection of Clostridium perfringens enterotoxin and Clostridiumdifficile toxins in patients with antibiotic-associated diarrhea. Eur. J.Clin. Microbiol. 20, 676–677.

Antharam, V.C., Li, E.C., Ishmael, A., Sharma, A., Mai, V., Rand, K.H., Wang,G.P., 2013. Intestinal dysbiosis and depletion of butyrogenic bacteriain Clostridium difficile infection and nosocomial diarrhea. J. Clin.Microbiol. 51, 2884–2892.

Bearson, S.M., Allen, H.K., Bearson, B.L., Looft, T., Brunelle, B.W., Kich, J.D.,Tuggle, C.K., Bayles, D.O., Alt, D., Levine, U.Y., Stanton, T.B., 2013.Profiling the gastrointestinal microbiota in response to Salmonella:low versus high Salmonella shedding in the natural porcine host.Infect. Genet. Evol. 16C, 330–340.

Brynestad, S., Granum, P.E., 2002. Clostridium perfringens and foodborneinfections. Int. J. Food Microbiol. 74, 195–202.

Cave, N.J., Marks, S.L., Kass, P.H., Melli, A.C., Brophy, M.A., 2002. Evaluationof a routine diagnostic fecal panel for dogs with diarrhea. J. Am. Vet.Med. Assoc. 221, 52–59.

Craven, M., Egan, C.E., Dowd, S.E., McDonough, S.P., Dogan, B., Denkers,E.Y., Bowman, D., Scherl, E.J., Simpson, K.W., 2012. Inflammationdrives dysbiosis and bacterial invasion in murine models of ilealCrohn’s disease. PLoS ONE 7, e41594.

Duboc, H., Rajca, S., Rainteau, D., Benarous, D., Maubert, M.A., Quervain, E.,Thomas, G., Barbu, V., Humbert, L., Despras, G., Bridonneau, C.,Dumetz, F., Grill, J.P., Masliah, J., Beaugerie, L., Cosnes, J., Chazouilleres,O., Poupon, R., Wolf, C., Mallet, J.M., Langella, P., Trugnan, G., Sokol, H.,Seksik, P., 2013. Connecting dysbiosis, bile-acid dysmetabolism andgut inflammation in inflammatory bowel diseases. Gut 62, 531–539.

Goldstein, M.R., Kruth, S.A., Bersenas, A.M., Holowaychuk, M.K., Weese,J.S., 2012. Detection and characterization of Clostridium perfringens inthe feces of healthy and diarrheic dogs. Can. J. Vet. Res. 76, 161–165.

Gurjar, A.A., Hegde, N.V., Love, B.C., Jayarao, B.M., 2008. Real-time multi-plex PCR assay for rapid detection and toxintyping of Clostridiumperfringens toxin producing strains in feces of dairy cattle. Mol. Cell.Probes 22, 90–95.

Hall, E.J., 2011. Antibiotic-responsive diarrhea in small animals. Vet. Clin.North Am. Small Anim. Pract. 41, 273–286.

Kruth, S.A., Prescott, J.F., Welch, M.K., Brodsky, M.H., 1989. Nosocomialdiarrhea associated with enterotoxigenic Clostridium perfringens in-fection in dogs. J. Am. Vet. Med. Assoc. 195, 331–334.

Lahti, P., Heikinheimo, A., Johansson, T., Korkeala, H., 2008. Clostridiumperfringens type A strains carrying a plasmid-borne enterotoxin gene(genotype IS1151-cpe or IS1470-like-cpe) as a common cause of foodpoisoning. J. Clin. Microbiol. 46, 371–373.

Marks, S.L., Kather, E.J., Kass, P.H., Melli, A.C., 2002. Genotypic andphenotypic characterization of Clostridium perfringens and Clostridiumdifficile in diarrheic and healthy dogs. J. Vet. Intern. Med. 16, 533–540.

Marks, S.L., Melli, A., Kass, P.H., Jang, S.S., Barkhoodarian, A., Hirsh, D.C.,1999. Evaluation of methods to diagnose Clostridium perfringens-associated diarrhea in dogs. J. Am. Vet. Med. Assoc. 214, 357–360.

Marks, S.L., Melli, A., Kass, P.H., Jang, S.S., Barkhoodarian, A., Hirsh, D.C.,2000. Influence of storage and temperature on endospore and en-

terotoxin production by Clostridium perfringens in dogs. J. Vet. Diagn.Invest. 12, 63–67. apparent in dogs with acute diarrhea. Although, the

Mar

Mcc

Miy

Pack

Ross

Rou

Schl

Such

Such

Such

Such

Y. Minamoto et al. / Veterinary Microbiology 174 (2014) 463–473 473

ks, S.L., Rankin, S.C., Byrne, B.A., Weese, J.S., 2011. Enteropathogenicbacteria in dogs and cats: diagnosis, epidemiology, treatment, andcontrol. J. Vet. Intern. Med. 25, 1195–1208.lane, B.A., 1992. Clostridium perfringens enterotoxin – structure, actionand detection. J. Food Saf. 12, 237–252.amoto, K., Li, J., McClane, B.A., 2012. Enterotoxigenic Clostridiumperfringens: detection and identification. Microbes Environ. 27,343–349.ey, C.D., Sartor, R.B., 2009. Commensal bacteria, traditional and

opportunistic pathogens, dysbiosis and bacterial killing in inflamma-tory bowel diseases. Curr. Opin. Infect. Dis. 22, 292–301.i, G., Pengo, G., Caldin, M., Palumbo Piccionello, A., Steiner, J.M.,

Cohen, N.D., Jergens, A.E., Suchodolski, J.S., 2014. Comparison ofmicrobiological, histological, and immunomodulatory parametersin response to treatment with either combination therapy withprednisone and metronidazole or probiotic VSL#3 strains in dogswith idiopathic inflammatory bowel disease. PLOS ONE 9, e94699.nd, J.L., Mazmanian, S.K., 2009. The gut microbiota shapes intestinalimmune responses during health and disease. Nat. Rev. Immunol. 9,313–323.egel, B.J., Van Dreumel, T., Slavic, D., Prescott, J.F., 2012. Clostridiumperfringens type A fatal acute hemorrhagic gastroenteritis in a dog.Can. Vet. J. 53, 555–557.odolski, J., Dowd, S.E., Steiner, J.M., Jergens, A., 2012a. High-through-

put 454 pyrosequencing analysis reveals dysbiosis of the mucosa-associated microbiota in dogs with inflammatory bowel disease.Gastroenterology 142, S678–S679.odolski, J.S., Camacho, J., Steiner, J.M., 2008. Analysis of bacterial

diversity in the canine duodenum, jejunum, ileum, and colon bycomparative 16S rRNA gene analysis. FEMS Microbiol. Ecol. 66,567–578.odolski, J.S., Dowd, S.E., Wilke, V., Steiner, J.M., Jergens, A.E., 2012b.

16s rRNA gene pyrosequencing reveals bacterial dysbiosis in theduodenum of dogs with idiopathic inflammatory bowel disease. PloSONE 7.odolski, J.S., Markel, M.E., Garcia-Mazcorro, J.F., Unterer, S.,

Heilmann, R.M., Dowd, S.E., Kachroo, P., Ivanov, I., Minamoto, Y.,

Dillman, E.M., Steiner, J.M., Cook, A.K., Toresson, L., 2012c. The fecalmicrobiome in dogs with acute diarrhea and idiopathic inflammatorybowel disease. PloS ONE 7, e51907.

Theriot, C.M., Koenigsknecht, M.J., Carlson Jr., P.E., Hatton, G.E., Nelson,A.M., Li, B., Huffnagle, G.B.J., Young, Z.L.V.B., 2014. Antibiotic-inducedshifts in the mouse gut microbiome and metabolome increase sus-ceptibility to Clostridium difficile infection. Nat. Commun. 5, 3114.

Thiede, S., Goethe, R., Amtsberg, G., 2001. Prevalence of beta2 toxin geneof Clostridium perfringens type A from diarrhoeic dogs. Vet. Rec. 149,273–274.

Tupler, T., Levy, J.K., Sabshin, S.J., Tucker, S.J., Greiner, E.C., Leutenegger,C.M., 2012. Enteropathogens identified in dogs entering a Floridaanimal shelter with normal feces or diarrhea. J. Am. Vet. Med. Assoc.241, 338–343.

Unterer, S., Busch, K., Leipig, M., Hermanns, W., Wolf, G., Straubinger, R.K.,Mueller, R.S., Hartmann, K., 2014. Endoscopically visualized lesions,histologic findings, and bacterial invasion in the gastrointestinalmucosa of dogs with acute hemorrhagic diarrhea syndrome. J. Vet.Intern. Med. 28, 52–58.

Watanabe, M., Hitomi, S., Sawahata, T., 2008. Nosocomial diarrhea causedby Clostridium perfringens in the Tsukuba-Tsuchiura district, Japan. J.Infect. Chemother. 14, 228–231.

Weese, J.S., Staempfli, H.R., Prescott, J.F., Kruth, S.A., Greenwood, S.J.,Weese, H.E., 2001. The roles of Clostridium difficile and enterotoxigenicClostridium perfringens in diarrhea in dogs. J. Vet. Intern. Med. 15,374–378.

Weingarden, A.R., Chen, C., Bobr, A., Yao, D., Lu, Y., Nelson, V.M., Sadowsky,M.J., Khoruts, A., 2014. Microbiota transplantation restores normalfecal bile acid composition in recurrent Clostridium difficile infection.Am. J. Physiol. Gastrointest. Liver Physiol. 306, G310–G319.

Wise, M.G., Siragusa, G.R., 2005. Quantitative detection of Clostridiumperfringens in the broiler fowl gastrointestinal tract by real-time PCR.Appl. Environ. Microbiol. 71, 3911–3916.

Xenoulis, P.G., Palculict, B., Allenspach, K., Steiner, J.M., Van House, A.M.,Suchodolski, J.S., 2008. Molecular-phylogenetic characterization ofmicrobial communities imbalances in the small intestine of dogs withinflammatory bowel disease. FEMS Microbiol. Ecol. 66, 579–589.