GUT MICROBES 2020, VOL. 12, NO. 1, e1785246 (15 pages) https://doi.org/10.1080/19490976.2020.1785246 RESEARCH PAPER Dysbiosis in a canine model of human fistulizing Crohn’s disease Ana Maldonado-Contreras a, Lluís Ferrerb*, Caitlin Cawleya, Sarah Crainb#, Shakti Bhattaraia, Juan Toscanoa&, Doyle V. Ward a, and Andrew Hoffmanb$ a Department of Microbiology and Physiological Systems, University of Massachusetts Medical School, Worcester, MA, USA; bCummings School of Veterinary Medicine, Tufts University, North Grafton, MA, USA ABSTRACT ARTICLE HISTORY Crohn’s disease (CD) is a chronic immune-mediated inflammatory condition caused by the loss of Received 28 October 2019 mucosal tolerance toward the commensal microbiota. On average, 29.5% and 42.7% CD patients Revised 17 April 2020 experience perianal complications at 10 and 20 y after diagnosis, respectively. Perianal CD (pCD) Accepted 8 June 2020 result in high disease burden, diminished quality of life, and elevated health-care costs. Overall pCD KEYWORDS are predictors of poor long-term outcomes. Animal models of gut inflammation have failed to fully fistulizing Crohn’s disease; recapitulate the human manifestations of fistulizing CD. Here, we evaluated dogs with spontaneous microbiome; dysbiosis; canine anal furunculosis (CAF), a disease with clinical similarities to pCD, as a surrogate model for perianal fistulas; canine understanding the microbial contribution of human pCD pathophysiology. furunculosis By comparing the gut microbiomes between dogs suffering from CAF (CAF dogs) and healthy dogs, we show CAF-dog microbiomes are either very dissimilar (dysbiotic) or similar (healthy-like), yet unique, to healthy dog’s microbiomes. Compared to healthy or healthy-like CAF microbiomes, dysbiotic CAF microbiomes showed an increased abundance of Bacteroides vulgatus and Escherichia coli and a decreased abundance of Megamonas species and Prevotella copri. Our results mirror what have been reported in previous microbiome studies of patients with CD; particularly, CAF dogs exhibited two distinct microbiome composition: dysbiotic and healthy-like, with determinant bacterial taxa such as E. coli and P. copri that overlap what it has been found on their human counterpart. Thus, our results support the use of CAF dogs as a surrogate model to advance our understanding of microbial dynamics in pCD. Introduction pCD, specifically PRDM1, NOD2, and ATG16L1.4,5 Immunologically, patients with pCD exhibit an up- Inflammatory bowel diseases (IBD), including regulated expression of TNF-α compared to CD Crohn’s disease (CD), are immune-mediated inflam­ patients without perianal complications or healthy matory condition affecting more than a million indi­ controls.6 The role of the microbiome has gained viduals in the US. On average, 29.5% and 42.7% CD attention in recent years as evidence indicating its patients experience perianal complications at 10 and involvement emerges.7–9 Particularly, treatment with 20 y after diagnosis, respectively.1,2 Perianal Crohn’s antibiotics (reviewed in2,10), fecal diversion,7,8 and disease, or pCD, represents one of the most clinically fecal transplants9 have proven to be effective for significant complications and is a predictor of poor managing the disease. Definition of the pathogenesis long-term outcome in CD.3 Perianal fistulas are and pathophysiology of pCD is key to further most common in patients suffering from colonic advance treatment strategies for patients. One of CD with rectal involvement (L2). The etiology of the main challenges of pCD research is the lack of pCD is not well defined yet is thought to be a relevant, reliable, and reproducible animal model. a combination of genetic and immune factors, and Here, we aim to evaluate the usefulness of dogs as the microbiome. Several genetic variants have been a model system to elucidate the contribution of the also associated with increased risk of development of microbiota to the pathophysiology of pCD. CONTACT Ana Maldonado-Contreras ana.maldonado@umassmed.edu Department of Microbiology and Physiological Systems, University of Massachusetts Medical School, Worcester, MA, 01655, USA *Present address: Departament de Medicina i Cirurgia Animals Veterinary School, Universitat Autonoma de Barcelona, Barcelona, Spain, 08193 # Present address: Ocean State Veterinary Specialists & Present address: East Greenwich, RI, 02818; Cedars-Sinai Medical Center, Los Angeles, CA, 90048 $ Present address: School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 19104 © 2020 The Author(s). Published with license by Taylor & Francis Group, LLC. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. e1785246-2 A. MALDONADO-CONTRERAS ET AL. We have postulated that canine anal furunculosis with CAF.21,26-28 Dietary treatment with “novel pro­ (CAF) can be used as a model for studying the tein” -refer to foods with protein sources not con­ pathophysiology and treatment of pCD.11,12 CAF sumed before by the dog with CAF- such as venison, and pCD exhibit common clinical signs such as lamb, fish, bison, duck-show modest results in CAF focal or multifocal dissecting sinus tracts (external treatment when used in combination with immuno­ openings) and ulcerations on the perianal area. The suppressants, antibiotics, and/or surgery.21,26-28 American Gastroenterological Association pro­ Despite of having positive clinical responses to posed a classification of “simple” and “complex” the aforementioned treatments, the majority of both fistulas13 in pCD. Simple fistulas include superficial, pCD and CAF patients experience relapses of the intersphincteric, or intrasphincteric fistulas below perinal manifestations. the dentate line, with a single tract, and absence of Of note, the German Shepherd Dog (GSD) appears perianal complications. Complex fistulas stretch to be overrepresented within the population of dogs above the dentate line (intersphincteric, trans­ with CAF, with more than 80% of dogs suffering from phincteric, extrasphincteric, suprasphincteric), CAF (CAF dogs) being GSD.29,30 GSD is also suscep­ with many tracts, and may be related to perianal tible to inflammatory bowel disease (IBD),31,32 sys­ abscesses, rectal stricture, proctitis, or can extend to temic aspergillosis,33–36 and deep pyoderma,37,38 the bladder or vagina.1,14 Similarly, CAF includes suggesting that GSD might have a broadly dysfunc­ epithelial-lined single or multiple (such as complex tional immune response to microbial exposure at fistulas in pCD) tracts that develop in the perianal epithelial surfaces, similar to human CD patients. tissue including superficial, intersphincteric, or Dogs have already proven useful for studying various intrasphincteric fistulas that can extend to the rectal other spontaneously occurring disorders similar to lumen15,16 but not to other areas (compared to those affecting humans,11,39-42 and the dog micro­ simple fistulas in pCD). biome is more similar to that of humans than it is to Treatment for both CAF and pCD requires the microbiomes of other animals that are generally a combination of immunosuppressive drugs (i.e., used as models of microbiome-centered diseases such cyclosporin, and tacrolimus),17 adjuvant antibiotic as IBD (e.g., mice and pigs).43 Dogs also sponta­ therapy, and/or surgery (i.e., fistulectomy, fistulot­ neously develop IBD31 and can experience common omy, drainage). Cyclosporin has been effective in disease complications, including perianal fistulas,16 managing both pCD and CAF with up to 78% and and exhibit IBD-associated microbiome alterations 100% of human (reviewed18) and dogs (reviewed19) that resemble changes observed in humans with responding to treatment, respectively. In addition, IBD.44,45 Thus, dogs represent an ideal model for positive clinical response to tacrolimus is similar in studying IBD and in particular pCD. pCD18 and dogs with CAF19 (44–78% and 50%, Given that the intersection of immunity and the respectively). Antimicrobials alone do not seem microbiota seems to be at the heart of both pCD effective for the management of pCD nor for CAF. and CAF, we sought to explore naturally occurring In humans, antibiotics are administrated as initial CAF as a surrogate model for human pCD. The aim treatment or in combination with other therapies of this study was to characterize the bacterial such as immunosuppressants.2 A few studies have microbiome structure of dogs suffering from CAF. reported the use of metronidazole in combination We posit that characterizing the microbiomes of with other therapies (azathioprine, tacrolimus oint­ CAF dogs will help clarify the pathophysiology ment, prednisone) for the treatment of CAF.20,21 of CAF and advise on the translational significance Treatment with diet for both pCD and CAF is still of CAF as an animal model for pCD. in debate. A dietary guidance for patients with IBD has been recently published highlighting the lack of strong evidence for foods to include or avoid for patients with Results and discussion IBD.22 Most IBD-friendly whole food diets involved CAF is associated with gut microbiota changes exclusion of carbohydrates, inclusion of lean protein, fruits, and vegetables.23–25 Similarly, few studies have A total of 20 dogs were recruited for this study: eight demonstrated adverse reaction of foods in some dogs healthy control dogs (HC dogs), and 12 CAF dogs GUT MICROBES e1785246-3 (Table 1). We collected 4–6 fecal samples per dog, (Figure 1A, Shannon diversity, Kruskal–Wallis test, totaling 70 CAF samples and 38 HC samples. p value = .06), similar to previous results with IBD Shotgun sequencing of the samples revealed the pre­ dogs.46 As expected, there was also generally lower sence of the following bacterial phyla (relative abun­ alpha diversity in older dogs compared to younger dance indicated as percentage of total dogs (Spearman correlation, p value = .1). No sig­ reads): Firmicutes (48%), Bacteroidetes (31%), nificant differences in alpha diversity were observed Proteobacteria (12%), Actinobacteria (7%), and by weight. Fusobacteria (2%). When we analyzed differences After filtering out low abundance genera (i.e., in microbial composition between dogs according genera not detected four times or more in at least to their disease status, we found slightly lower 20% of the samples), we performed gneiss analysis,47 alpha diversity in CAF dogs compared to HC dogs which applies the concept of balance trees to Table 1. Characteristics of the recruited dogs according to disease status and cluster. HC dogs (n = 8, 38 CAF dogs (n = 12, 70 CAF dogs, near cluster (n = 7, CAF dogs, far cluster (n = 5, 43 samples) samples) 65 samples) samples) Dogs description Average age (years) ± STDEV 7.3 ± 1.43 4.3 ± 1.28 7.9 ± 1.53 7 ± 0.89 Average weight (kg) ± STDEV 36.9 ± 7.93 41.1 ± 7.31 42.4 ± 9.37 39.2 ± 2.77 Female sex 4 (50%) 3 (25%) 1 (14.3%) 2 (40%) Number of fistulas 0–2 0 4 3 1 3–5 0 5 3 2 >5 0 3 0 3 Pathology findings Morphological features ND Crypt hyperplasia 3 (25%) 1 (14.3%) 2 (40%) Crypt dilation/distortion (% positive) 8 (66.7%) 5 (71.4%) 3 (60%) Fibrosis/atrophy (% positive) 8 (66.7%) 5 (71.4%) 3 (60%) Inflammation ND Lamina Propria lymphocytes and plasma 3 (25%) 1 (14.3%) 2 (40%) cells (% positive) Lamina Propria Eosinophils (% positive) 4 (33.3%) 2 (28.6%) 2 (40%) Final diagnosis ND Normal mucosal (% positive) 3 (25%) 2 (28.6%) 1 (20%) Lymphoplasmatic inflammatory 2 (16.7%) 0 2 (40%) (% positive) Eosinophilic inflammatory (% positive) 4 (33.3%) 2 (28.6%) 2 (40%) Mucosal atrophy/fibrosis 5 (41.7%) 3 (42.9%) 2 (40%) (non-inflammatory) (% positive) s d ro cola eiu rmis gale sifie cop leb nifo rme las sp s fu ype unc ide s ro ide ona sh s ero cte gam ona ona B act Ba Me gam gam Me Me Figure 1. Analysis of the microbiota composition in samples from CAF dogs and HC dogs. A) Shannon index of alpha diversity for HC dogs (green) and CAF dogs (orange). B) Relative abundance of the bacterial taxa found to discriminate between dogs with different disease status: HC dogs (green) and CAF dogs (orange); thick black lines indicate mean abundance. T-test, p values are expressed as: ****<0.0001, ***<0.001, and #0.1 (Corrected for multiple comparisons (5 bacterial taxa found significantly different by disease status) using the Bonferroni-Dunn method, with alpha = 0.05). e1785246-4 A. MALDONADO-CONTRERAS ET AL. compositional data to identify microbial subcommu­ (Supplementary Figure 1) based on the Bray nities that covary with environmental variables. The Curtis dissimilarity index calculated from the rela­ gneiss analysis considers the log-ratio abundances of tive abundance of each bacterial taxa (n = 61). subcommunities within the microbiome to indicate Clustering results are visualized the with non- taxa whose abundances change relative to other taxa metric multidimensional scaling (NMDS). As seen with respect to a variable of interest; in our analysis, in Figure 2A, CAF dogs were divided into two the variable of interest was disease status (CAF ver­ clusters: healthy-like cluster, referred to hereafter sus HC). We used linear mixed-effect models to as “near cluster,” gathering in close proximity to account for inter-individual microbiome variability HC dogs; and dysbiotic cluster, referred to hereafter in this analysis. When we analyzed the microbiomes as “far cluster,” positioned significantly more dis­ according to disease status, we found that tant from the HC dogs (Bray Curtis distance, Megamonas hypermegale, Megamonas unclassified, PERMANOVA, R2 = 0.252, p value = .001. Bacteroides coprocola, Megamonas funiformis, and Supplementary Figure 2). Although the average Bacteroides plebeius were differentially abundant in silhouette width in this analysis indicated a weak CAF dogs compared to HC dogs (Table 2). We structure, this differentiation has been repeatedly further confirmed that there were significant differ­ reported in patients suffering from CD.44,48-50 ences in the proportions of these five bacterial taxa Taken together, these data suggest that the micro­ between HC dogs and CAF dogs, with lower abun­ biota of CAF dogs can range from one that closely dance in CAF dogs (Figure 1B, false discovery rate resembles that of a healthy individual to a dysbiotic [FDR], t-test). We also applied analysis of the com­ microbiota that differs markedly from that of position of microbiomes (ANCOM) to identify dis­ a healthy individual, similar what it was been criminant bacterial taxa for CAF dogs versus HC observed repeatedly in CD patients. dogs. We identified M. funiformis (W = 51), To identify determinant taxa associated with M. hypermegale (W = 52), and M. unclassified each cluster, we performed additional analyses (W = 60) as discriminating between CAF dogs and with gneiss and ANCOM. We found that CAF HC dogs. Altogether, these two robust analyses con­ dogs belonging to the near cluster had different curred that the abundance of Megamonas species can dominance of M. funiformes and Megamonas discriminate between HC dog samples and CAF-dog unclassified compared to HC dogs (Table 2). samples. Moreover, ANCOM results identified P. copri (W = 53) as a bacterial taxon that differentiated between CAF dogs belonging to the near cluster Dogs suffering from CAF exhibit a dichotomous and HC dogs. bacterial community structure When comparing CAF dogs belonging to the far We, next, investigated the microbiome patterns cluster and HC dogs, both gneiss and ANCOM further. Using partitioning around medoids with identified B. coprocola (W = 52), M. funiformis the estimation of the number of clusters (PAMK), (W = 49), M. hypermegale (W = 48), and we determined the optimal number of clusters M. unclassified (W = 52) as distinguishing between Table 2. Bacterial taxa with significantly different dominance between samples grouped by disease status or cluster. Comparison Determinant bacterial taxa gneiss, FDR-corrected p value Group showing increased abundance HC dogs versus CAF dogs Bacteroides plebeius 0.00137614 HC dogs Megamonas hypermegale Bacteroides coprocola Megamonas unclassified Megamonas funiformis HC dogs versus CAF dogs in near cluster Megamonas funiformis 0.01357096 HC dogs Megamonas unclassified CAF dogs in near cluster versus CAF dogs in far cluster Bacteroides plebeius 0.00001034 CAF dogs in near cluster Megamonas hypermegale Bacteroides coprocola Megamonas unclassified Megamonas funiformis GUT MICROBES e1785246-5 Figure 2. Analysis of the microbiota clusters. A) NMDS visualization of samples from HC and CAF dogs. Bacterial taxa present were quantified by MetaPhlAn, distances were calculated using Bray Curtis dissimilarity index, and samples were plotted based on NMDS. Clusters were defined using PAMK and are colored green (HC dogs), orange (CAF dogs in the near cluster) and red (CAF dogs in the far cluster). Relative abundance of the bacterial taxa found to discriminate between dogs, separated by cluster assignment: B) Bacterial taxa overrepresented in HC dogs or CAF dogs in the near cluster compared to CAF dogs in the far cluster; C) Bacterial taxa overrepresented in CAF dogs in the far cluster compared to HC dogs or CAF dogs in the near cluster. Clusters are colored as described for panel A, and thick black lines indicate mean abundance. FDR t-test, p values are expressed as: ****<0.0001, ***<0.001, ** <0.005, * <0.05, and #0.1 (Corrected for multiple comparisons (8 bacterial taxa found significantly different among clusters and HC) using the Bonferroni-Dunn method, with alpha = 0.05). the two groups (Table 2). One additional bacterial found that the two clusters differed in the domi­ taxon was identified as being differentially domi­ nance of M. hypermegale, M. unclassified, nant between CAF dogs belonging to the far cluster B. coprocola, M. funiformis, and B. plebeius and HC dogs only by gneiss: B. plebeius. (Table 2). Moreover, ANCOM results identified We also investigated differences between CAF six bacterial taxa that distinguished between the dogs belonging to the near and far clusters. We near and far clusters, specifically, B. coprocola e1785246-6 A. MALDONADO-CONTRERAS ET AL. (W = 50), B. plebeius (W = 51), Bacteroides vulgatus Figure 3A). Moreover, Helicobacter canis domi­ (W = 52), P. copri (W = 53), Megamonas unclassi­ nance was statistically different between CAF dogs fied (W = 48), and Escherichia coli (W = 48). according to fistula number (gneiss, FDR corrected Particularly, we found that the relative abundance p value <.0001). To determine the direction of of six of those bacterial taxa decreased (Figure 2B) dominance, we performed correlation analysis while two increased (Figure 2C) in the far cluster between H. canis relative abundance and number compared to HC dogs and the CAF dogs in the near of fistulas. We found a negative correlation between cluster. H. canis relative abundance and the number of Overall, these analyses established that higher fistulas in CAF dogs (Table 3). abundance of Megamonas spp., P. copri, It has been suggested that increased host DNA B. coprocola, and B. plebeius is associated with content in fecal samples may be an indicator of healthy and healthy-like (near cluster CAF-dog) inflammation-associated shedding of epithelial or samples. On the other hand, increased abundance blood cells. Indeed, a previous study found high of E. coli and B. vulgatus is associated with dysbiosis levels of human DNA in stool samples from pedia­ (far cluster CAF dogs). tric patients with active CD.44 Thus, we determined the abundance of dog DNA sequences present in each stool sample analyzed as a percentage of total Clinical manifestations of CAF are associated with reads. We found that samples from CAF dogs con­ dysbiosis tained a higher percentage (mean = 4.09%, mini­ CAF has a clinical appearance similar to that of mum = 0.02%, maximum = 44.68%) of host DNA perianal fistulas in humans, a complication fre­ sequences than did samples from HC dogs quently associated with CD. Specifically, dogs suf­ (Mean = 0.41%, minimum = 0.02%, maxi­ fering from CAF develop epithelial-lined sinus mum = 2.38%, Mann–Whitney test, p value = .001, tracts around the perianal tissue including super­ Figure 3B). There was no significant difference ficial, intersphincteric, or intrasphincteric fistulas between the abundance of host DNA sequences that can extend to the rectal lumen.15,16 These recovered in samples from the near versus far clus­ ulcerative tracts vary in diameter, depth, and con­ ter. This result suggests that, compared to their nectivity, and can affect the entire area around the healthy counterparts, CAF dogs have a higher anus. These lesions are usually referred to as fistu­ degree of shedding of epithelial or blood cells asso­ las. The CAF dogs recruited in this study presented ciated with inflammation, similar to what has been with varying numbers of fistulas (average, 3.3 ± 1.4. previously observed in CD humans. Table 1) but no anal sac disease was observed. We In addition to fistulas, GSD with CAF frequently found that CAF dogs in the far cluster exhibited present clinical and histologic evidence of colitis. a trend of higher average number of fistulas than The major microarchitectural changes accompany­ those in the near cluster (4 vs. 2.87, respectively, ing colonic inflammation in dogs include crypt A) B) C) 8 50 3 # Shannon diversity index 40 6 Number of fistulas % Dog DNA 2 30 4 20 1 2 10 0 0 0 Near cluster Far cluster HC dogs CAF dogs HC dogs CAF dogs/fibrosis Figure 3. Clinical manifestation of CAF. A) The number of fistulas presented by CAF dogs according to cluster (orange, near cluster; red, far cluster). B) Percentage of dog DNA reads in metagenomic sequencing samples grouped by disease status. C) Shannon index of alpha diversity in samples from HC dogs (green, n = 8) or CAF dogs exhibiting fibrosis (orange pattern, n = 6). T-test, p values are expressed as: ****<0.0001, ***<0.001, ** <0.005, * <0.05, and #0.1. GUT MICROBES e1785246-7 Table 3. Bacterial taxa that correlate with clinical manifestations of CAF. Spearman gneiss, FDR-corrected Disease manifestation/assessment Determinant bacterial taxa correlation p value p value Gross examination Number of fistulas Helicobacter canis −0.4715 < 0.001 0.00000003 Pathology findings CAF dogs with a combination of crypt hyperplasia, Ruminococcus torques −0.355 0.002 < 0.05 crypt dilation/distortion, and/or fibrosis/atrophy (n = 7) CAF dogs with a combination of microarchitectural Subdoligranulum unclassified −0.1639 0.1751 0.02305158 changes and presence of inflammatory markers Paraprevotella unclassified −0.3336 0.0048 (n = 5) hyperplasia, dilation/distortion, and mucosal atro­ infiltration (Table 3). Recently, depletion of phy or fibrosis.51 Thus, we also assessed morpho­ Subdoligranulum species in dysbiotic CD patients logical features and other immune-associated was associated with bile acid dysregulation.50 inflammatory markers on the colonic mucosa of Based on the presence of major microarchitec­ the CAF dogs (Table 1). tural and immune cells in the lamina propria we First, we observed that there was a positive correla­ calculated pathology scores as described before.51 tion between histopathology scores and number of We observed that the pathology scores from CAF fistulas (Spearman correlation, p value <.05). Further, dogs belonging to the near cluster (median = 2, we found that 75% of the CAF dogs included in this minimum = 0, maximum = 5) were comparable to study exhibited morphological evidence of either those of the far cluster (median = 3, minimum = 0, cryptal hyperplasia, dilation/distortion, or fibrosis. maximum = 5, Mann–Whitney test, p value >.05). CAF dogs with those major microarchitectural To investigate whether CAF dogs exhibited periph­ changes showed differential dominance of the same eral signs of inflammation, complete blood counts bacterial taxa that were associated with general CAF (CBCs) were performed. CBC indicators were within disease status, namely, Megamonas species the normal range for all tested CAF dogs; thus, no sign (M. hypermegale, M. unclassified, and M. funiformis), of serious inflammation or infection was detectable by B. plebeius, and B. coprocola (data not shown). In CBC analysis (data not shown). Although hemoglobin addition to these taxa, Ruminococcus torques, known levels were lower in CAF dogs belonging to the far to be decreased in CD,44,48,52,53 was also identified as cluster than in those in the near cluster, we did not a determinant taxon between CAF dogs with major identify any bacterial taxa associated with hemoglobin microarchitectural changes and HC dogs (gneiss, FDR levels, nor with any of the analyzed blood inflamma­ corrected p value <.05). Moreover, we found that tory markers. R. torques relative abundance is negatively correlated In sum, there is an inverse correlation between with microarchitectural changes (Table 3). H. canis abundance and the number of fistulas. In Interestingly, CAF dogs presenting fibrosis (50%) addition to the bacterial taxa associated with disease exhibited significantly decreased alpha diversity com­ status, CAF dogs exhibiting microarchitectural pared to HC dogs (Figure 3C, Shannon diversity, changes related to inflammation also exhibit Kruskal–Wallis test, p value <.05). decreased R. torques abundance. Furthermore, CAF Furthermore, CAF dogs exhibiting major micro­ dogs presenting microarchitectural defects combined architectural combined with infiltration of immune with immune cell infiltration in the lamina propria cells (i.e., eosinophils, lymphocytes, or plasma cells) have decreased abundance of Subdoligranulum in the lamina propria (41.6%) showed differential unclassified and Paraprevotella unclassified. abundance of Subdoligranulum unclassified and Paraprevotella unclassified when compared to CAF Microbial gene pathways change in accordance with dogs without those clinical manifestations (gneiss, disease status FDR corrected p value = .02). Particularly, those bacterial taxa were negatively correlated with both The genomic content of the microbiota was deter­ major microarchitectural changes and immune cell mined using HUMAnN2.54 After filtering for gene e1785246-8 A. MALDONADO-CONTRERAS ET AL. Figure 4. Microbial gene pathway analysis. A) Bray Curtis dissimilarity index beta diversity plot of dogs samples colored by clustering assignation: HC dogs (green) and CAF dogs belonging to the near (orange) or far (red) cluster. Bacterial pathways were quantified by HUMAnN2 and samples were plotted based on principal component analysis. B) The top 20 gene pathways that most strongly distinguish between CAF-dog samples and HC dog samples, identified using Random Forest. pathways that were not present in at least 5% of the from CAF in order to explore the usefulness of samples, we performed clustering analysis on 1,532 dogs as models for pCD, a serious complication of filtered pathways. Clustering of the gene pathway CD that represents an unmet clinical need. abundances separated CAF dogs belonging to Other studies have aimed to characterize the either the near cluster or the far cluster from HC microbiome composition of dogs with IBD with dogs (Figure 4A, PERMANOVA, R2 = 0.19, varying symptoms and from different breeds55-58 p value = .001). and their similarities with those from humans Random Forest was used to identify microbial with IBD.46 However, this is the first study that gene pathways that best distinguished HC dogs aims to describe the similarities of dogs suffering from CAF dogs, predicting differences with an from a specific clinical manifestation: canine fur­ OOB estimate of error rate of 6.48% (Figure 4B). unculosis (CAF), to compare it to those from As expected, most of the identified gene pathways humans with CD. Moreover, we only included in originated from bacterial genera that were found to the study German Shepherd Dogs, a breed suscep­ be predictive of disease status, specifically, tible to IBD with genetic polymorphisms compar­ Megamonas, Bacteroides, and Helicobacter. able to those found in their IBD human Altogether, the metabolic capacity of the micro­ counterpart.59,60 biome varied along with disease status or cluster Interestingly, we found that similar to the assignation. Hence, most of the discriminant gene microbiome of pediatric44 and adult50 patients pathways belong exclusively to bacterial taxa with IBD, the microbiomes of CAF dogs can be depleted in CAF dogs. Although it is uncertain either healthy-like or dysbiotic. The most recent that this phenomenon causes or is a consequence Human Microbiome Project (HMP) study includ­ of disease, characterization of these changes will ing 132 IBD patients and 2,965 stool samples col­ enlighten our understanding of the microbial lected over a year showed the same bi-modal dynamics in disease. microbiome pattern among patients with IBD.50 Specifically, the authors found that differences between dysbiotic and healthy-like microbiomes Discussion were more pronounced in patients with CD. Our understanding of the pathophysiology of pCD Moreover, dysbiotic microbiomes in patients is not complete albeit their high prevalence and with CD were characterized by depletion of obli­ burden for patients with CD. The major challenge gate anaerobes and enrichment of facultative- to study the pathogenesis of this manifestation is anaerobes.50 Similarly, we found that decreased the absence of relevant animal models. Here, we abundance of obligate anaerobes such as characterized the microbiome of dogs suffering Megamonas species and P. copri and increased GUT MICROBES e1785246-9 abundance of facultative-anaerobes such as E. coli of inflammation and aberrant immune response in distinguish between samples from CAF dogs with the gut.75–79 Therefore, these two strains seemed to a dysbiotic microbiome and those from HC dogs. be favored in the context of CAF and CD. Megamonas is a common gut commensal Concomitantly with what has been observed in microbe of carnivore animals. Increased abundance patients with CD, we found that dysbiotic micro­ of Megamonas has been found in healthy cats61 and biomes are associated with host independent mar­ in dogs consuming inulin-rich diets.62 Few studies kers of disease.50 Here, CAF dogs with dysbiotic have shown the relevance of Megamonas in human microbiomes tend to have a higher number of health.63,64 Megamonas species are important pro­ fistulas. We also found that markers of inflamma­ pionate and acetate producers.65,66 Thus, based on tion such as microarchitectural changes and these results, a greater emphasis on Megamonas, immune cell infiltration in the lamina propria and its potential impact on gastrointestinal health, were negatively correlated with the abundance of may be justified in the future. Ruminococcus torques, Subdoligranulum unclassi­ Moreover, as seen in CAF dogs with a dysbiotic fied, and Paraprevotella unclassified. Of note, microbiome, pediatric patients with CD exhibiting Ruminococcus torques is a butyrate-producing a dysbiotic microbiome harbored significantly less Clostridia shown to induce anti-inflammatory abundance of P. copri compared to healthy control activity and is known to be decreased in individuals.44 In the HMP study, however, abun­ CD.44,48,52,53 Similarly, the HMP study50 estab­ dance of P. copri was not significantly different lished that Subdoligranulum species – also with between CD patients and healthy controls but the butyrate-producing capacities – were both reduced latter showed an expansion and relaxation of in dysbiotic CD patients and associated dysregula­ P. copri abundance over time. This organism is of tion of IBD-linked metabolites (i.e., bile acids, car­ particular interest given that it has been long asso­ nitines, and polyunsaturated fatty acids). ciated with health but also more recently with new- In sum, we found that dogs suffering from CAF onset of rheumatoid arthritis.67 Thus, further exhibit gut bacterial community structures that research on the role of P. copri in IBD is needed. resemble those observed in CD patients. In the HMP study, the depleted anaerobe species Moreover, we found overlapping bacteria species in the far dysbiotic cluster were Faecalibacterium and their metabolic capacity associated with dys­ praustnizii and Ruminococcus hominis.50 Thus, biosis in both CAF and human CD. Therefore, we although depleted bacterial species might not over­ propose to use dogs suffering from CAF as lap between CAF and CD, we do observe a surrogate model to: (1) study pCD to further a microbiome composition trend, namely elucidate the role of the microbiome in this clinical a dysbiotic microbiome characterized by depletion complication, and (2) explore the pre-clinical effi­ of obligate anaerobes, that is similar between the cacy of microbiome-centered therapeutics to two diseases. treat pCD. On the other hand, we did find that increased abundance of E.coli overlaps between CAF and CD. Increased abundance of Proteobacteria, specifically Methods and materials E. coli, has been definitively demonstrated in Subjects and sample collection patients with CD.68–72 In addition, the HMP study reported that CD patients exhibiting a dysbiotic All dogs in this study were GSD that had not microbiome harbor increased abundance of received antibiotic treatment within the preceding E. coli.50 3 months and were treated with a special diet (novel We also found an enrichment of Bacteroides protein or hypoallergenic) for at least 8 weeks with­ vulgatus in the far dysbiotic cluster. Increased out clinical response. CAF dogs were recruited abundance of B. vulgatus has previously been from Cummings School of Veterinary Medicine at observed in dogs46,56-58 and in patients with Tufts University. Inclusion criteria included IBD.73,74 Bacteroides species, namely B. vulgatus, German Shepard dogs with a clinical diagnosis of are thought to be responsible for the development perianal fistulas with clinical signs of tenesmus, e1785246-10 A. MALDONADO-CONTRERAS ET AL. dyschezia, and partial or complete relapse from Metagenomic profiling cyclosporine A therapy. HC dogs were selected Community composition was profiled using based on the absence of perianal fistulas, nonexis­ MetaPhlan281 to determine the per sample relative tence of diarrhea or adverse gastrointestinal symp­ abundance of species composition of bacterial, fun­ toms, and lack of immunosuppressant treatment. gal, parasitic, and viral genomes per sample. Then, Dog information including age, weight, gender, to assess the content of metabolic and functional number of fistulas, treatment, and antibiotic usage genes and metabolic pathways we used was obtained from clinical records. HUMAnN2.54 Naturally passed feces were collected daily for a period of 1 week from eight HC dogs and from 12 CAF dogs. Dog owners were provided with Statistical analysis tubes (OMNI•gene-GUT| OMR-200, DNAgenotek, Canada) for sample collection and were instructed to For analysis, we considered only taxa that were place a clean pad on the surface before dogs’ bowel detected in at least 10% of samples and had movement to avoid sample contamination from soil, a relative abundance of at 0.2% or greater in at grass, or other surfaces. Samples were freshly col­ least one sample. Community patterns were ana­ lected into the tubes provided. Samples were stored lyzed using partitioning around medoids with the at room temperature until they were delivered at the estimation of the number of clusters (PAMK) to next clinical appointment, which occurred within find the optimal number of clusters as performed 2 weeks of sampling. During that next appointment, previously44 and visualized after multidimensional biopsies from the distal colon and blood samples scaling (MDS). Clustering analyses and visualiza­ were obtained from all CAF dogs for histopathologi­ tions were performed in Phyloseq 1.26.1 and the cal analysis and CBC, respectively. To determine R package cluster v1.4–1 to estimate microbiome colonic inflammation, an experienced pathologist patterns using PAMK with optimum average sil­ blind-scored the biopsies based on the infiltration houette width.82,83 Briefly, the PAM algorithm is of eosinophils, lymphocytes, or plasma cells in the based on the search for “k” representative objects or lamina propria and morphological changes, includ­ k-medoids among the observations of the data set. ing crypt dilation/distortion, fibrosis/atrophy, or In k-medoids clustering, each cluster is represented crypt hyperplasia, as previously described.80 by one of the data point in the cluster. These points are named cluster medoids. After finding a set of k-medoids, clusters are constructed by assigning each observation to the nearest medoid. To esti­ DNA isolation and sequencing mate the optimal number of clusters, we estimated DNA isolation was performed using the the average silhouette method. Here, we compute MagAttract PowerSoil DNA Kit (Qiagen, # 27100- the PAM algorithm using different values of clus­ 4-EP) on Eppendorf epMotion 5075 liquid handlers ters k. Next, the average cluster silhouette is calcu­ following the manufacturer’s instructions. DNA lated according to the number of clusters. A high sequencing libraries were prepared using the average silhouette width indicates a good cluster­ Nextera XT DNA Library Preparation Kit ing. The optimal number of clusters k is the one (Illumina, #FC-131-1096) and were sequenced on that maximizes the average silhouette over a range the Illumina NextSeq 500 platform as 150-nt of possible values for k. paired-end reads. We obtained an average of Shannon diversity index (alpha diversity) and 3.8 M reads per sample (minimum: 1 M reads and Bray Curtis dissimilarity (beta diversity) were calcu­ maximum: 7.7 M reads). Read data were quality lated using QIIME2 1.8.9. To assess the significance trimmed and filtered of host DNA using of differences in diversity metrics between samples KneadData (https://bitbucket.org/biobakery/knead according to either disease status or cluster assigna­ data/wiki/Home) against a prebuilt bowtie2 index tion, we used Kruskal–Wallis, PERMANOVA, for Canis familiaris reference genome, build 3.1. ADONIS, and Mantel tests. Determinant bacterial GUT MICROBES e1785246-11 species were identified by the robust algorithms Author contributions gneiss47 and ANCOM84 implemented in QIIME2. A.M-C, A.H., and L.F. collectively conceptualized the manu­ In gneiss, we used linear-mixed models (LME) to script. A.M-C, D.V.W, J.T, and S.C contributed to dog recruit­ account for the random effect of patients with mul­ ment, sample management, sequencing, and the data analysis. S. tiple samples. LME are used to test the relationship B helped with biostatistics analyses. S.C performed colonosco­ between a single response variable and one or more pies and histopathology. A.M-C wrote the manuscript. All independent variables, where observations are authors reviewed and edited the final version of the text. made across dependent samples. We then evaluated the correlations (with FRD correction) of 61 bac­ Availability of supporting data terial taxa with each variable separately, specifically disease status (CAF vs HC), cluster assignation Sequence files for all samples used in this study have been deposited in the NCBI SRA (SRA: SRP191145 and Bioproject (CAF near cluster, CAF far cluster, and HC), or PRJNA53120). Metadata and MetaPhlan tables – with corre­ clinical manifestations (number of fistulas, crypt sponding taxonomic classifications – have been included as hyperplasia, crypt dilation/distortion, fibrosis/atro­ Additional files 1 and 2. phy). To understand the direction of dominance found in gneiss we performed Spearman correla­ tions of the determinant bacterial taxa with the Competing interest statements clinical manifestation of interest. The authors declare no competing interest. Additionally, we use ANCOM adjusting for dependent repeated measurements.84 For ANCOM, we analyzed 61 bacterial taxa, and only bacterial taxa Ethics approval and consent to participate reporting W > 45 were considered significant. The protocol for sample collection was approved by the Finally, we used Random Forest (R package, Clinical Research Review Committee of the Cummings randomForest 4.6–14),6,8a supervised learning algo­ School of Veterinary Medicine at Tufts University rithm, to identify gene pathways that discriminated (CRRC#007-015). Each pet owner consented to participate in the study. samples by disease status as previously described.44 Abbreviations Acknowledgments ANCOM Analysis of Composition of Microbiomes We thank members of the clinical staff, namely Diane Welsh CD Crohn’s disease for her help with dog identification and recruitment for the CAF canine anal furunculosis study. HC Healthy controls IBD Inflammatory Bowel Diseases Funding References Financial support for this research was provided by the Office of Faculty Affairs at the University of Massachusetts Medical 1. Panes J, Reinisch W, Rupniewska E, Khan S, Forns J, School and the American Gastroenterological Association to Khalid JM, Bojic D, Patel H. Burden and outcomes for A.M-C.The Shipley Foundation provided support to A.H; The complex perianal fistulas in Crohn’s disease: systematic Shipley Foundation [The Shipley Foundation]; Office of review. World Journal of Gastroenterology: WJG. Faculty Affairs, University of Massachusetts Medical School 2018;24:4821–4834. doi:10.3748/wjg.v24.i42.4821. [Faculty Diversity Program] provided support to AM-C; 2. Kelley KA, Kaur T, Tsikitis VL. Perianal Crohn’s dis­ American Gastroenterological Association (US) [Research ease: challenges and solutions. Clin Exp Gastroenterol. Scholar Award]; provided support to AM-C 2017;10:39–46. doi:10.2147/CEG.S108513. 3. Beaugerie L, Seksik P, Nion-Larmurier I, Gendre JP, Cosnes J. Predictors of Crohn’s disease. ORCID Gastroenterology. 2006;130:650–656. doi:10.1053/j. gastro.2005.12.019. Ana Maldonado-Contreras http://orcid.org/0000-0002- 4. Cleynen I, González JR, Figueroa C, Franke A, 5967-9623 McGovern D, Bortlík M, Crusius BJA, Vecchi M, Doyle V. Ward http://orcid.org/0000-0002-4952-824X Artieda M, Szczypiorska M, et al. Genetic factors e1785246-12 A. MALDONADO-CONTRERAS ET AL. conferring an increased susceptibility to develop Small Anim Pract. 2019;49:53–65. doi:10.1016/j. Crohn’s disease also influence disease phenotype: cvsm.2018.08.006. results from the IBDchip European Project. Gut. 17. Hardie RJ, Gregory SP, Tomlin J, Sturgeon C, 2013;62:1556–1565. doi:10.1136/gutjnl-2011-300777. Lipscomb V, Ladlow J. Cyclosporine treatment of anal 5. Schnitzler F, Friedrich M, Wolf C, Stallhofer J, furunculosis in 26 dogs. J Small Anim Pract. Angelberger M, Diegelmann J, Olszak T, Tillack C, 2005;46:3–9. 10.1111/j.1748-5827.2005.tb00267.x. Beigel F, Göke B, et al. The NOD2 Single nucleotide poly­ 18. Gold SL, Cohen-Mekelburg S, Schneider Y, Steinlauf A. morphism rs72796353 (IVS4+10 A>C) is a predictor for Perianal fistulas in patients with Crohn’s disease, part 1: perianal fistulas in patients with Crohn’s disease in the current medical management. Gastroenterol Hepatol absence of other NOD2 mutations. PLoS One. 2015;10: (N Y). 2018;14:470–481. e0116044. doi:10.1371/journal.pone.0116044. 19. Adam P, Patterson KLC. Managing anal furunculosis in 6. Ruffolo C, Scarpa M, Faggian D, Romanato G, De dogs. Internal Medicine Compendium. 2005;27. https:// Pellegrin A, Filosa T, Prando D, Polese L, Scopelliti M, pdfs.semanticscholar.org/1fe4/ Pilon F, et al. Cytokine network in chronic perianal 5115545e42e8133503e87a1a4343c46ce4ff.pdf. Crohn’s disease and indeterminate colitis after colectomy. 20. Tisdall PL, Hunt GB, Beck JA, Malik R. Management of J Gastrointest Surg. 2007;11:16–21. doi:10.1007/s11605-006- perianal fistulae in five dogs using azathioprine and metro­ 0021-y. nidazole prior to surgery. Aust Vet J. 1999;77:374–378. 7. Mennigen R, Heptner B, Senninger N, Rijcken E. doi:10.1111/j.1751-0813.1999.tb10307.x. Temporary fecal diversion in the management of color­ 21. Stanley BJ, Hauptman JG. Long-term prospective eva­ ectal and perianal Crohn’s disease. Gastroenterol Res luation of topically applied 0.1% tacrolimus ointment Pract. 2015;2015(286315). doi:10.1155/2015/286315. for treatment of perianal sinuses in dogs. J Am Vet Med 8. Rehg KL, Sanchez JE, Krieger BR, Marcet JE. Fecal Assoc. 2009;235:397–404. doi:10.2460/javma.235.4.397. diversion in perirectal fistulizing Crohn’s disease is an 22. Levine A, Rhodes JM, Lindsay JO, Abreu MT, underutilized and potentially temporary means of suc­ Kamm MA, Gibson PR, Gasche C, Silverberg MS, cessful treatment. Am Surg. 2009;75:715–718. Mahadevan U, Boneh RS, et al. Dietary Guidance for 9. Zhang FM, Wang HG, Wang M, et al. Fecal microbiota Patients with Inflammatory Bowel Disease from the transplantation for severe enterocolonic fistulizing International Organization for the study of inflamma­ Crohn’s disease. World J Gastroenterol. tory bowel disease. Clin Gastroenterol Hepatol. 2020. 2013;19:7213–7216. doi:10.3748/wjg.v19.i41.7213. doi:10.1016/j.cgh.2020.01.046. 10. Panes J, Rimola J. Perianal fistulizing Crohn’s disease: 23. Marlow G, Ellett S, Ferguson IR, Zhu S, Karunasinghe N, pathogenesis, diagnosis and therapy. Nat Rev Jesuthasan AC, Han D, Fraser AG, Ferguson LR. Gastroenterol Hepatol. 2017. doi:10.1038/ Transcriptomics to study the effect of a Mediterranean- nrgastro.2017.104. inspired diet on inflammation in Crohn’s disease patients. 11. Ferrer L, Kimbrel EA, Lam A, Falk EB, Zewe C, Hum Genomics. 2013;7(24). doi:10.1186/1479-7364-7-24. Juopperi T, Lanza R, Hoffman A. Treatment of perianal 24. Suskind DL, Wahbeh G, Gregory N, Vendettuoli H, fistulas with human embryonic stem cell-derived Christie D. Nutritional therapy in pediatric Crohn dis­ mesenchymal stem cells: a canine model of human ease: the specific carbohydrate diet. J Pediatr fistulizing Crohn’s disease. Regen Med. 2016;11:33–43. Gastroenterol Nutr. 2014;58:87–91. doi:10.1097/ doi:10.2217/rme.15.69. MPG.0000000000000103. 12. Hoffman AM, Dow SW. Concise review: stem cell trials 25. Olendzki BC, Silverstein TD, Persuitte GM, Ma Y, using companion animal disease models. Stem Cells. Baldwin KR, Cave D. An anti-inflammatory diet as 2016;34:1709–1729. doi:10.1002/stem.2377. treatment for inflammatory bowel disease: a case series 13. Sandborn WJ, Fazio VW, Feagan BG, Hanauer SB. report. Nutr J. 2014;13(5). doi:10.1186/1475-2891-13-5. American gastroenterological association clinical practice, 26. Proverbio D, Perego R, Spada E, Ferro E. Prevalence of C. AGA technical review on perianal Crohn’s disease. adverse food reactions in 130 dogs in Italy with derma­ Gastroenterology. 2003;125:1508–1530. doi:10.1016/j. tological signs: a retrospective study. J Small Anim Pract. gastro.2003.08.025. 2010;51:370–374. doi:10.1111/j.1748-5827.2010.00951.x. 14. Taxonera C, Schwartz DA, Garcia-Olmo D. Emerging 27. Lombardi RL, Marino DJ. Long-term evaluation of treatments for complex perianal fistula in Crohn’s dis­ canine perianal fistula disease treated with exclusive ease. World Journal of Gastroenterology: WJG. fish and potato diet and surgical excision. J Am 2009;15:4263–4272. doi:10.3748/wjg.15.4263. Anim Hosp Assoc. 2008;44:302–307. doi:10.5326/ 15. Jamieson PM, Simpson JW, Kirby BM, Else RW. 0440302. Association between anal furunculosis and colitis in the 28. Harkin KR, Walshaw R, Mullaney TP. Association of dog: preliminary observations. J Small Anim Pract. perianal fistula and colitis in the German shepherd dog: 2002;43:109–114. doi:10.1111/j.1748-5827.2002.tb00039.x. response to high-dose prednisone and dietary therapy. 16. Cain CL. Canine perianal fistulas: clinical presentation, J Am Anim Hosp Assoc. 1996;32:515–520. doi:10.5326/ pathogenesis, and management. Vet Clin North Am 15473317-32-6-515. GUT MICROBES e1785246-13 29. Day MJ, W. BMQ. Pathology of surgically resected Prog Neuropsychopharmacol Biol Psychiatry. tissue from 305 cases of anal furunculosis in the dog. 2000;24:727–776. 10.1016/S0278-5846(00)00104-4. Journal of Small Animal Practice. 1992;33:583–589. 10. 43. Coelho LP, Kultima JR, Costea PI, Fournier C, Pan Y, 1111/j.1748-5827.1992.tb01062.x. Czarnecki-Maulden G, Hayward MR, Forslund SK, 30. Killingsworth CR, Walshaw R, Dunstan RW, Rosser EJ Schmidt TSB, Descombes P, et al. Similarity of the dog Jr. Bacterial population and histologic changes in dogs and human gut microbiomes in gene content and with perianal fistula. Am J Vet Res. 1988;49:1736–1741. response to diet. Microbiome. 2018;6(72). doi:10.1186/ 31. Kathrani A, House A, Catchpole B, Murphy A, s40168-018-0450-3 German A, Werling D, Allenspach K. Polymorphisms 44. Lewis JD, Chen E, Baldassano R, Otley A, Griffiths A, in the TLR4 and TLR5 gene are significantly associated Lee D, Bittinger K, Bailey A, Friedman E, Hoffmann C, with inflammatory bowel disease in German shepherd et al. Inflammation, antibiotics, and diet as environ­ dogs. PLoS One. 2010;5(e15740). doi:10.1371/journal. mental stressors of the gut microbiome in pediatric pone.0015740. Crohn’s disease. Cell Host Microbe. 2015;18:489–500. 32. Kathrani A, Lee H, White C, Catchpole B, Murphy A, doi:10.1016/j.chom.2015.09.008. German A, Werling D, Allenspach K. Association 45. Halfvarson J, Brislawn CJ, Lamendella R, Vázquez- between nucleotide oligomerisation domain two Baeza Y, Walters WA, Bramer LM, D’Amato M, (Nod2) gene polymorphisms and canine inflammatory Bonfiglio F, McDonald D, Gonzalez A, et al. Dynamics bowel disease. Vet Immunol Immunopathol. of the human gut microbiome in inflammatory bowel 2014;161:32–41. doi:10.1016/j.vetimm.2014.06.003. disease. Nat Microbiol. 2017;2(17004). doi:10.1038/ 33. Bruchim Y, Elad D, Klainbart S. Disseminated aspergil­ nmicrobiol.2017.4 losis in two dogs in Israel. Mycoses. 2006;49:130–133. 46. Vazquez-Baeza Y, Hyde ER, Suchodolski JS, doi:10.1111/j.1439-0507.2006.01168.x. Knight R. Dog and human inflammatory bowel dis­ 34. Day MJ, Penhale WJ, Kabay MJ, Robinson WF, ease rely on overlapping yet distinct dysbiosis Huxtable CRR, Eger CE, Shaw SE, Mills JN, networks. Nat Microbiol. 2016;1(16177). Wyburn RS. Disseminated aspergillosis in dogs. Aust doi:10.1038/nmicrobiol.2016.177. Vet J. 1986;63:55–59. 10.1111/j.1751-0813.1986. 47. Morton JT, Sanders J, Quinn RA, McDonald D, tb02924.x. Gonzalez A, Vázquez-Baeza Y, Navas-Molina JA, 35. Cook E, Meler E, Garrett K, Long H, Mak K, Song SJ, Metcalf JL, Hyde ER, et al. Balance trees reveal Stephens C, Thompson A. Disseminated microbial niche differentiation. mSystems. 2017;2. Chrysosporium infection in a German shepherd dog. doi:10.1128/mSystems.00162-16. Med Mycol Case Rep. 2015;10:29–33. doi:10.1016/j. 48. Gevers D, Kugathasan S, Denson L, Vázquez-Baeza Y, mmcr.2016.01.002. Van Treuren W, Ren B, Schwager E, Knights D, Song S, 36. Pastor J, Pumarola M, Cuenca R, Lavin S. Systemic Yassour M, et al. The treatment-naive microbiome in aspergillosis in a dog. Vet Rec. 1993;132:412–413. 10. new-onset Crohn’s disease. Cell Host Microbe. 1136/vr.132.16.412. 2014;15:382–392. doi:10.1016/j.chom.2014.02.005. 37. Rosser EJ Jr. German shepherd dog pyoderma: 49. Frank DN, St. Amand AL, Feldman RA, Boedeker EC, a prospective study of 12 dogs. J Am Anim Hosp Harpaz N, Pace NR. Molecular-phylogenetic character­ Assoc. 1997;33:355–363. doi:10.5326/15473317-33- ization of microbial community imbalances in human 4-355. inflammatory bowel diseases. Proc Natl Acad Sci U S A. 38. Miller WH. Deep pyoderma in two German shepherd 2007;104:13780–13785. doi:10.1073/pnas.0706625104. dogs associated with a cell mediated immunodeficiency. 50. Lloyd-Price J, Arze C, Ananthakrishnan AN, J Am Anim Hosp Assoc. 1991;27. Schirmer M, Avila-Pacheco J, Poon TW, Andrews E, 39. Vail DM, MacEwen EG. Spontaneously occurring Ajami NJ, Bonham KS, Brislawn CJ, et al. Multi-omics tumors of companion animals as models for human of the gut microbial ecosystem in inflammatory bowel cancer. Cancer Invest. 2000;18:781–792. 10.3109/ diseases. Nature. 2019;569:655–662. doi:10.1038/ 07357900009012210. s41586-019-1237-9. 40. Cadieu E, Ostrander EA. Canine genetics offers new 51. Robert J, Washabau MJD. Canine and feline gastroen­ mechanisms for the study of human cancer. Cancer terology. USA: Elsevier Saunders; 2013. Epidemiol Biomarkers Prev. 2007;16:2181–2183. 52. Takahashi K, Nishida A, Fujimoto T, Fujii M, Shioya M, doi:10.1158/1055-9965.EPI-07-2667. Imaeda H, Inatomi O, Bamba S, Andoh A, Sugimoto M, 41. Jonasdottir TJ, Mellersh CS, Moe L, Heggebo R, et al. Reduced abundance of butyrate-producing bac­ Gamlem H, Ostrander EA, Lingaas F. Genetic mapping teria species in the fecal microbial community in of a naturally occurring hereditary renal cancer syn­ Crohn’s disease. Digestion. 2016;93:59–65. drome in dogs. Proc Natl Acad Sci U S A. doi:10.1159/000441768. 2000;97:4132–4137. doi:10.1073/pnas.070053397. 53. Joossens M, Huys G, Cnockaert M, De Preter V, 42. Overall KL. Natural animal models of human psychia­ Verbeke K, Rutgeerts P, Vandamme P, Vermeire S. tric conditions: assessment of mechanism and validity. Dysbiosis of the faecal microbiota in patients with e1785246-14 A. MALDONADO-CONTRERAS ET AL. Crohn’s disease and their unaffected relatives. Gut. disease. Medicine (Baltimore). 2014;93(e51). 2011;60:631–637. doi:10.1136/gut.2010.223263. doi:10.1097/MD.0000000000000051 54. Abubucker S, Segata N, Goll J, Schubert AM, Izard J, 65. Zhao C, Dong H, Zhang Y, Li Y. Discovery of potential Cantarel BL, Rodriguez-Mueller B, Zucker J, genes contributing to the biosynthesis of short-chain Thiagarajan M, Henrissat B, et al. Metabolic reconstruc­ fatty acids and lactate in gut microbiota from systematic tion for metagenomic data and its application to the investigation in E. coli. NPJ Biofilms Microbiomes. human microbiome. PLoS Comput Biol. 2012;8 2019;5(19). doi:10.1038/s41522-019-0092-7. (e1002358). doi:10.1371/journal.pcbi.1002358 66. Polansky O, Sekelova Z, Faldynova M, Sebkova A, Sisak F, 55. Guard BC, Barr JW, Reddivari L, Klemashevich C, Rychlik I. Important metabolic pathways and biological Jayaraman A, Steiner JM, Vanamala J, Suchodolski JS. processes expressed by chicken cecal microbiota. Appl Characterization of microbial dysbiosis and metabolo­ Environ Microbiol. 2015;82:1569–1576. doi:10.1128/ mic changes in dogs with acute diarrhea. PLoS One. AEM.03473-15. 2015;10(e0127259). doi:10.1371/journal.pone.0127259. 67. Scher JU, Sczesnak A, Longman RS, Segata N, Ubeda C, 56. Suchodolski JS, Xenoulis PG, Paddock CG, Steiner JM, Bielski C, Rostron T, Cerundolo V, Pamer EG, Jergens AE. Molecular analysis of the bacterial micro­ Abramson SB, et al. Expansion of intestinal Prevotella biota in duodenal biopsies from dogs with idiopathic copri correlates with enhanced susceptibility to inflammatory bowel disease. Vet Microbiol. arthritis. Elife. 2013;2(e01202). doi:10.7554/eLife.01202 2010;142:394–400. doi:10.1016/j.vetmic.2009.11.002. 68. Martinez C, Antolin M, Santos J, Torrejon A, Casellas F, 57. Suchodolski JS, Dowd SE, Wilke V, Steiner JM, Borruel N, Guarner F, Malagelada J-R. Unstable composi­ Jergens AE. 16S rRNA gene pyrosequencing reveals tion of the fecal microbiota in ulcerative colitis during clin­ bacterial dysbiosis in the duodenum of dogs with idio­ ical remission. Am J Gastroenterol. 2008;103:643–648. pathic inflammatory bowel disease. PLoS One. 2012;7 doi:10.1111/j.1572-0241.2007. (e39333). doi:10.1371/journal.pone.0039333. 01592.x. 58. Minamoto Y, Otoni CC, Steelman SM, 69. Baumgart M, Dogan B, Rishniw M, Weitzman G, Büyükleblebici O, Steiner JM, Jergens AE, Bosworth B, Yantiss R, Orsi RH, Wiedmann M, Suchodolski JS. Alteration of the fecal microbiota and McDonough P, Kim SG, et al. Culture independent serum metabolite profiles in dogs with idiopathic analysis of ileal mucosa reveals a selective increase in inflammatory bowel disease. Gut Microbes. 2015;6(33–­ invasive Escherichia coli of novel phylogeny relative to 47). doi:10.1080/19490976.2014.997612. depletion of Clostridiales in Crohn’s disease involving 59. Massey J, Short AD, Catchpole B, House A, Day MJ, Lohi H, the ileum. Isme J. 2007;1:403–418. doi:10.1038/ Ollier WER, Kennedy LJ. Genetics of canine anal furuncu­ ismej.2007.52. losis in the German shepherd dog. Immunogenetics. 70. Kotlowski R, Bernstein CN, Sepehri S, Krause DO. High 2014;66:311–324. doi:10.1007/s00251-014-0766-5. prevalence of Escherichia coli belonging to the B2+D 60. Peiravan A, Bertolini F, Rothschild MF, Simpson KW, phylogenetic group in inflammatory bowel disease. Gut. Jergens AE, Allenspach K, Werling D. Genome-wide 2007;56:669–675. doi:10.1136/gut.2006.099796. association studies of inflammatory bowel disease in 71. Martin HM, Campbell BJ, Hart CA, Mpofu C, Nayar M, German shepherd dogs. PLoS One. 2018;13(e0200685). Singh R, Englyst H, Williams HF, Rhodes JM. Enhanced doi:10.1371/journal.pone.0200685. Escherichia coli adherence and invasion in Crohn’s 61. Suchodolski JS, Foster ML, Sohail MU, Leutenegger C, disease and colon cancer. Gastroenterology. Queen EV, Steiner JM, Marks SL. The fecal microbiome 2004;127:80–93. 10.1053/j.gastro.2004.03.054. in cats with diarrhea. PLoS One. 2015;10(e0127378). 72. Nishino K, Nishida A, Inoue R, Kawada Y, Ohno M, doi:10.1371/journal.pone.0127378. Sakai S, Inatomi O, Bamba S, Sugimoto M, 62. Beloshapka AN, Dowd SE, Suchodolski JS, Steiner JM, Kawahara M, et al. Analysis of endoscopic brush sam­ Duclos L, Swanson KS. Fecal microbial communities of ples identified mucosa-associated dysbiosis in inflam­ healthy adult dogs fed raw meat-based diets with or matory bowel disease. J Gastroenterol. 2018;53:95–106. without inulin or yeast cell wall extracts as assessed by doi:10.1007/s00535-017-1384-4. 454 pyrosequencing. FEMS Microbiol Ecol. 73. Fujita H, Eishi Y, Ishige I, Saitoh K, Takizawa T, 2013;84:532–541. doi:10.1111/1574-6941.12081. Arima T, Koike M. Quantitative analysis of bacterial 63. Zhang X, Shen D, Fang Z, Jie Z, Qiu X, Zhang C, et al. DNA from Mycobacteria spp., Bacteroides vulgatus, Human gut microbiota changes reveal the progression and Escherichia coli in tissue samples from patients of glucose intolerance. PLoS One. 2013;8(e71108). with inflammatory bowel diseases. J Gastroenterol. doi:10.1371/journal.pone.0071108 2002;37:509–516. doi:10.1007/s005350200079. 64. Chen L, Wang W, Zhou R, Ng SC, Li J, Huang M, 74. Lucke K, Miehlke S, Jacobs E, Schuppler M. Prevalence Zhou F, Wang X, Shen BA, Kamm M, et al. of Bacteroides and Prevotella spp. in ulcerative colitis. Characteristics of fecal and mucosa-associated micro­ J Med Microbiol. 2006;55:617–624. doi:10.1099/ biota in Chinese patients with inflammatory bowel jmm.0.46198-0. GUT MICROBES e1785246-15 75. Bloom SM, Bijanki V, Nava G, Sun L, Malvin N, Immunol. 2003;47:371–378. doi:10.1111/j.1348- Donermeyer D, Dunne W, Allen P, Stappenbeck T. 0421.2003.tb03368.x. Commensal Bacteroides species induce colitis in 80. Jergens AE, Evans RB, Ackermann M, Hostetter J, host-genotype-specific fashion in a mouse model of Willard M, Mansell J, Bilzer T, Wilcock B, inflammatory bowel disease. Cell Host Microbe. Washabau R, Hall EJ, et al. Design of a simplified his­ 2011;9:390–403. doi:10.1016/j.chom.2011.04.009. topathologic model for gastrointestinal inflammation in 76. Bamba T, Matsuda H, Endo M, Fujiyama Y. The dogs. Vet Pathol. 2014;51:946–950. doi:10.1177/ pathogenic role of Bacteroides vulgatus in patients 0300985813511123. with ulcerative colitis. J Gastroenterol. 1995;30 81. Truong DT, Franzosa EA, Tickle TL, Scholz M, :45–47. Weingart G, Pasolli E, Tett A, Huttenhower C, 77. Breeling JL, Onderdonk AB, Cisneros RL, Kasper DL. Segata N. MetaPhlAn2 for enhanced metagenomic Bacteroides vulgatus outer membrane antigens asso­ taxonomic profiling. Nat Methods. 2015;12:902–903. ciated with carrageenan-induced colitis in guinea pigs. doi:10.1038/nmeth.3589. Infect Immun. 1988;56:1754–1759. 10.1128/IAI.56.7. 82. Rousseeuw P. Silhouettes: A graphical aid to the inter­ 1754-1759.1988. pretation and validation of cluster analysis. J Comput 78. Matsuda H, Fujiyama Y, Andoh A, Ushijima T, Appl Math. 1987;20:53–65. 10.1016/0377-0427(87) Kajinami T, Bamba T. Characterization of antibody 90125-7. responses against rectal mucosa-associated bacterial 83. Kaufman L, A. R. PJ. Finding groups in data: an intro­ flora in patients with ulcerative colitis. duction to cluster analysis. USA and Canada: John J Gastroenterol Hepatol. 2000;15:61–68. doi:10.1046/ Wiley & Sons, Inc; 1990. j.1440-1746.2000.02045.x. 84. Mandal S, Van Treuren W, White RA, Eggesbo M, 79. Shiba T, Aiba Y, Ishikawa H, Ushiyama A, Takagi A, Knight R, Peddada SD. Analysis of composition of Mine T, Koga Y. The suppressive effect of bifidobac­ microbiomes: a novel method for studying microbial teria on bacteroides vulgatus, a putative pathogenic composition. Microb Ecol Health Dis. 2015;26(27663). microbe in inflammatory bowel disease. Microbiol doi:10.3402/mehd.v26.27663