APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Aug. 1977, P. 194-206 Copyright C 1977 American Society for Microbiology

Vol. 34, No. 2 Printed in U.S.A.

Bacterial Association in the Gastrointestinal Tract of Beagle Dogs C. P. DAVIS,'* D. CLEVEN, E. BALISH, AND C. E. YALE Departments of Surgery and Medical Microbiology, University of Wisconsin Center for Health Sciences, Madison, Wisconsin 53706

Received for publication 3 March 1977

Nine male beagle dogs, housed in either a conventional or locked environment for 2.5 years, were killed, and the bacterial flora present in various regions of each gastrointestinal tract was assessed by culture techniques, light microscopy, and scanning electron microscopy. All dogs possessed a complex microflora in their colons; in almost every dog anaerobes predominated. The highest number of bacteria cultured was 1010/g (dry weight) oftissue and contents; highest counts obtained with a Petroff-Hauser counting chamber were 1010/ml (wet weight). Although there was a consistency in the detectable genera, there were also noticeable differences in the flora of dogs housed under different environmental conditions. These differences included qualitative and quantitative changes in the flora as well as alterations in the distribution and localization of microorganisms along the gastrointestinal tract and in the crypts of Lieberkuhn. No bacterial layers were detected on the surfaces of stomach or proximal bowel in any of the dogs. Dogs housed in a conventional, open, environment had bacteria that occurred in layers on their ceca and colons and in their crypts of Lieberkuhn; however, dogs housed under "locked" environmental conditions did not possess them or had them less frequently. Dogs removed from the locked environment and kept (30 days) in conventional housing conditions were the only ones with detectable segmented filamentous microbes in their ilea. This study shows that the microbial flora does not simplify when dogs are housed in a locked environment. Indeed, it may increase in complexity and cause alterations in the bacterial flora that is associated closely with gastrointestinal epithelial cells and crypts of Lieberkuhn.

Studies of the microbial flora of murine species have shown that the adult gastrointestinal (GI) tract contains many microbial species that live in relatively stable populations that are localized in specific regions of the GI tract (2022). In the murine species, some microbial populations are known to associate intimately with epithelial cells by either direct attachment to epithelial cell membranes (8, 9, 11, 12) or layer formation (6, 20, 23). In contrast to what is known about bacterial localization in the GI tract of murine species, little is known about the bacterial attachment and layering in the GI tracts of higher mammals. Most of the available information on the microbial flora of higher mammals has been obtained by culturing fecal samples or lumenal aspirations, which have not yielded any data on bacterial populations that might be intimately associated with the GI surface or crypts of Lie-

berkuhn (10, 20). Also, there are very little data available on the impact of altered environments on the GI tract flora. However, recent studies by Holdeman et al. (13) and Tannock and Savage (24) indicate that stress (psychological, environmental, or dietary) may cause changes in the flora. The aims of this investigation were to determine if the GI tracts of adult, male, beagle dogs possessed localized bacterial populations and to ascertain if different housing conditions would significantly alter their GI flora. MATERIALS AND METHODS Animals and housing conditions. Nine male, purebred, beagle dogs, 9 to 12 months old, were obtained from a closed colony at the Argonne National Laboratory, Argonne, Ill. Two of the dogs (group 1) were housed in the conventional (open) dog-holding facilities at the University of Wisconsin. Seven dogs were housed and maintained in a "locked environment" that consisted of stainlesssteel housing units that are designed to maintain

' Present address: Department of Microbiology, University of Texas Medical Branch, Galveston, TX 77550.

194

VOL. 34, 1977 germ-free dogs. The latter system has been described in detail elsewhere (3). The dogs in the locked environment (group 3) were fed sterile water and diet, and all air entering the unit was filter sterilized. All entries into and out of the unit were made under sterile conditions. In essence, the locked-environment dogs were treated in a manner that prevented outside bacteriological contamination for 2.5 years (3). All nine dogs were fed a steamsterilized diet (Purina Dog Meal, Ralston Purina Co., St. Louis, Mo.) and water ad libitum. After 2.5 years, four dogs were removed from the locked environment and placed in regular (open) dog-holding facilities for 1 month (group 2). During the latter 1month transition period, two of the dogs were fed the sterilized diet, whereas the other two dogs were fed the same diet, but unsterilized, ad libitum. All group 1 and group 2 dogs (after removal from the locked environment) were given nonsterilized water ad libitum. All dogs were killed by an overdose of phenobarbital. We then examined and compared the microbial flora of dogs housed under conventional conditions (group 1), locked-environment conditions (group 3), and transitory conditions, in which dogs were removed from the locked environment and put into a conventional environment (group 2). Tissue sample locations. Two adjacent pieces of tissue from the cardiac region of the stomach, proximal small bowel, distal ileum, cecum, and colon were obtained from dogs by first clamping the width of the GI tracts with hemostats and then cutting, with sterile instruments, tissue pieces (about 1 to 2 cm2) from the GI tract. Upper small bowel tissues were removed from a region about 20 cm proximal to the ileocecal valve. Colonic tissues were taken from a region approximately 20 cm distal to the ileocecal valve. Microbiological methods. For the growth of strict anaerobes, separate pieces of ilea, ceca, and colons were immediately placed into tubes of prereduced transport broth (2) and, within 30 min, passed into an anaerobic glove box (1, 2) through a rapid-entry port, and homogenized in a blender. Pooled samples of either ileal, cecal, or colonic tissues in the groups 1 and 2 dogs housed together and the two group 3 dogs housed together were homogenized. Serial 10fold dilutions were made in Trypticase soy broth (5). Then, 0.1 ml of each dilution (10-' to 10-9) was plated onto prereduced A II agar (2), and the plates were incubated in a transparent, plastic incubator especially designed for the glove box (2a). The latter dilution tubes, after removal from the glove box through the air lock, were used to inoculate enriched and differential media for the detection and identification of aerobic and microaerophilic bacteria. The media and culture conditions were as described elsewhere (3, 5) except that S.F. agar was not used. All media used to grow the anaerobic bacteria were prereduced in the glove box at least 48 h prior to inoculation (2). The anaerobic bacteria were identified by their Gram reaction, biochemical tests, and volatile and nonvolatile fatty acids produced in PYG broth according to the identification protocol estab-

BACTERIA IN BEAGLE GI TRACTS

195

lished by the Virginia Polytechnic Institute Anaerobe Laboratory (14). Volatile and nonvolatile fatty acids were identified (14) with a gas chromatograph (Dohrmann, Mt. View, Calif.). Biochemical tests were done by a modification of the Minitek procedure described previously (5). Other media (litmus milk, chopped meat, gelatin, and egg yolk agar) were prereduced and inoculated with bacteria harvested from A II agar. All biochemical tests were done in the glove box at 35 to 37°C. All morphologically dissimilar colony-forming units (CFUs) were counted on plates yielding 1 to about 300 colonies and inoculated onto A II agar. Total counts of individual CFUs of both aerobes and anaerobes were tabulated. References to CFUs of specific genera represent CFUs per gram (dry weight) of tissue and contents. Stomachs and proximal small bowels were not examined by microbiological methods. Direct counts of bacteria were made from 100-fold dilutions of ileal, cecal, and colonic tissue homogenates with phase optics and a counting chamber (PetroffHauser, Philadelphia, Pa.). Sample preparation for microscopy. Pieces of tissue, taken from areas directly adjacent to those used for CFU determination, were either immediately placed in cacodylate-buffered glutaraldehyde (at 4°C) or put serosal-side down into a vial, which was immediately immersed in liquid nitrogen. This frozen sample was subsequently cleaved with a razor blade, sectioned in a cryostat, and stained with a tissue Gram stain while another piece of the cleaved sample was fixed in glutaraldehyde, as above. This technique was used to preserve delicate surface structures that are often removed by immediate fixation in glutaraldehyde (4). All samples fixed in glutaraldehyde were postfixed in osmium tetroxide, dehydrated in an ethanol and amyl acetate series, critical-point dried, metal coated, and examined with a scanning electron microscope (SEM). Details of the above procedure have already been described (4, 18). A Zeiss Universal microscope was used for all light micrography and either a JEOL or a JEM U3 SEM, operated at 20 kV, was used for all SEM. RESULTS

Genera and species of bacteria. Table 1 lists the genera and species of microorganisms cultivated from all of the dogs. Eighty-four species of bacteria representing 27 genera and five genera of fungi were isolated. This count also includes those organisms only partially identified. Although the quantitative and qualitative composition of the microbial flora differed from sample to sample, similarities of microbial colonization were observed when the most numerous bacteria and their localization sites were shown (Tables 1 and 2). The highest numbers (109 to 1010/g [dry weight]) of viable bacteria were from anaerobic genera (Bacteroides, Bifidobacterium, Peptostreptococcus, Eubacterium, Clostridium, and Peptococcus) and mi-

TABLE 1. Microorganisms isolated from ileal, cecal, and colonic homogenates of beagle cjogsa Highest CFUb Occurrence in dog Species groupsc No. detected Location

Streptococcus S. bovis S. acidominimus S. mitis Streptococcus species 2 S. salivarius S. dysgalactiae S. intermedius S. faecium Streptococcus species ld S. agalactiae S. faecalis Streptococcus species 3 Lactobacillus acidophilus L. casei subsp. rhamnosus L. leichmannii L. plantarum L. fermentum L. lactis L. crispatus L. casei subsp. casei

3 3 1 2 4 3 3 6 5 6 5 1

2 8 5 4 3 3 3 2 3 1 3

L. minutus L. helveticus L. salivarius subsp. salivarius

Bifidobacterium infantis lactentis B. adolescentis Bacteroides vulgatus B. distasonis B. corrodens B. amylophilus B. capillosus Bacteroides species B. furcosus Eubacterium ventriosum E. ruminantium Eubacterium species 1 E. lentum E. contortum E. cellulosolvens E. aerofaciens Clostridium Clostridium species 1 Clostridium species 2 C. inulinum C. irregularis C. cochlearium C. perfringens C. histolyticum Clostridium species 4 Clostridium species 3 Peptostreptococcus P. magnus P. micros P. productus Peptostreptococcus species 1 P. parvulus P. anaerobius Peptococcus P. constellatus

x x x x x

x x x x

x x X

x x

x x

x x x

x x X

x

10O0 1O"° 109

108 107 107 107 106 106 104 104 102 1010 108 108 108 105 105 106 106 105 104 103

2 x 1010 3 x 107

Ileum

Colon Cecum Colon Colon Colon Colon Colon Colon Ileum Ileum Ileum

1, 2 1

1, 2, 3 1 1, 2 2 2 2, 3 1, 2, 3 2 2, 3 2 1, 2, 3 3 2, 3 3 3 2, 3 2 2 3 3 2 2, 3 2 1, 2, 3 1, 2, 3

6 6 5 1 3 2 6

x 1010 x 1010 x 1010 x 109 x 109 x 109 x 108 x 109 x 109 x 109 x 109 x 108 x 108 x 107

Colon Cecum Colon Colon Colon Colon Colon Colon Cecum Colon Colon Colon Colon Colon Ileum Cecum

5 2 2 1 9 3 3 3 2

x 109 x 109 x 109 x 109 x 108 x 108 x 107 x 107 x 107

Colon Colon Ileum Colon Colon Cecum Cecum Ileum Cecum

1, 2, 3 1, 2, 3 2 3 3 3 3 2

6 5 3 6 3 3

x 109 x 109 x 108 x 107 x 107 x 107

Colon Colon Cecum Cecum Cecum Ileum

2, 3 1, 2 2 2 3

Colon

3

1 1 1 9 1 1 2

3 x 109 196

Colon Colon Colon Cecum Cecum Ileum Colon Colon Colon Ileum Ileum

2, 3 3 3 1, 2 3

1 1 1, 2, 3 3 3 2 2

2

1

TABLE 1.-Continued Highest CFUb

Species P. prevotii P. magnus

Fusobacterium F. necrogenes F. varium F. gonidiaformans

No. detected 3 x 108 6 x 107

Location Colon Ileum

Occurrence in dog groupsc group5 1 2

3 x 101

Cecum

3

3 x 108

Cecum

3

3 x 107

Cecum

3

Gaffkya anaerobica

3 x 108

Cecum

3

Veillonella parvula

2 x 108

Cecum

3

Acidaminococcus fermentans

2 x 108

Cecum

2

Corynebacterium C. pseudotuberculosis Corynebacterium species 1 Corynebacterium species 2 C. ovis

3 x 107 3 x 106 3 x 105 1 x 105

Ileum Ileum Ileum Ileum

1, 3 1, 3

Ruminococcus albus

3 x 107

Cecum

2

Propionibacterium acnes

7 x 107

Colon

2

Proteus mirabilis

5 x 106

Cecum

1, 3

Escherichia coli

4 x 106

Colon

1, 2, 3

Klebsiella pneumoniae

3 x 106

Colon

3

Enterobacter cloacae

1 x 106

Cecum

2, 3

Staphylococcus aureus

2 x 105

Ileum

1, 2, 3

Bacillus B. coagulans B. pantothenticus

1 x 105 3 x 104

Ileum Ileum

1 1

Anaerobiospirillum succiniciproducens

1 x 105

Colon

3

Moraxella (group M-5)

6 x 104

Ileum

2

Micrococcus species

5 x 104 1 x 104 4 x 103

Ileum Ileum Ileum

3 3 3

Othersd Gram-negative rod Gram variable

6 x 106 2 x 105

Cecum Ileum

3 2

Alternaria species

1 x 105

Colon

1

Rhodotorula rubra

9 x 103

Ileum

3

Cladosporium species

6 x 103

Ileum

3

Unidentified fungus

4 x 103

Colon

3

Micrococcus M. cryophiles M. varians

3 3

Coccobacillus

Trichosporon cutaneum 3 x 103 Ileum 3 a Ileal, cecal, and colonic tissues and contents were homogenized with a blender in an anerobic glove box to quantitate anaerobes. Homogenates were removed from the glove box and plated onto selective media for 197

198

APPL. ENVIRON. MICROBIOL.

DAVIS ET AL.

TABLE 1. -Continued

aerobes, facultative anaerobes, and fungi. b CFU per gram of tissue and contents (dry weight). c Indicates which dog groups (1, conventionally housed; 2, transitionally housed; 3, housed in a locked environment) possessed detectable microbes in their GI tracts. d Those organisms listed as either unknown, species, or other were either unidentifiable by our methods or did not correspond to any known species or subspecies. TABLE 2. Viable and microscopic counts of bacteria in the ilea, ceca, and colons of beagle dogs Region of tract Counting method counted

Environmental conditionsa

Group 3 3.1 x x 109 1.4 x 106, 4.0 x 108 Cultureb Ileum NDd 4.0 x x 109 1.1 x 109 Microscopicc 3 x 107 4.5 x x 109 2.4 x 109, 7.7 x 109 Cecum Culture 6.0 x 109 x 109 ND 6.2 x Microscopic x 109 3.4 x 101', 4.6 x 1010 8.6 x 1010 1.9 x Culture Colon 4.4 x 10"0 x 1010 ND 3.0 x Microscopic a Three different environmental conditions were used: group 1 dogs were housed in a conventional dogholding facility; group 2 dogs were housed in a locked environment (sterile food, water, and air) for about 2.5 years and then removed to a conventional environment for 1 month; and group 3 dogs were housed in the locked environment for 2.5 years. b Total CFU of all bacteria obtained per gram (dry weight) of tissue and lumenal contents from a single dog or two dogs in the same group. c Direct counts of undiluted or 100-fold dilutions of bacteria, observed with phase optics, were counted in a Petroff-Hauser counting chamber. d ND, Not done. Group 1 1.2 x 108

croaerophilic to anaerobic Lactobacillus species (Table 1). The highest number of facultative anaerobes was almost always streptococci. Microscopic observation of diluted contents indicated that ceca and ilea usually possessed about one to three logs less bacteria than did the colon (Table 2). When total microscopic counts were compared with total CFUs, the data suggested that most of the bacteria viewed with phase optics were viable (Table 2). The ilea of all three groups of dogs possessed a heterogeneous microbial flora. The ileal anaerobes also showed a wider variation in viable bacteria (103 to 109 [Table 3]) than did anaerobes isolated from either the cecum (107 to 109 [Table 4]) or the colon (108 to 1010 [Table 5]). The total number of aerobic and facultative bacteria in the ilea (105 to 107 [Table 3]) of the dogs was comparable to those found in their ceca (105 to 108 [Table 4]), but less numerous than those aerobes and facultative anaerobes found in the colons (107 to 1010 [Table 5]). Whereas 13 species of bacteria were isolated from the ilea of conventional dogs (group 1), 21 and 20 species were isolated from the ilea of group 2 and group 3 dogs, respectively. In addition, fungi (three genera) were mainly isolated from group 3 dogs. There was a remarkable difference in the microbial flora in the ceca of dogs housed conventionally (group 1) when compared with the two other groups of dogs (Table 4). Convention-

Group 2 107, 4.4 107, 7.6 108 1.9 107 1.5 108 , 9.6 1010, 1.8

ally housed dogs (group 1) possessed only two genera of microbes (Bacteroides and Streptococcus) in their ceca, whereas the dogs in groups 2 and 3 possessed 8 and 13 different genera, respectively (Table 4). The ceca of conventionally housed dogs also had lower total CFUs of anaerobes (one to two logs) and lower total CFUs of aerobes (about one-half to three logs) (Table 4) then did the ceca of dogs in groups 2 and 3. Four different microbial species were isolated from the ceca of conventional dogs, whereas isolated dogs of group 3 had 25 different species, and group 2 dogs had 18 different species present in their ceca. The dogs, regardless of housing conditions, generally showed a similar colonization pattern in their colons by the various genera of anaerobes. Dogs removed from the locked environment and then placed into a conventional environment (group 2) showed a decrease of 0.5 to 1 log in the total number of colonic anaerobes (Table 5). In addition, the dogs housed under conventional conditions (group 1) had a population of Streptococcus bovis, S. acidominus, and S. mitis that was 1 to 2 logs higher (109 to 1010) than that found in the colons of other group 2 and 3 dogs (Table 5). The dogs in confinement (group 3) had 26 different microbial species in their colon. The dogs in transition (i.e., group 2) had 22 different species of bacteria in their colon, whereas conventional (group 1) dogs had 16 different species in the colonic contents.

VOL. 34, 1977

BACTERIA IN BEAGLE GI TRACTS

199

TABLE 3. Predominant microbes isolated from the ilea of beagle dogs housed in conventional or locked

environments a Predominant microbes isolated from: Genera

Group 1 dogs

Group 2 dogs

Anaerobic and facultative

Bacteroides (1)b 107c Peptostreptococcus (1) 107

Total anaerobes

1.2 x l0"d

Aerobic and facultative

Total aerobes

Group 3 dogs

Clostridium (4) 107-9 Eubacterium (2) 108 Peptococcus (1) 107 Bifidobacterium (1) 108 Lactobacillus (3) 103-7 6.0 x 103, 4.4 x 109

Bacteroides (1) 107 Clostridium (2) 105-6 Eubacterium (2) 1067 Bifidobacterium (1) 101 Lactobacillus (4) 104-5 4.0 x 108, 9.0 x 105

Streptococcus (6) 1046 Staphylococcus (1) 105 Corynebacterium (2) 1067 Bacillus (2) 104-5

Streptococcus (6) 102-7 Staphylococcus (1) 105 Moraxcella (1) 104 Enterobacter (1) 102 Escherichia (1) 104

6.6 x 107

4.2 x 105, 3.1 x 107

Streptococcus (4) 104-5 Staphylococcus (1) 105 Corynebacterium (1) 105 Micrococcus (3) 1034 Klebsiella (1) 105 Rhodotorula (1) 103 Cladosporium (1) 103 Unidentified fungus (1) 103 6.5 x 105, 1.4 x 106

See footnote a, Table 2. Parentheses indicate the number of species found in the genus. c Indicates the total population number(s) of the species within the genus. d Indicates the total number of all genera (two numbers indicate that two different samples were examined). a

b

TABLE 4. Predominant microbes isolated from the ceca of beagle dogs housed in conventional or locked environments a Predominant microbes isolated from: Genera Group 1 dogs

Group 2 dogs

Group 3 dogs

Anaerobic and facultative

Bacteroides (l)b 107c

Bacteroides (3) 107-8 Peptostreptococcus (2) 107Clostridium (4) 1068 Eubacterium (2) 107-8 Bifidobacterium (2) 107-8 Lactobacillus (1) 107

Total anaerobes

3 X

4.5 x 108, 1.9 x 109

Bacteroides (4) 1019 Peptostreptococcus (1) 107 Clostridium (6) 107Eubacterium (1) 106 Lactobacillus (1) 106 Fusobacterium (3) 107-8 Veillonella (1) 108 Peptococcus (1) 108 2.4 x 109, 7.7 x 109

Aerobic and facultative

Streptococcus (3) 104-5

107d

Total aerobes 6.5 x 105 a-d See footnote a, Table 2 and footnotes

Streptococcus (3) 1067 Escherichia (1) 105

1.0 X 106, 3.1 x 107

Streptococcus (3) 108 Proteus (1) 106 Enterobacter (1) 106 Klebsiella (1) 105 Alternaria (1) 105 2.2 x 108, 8.0 x 108

b-d, Table 3.

Tables 3, 4, and 5 indicate that the dogs Fusobacterium and Veillonella species (10' to housed in the locked environments possessed a 108) in their ceca (Table 4), whereas the convenmore complex microbial flora (22 genera and 52 tional (group 1) and transitional (group 2) dogs species) in their ilea, ceca, and colons than the did not possess detectable CFUs of the latter conventionally housed group (12 genera and 26 two genera at any site sampled. There were species). The conventionally housed dogs other minor genera differences in the three (group 1) did not have detectable Bifidobacter- groups of dogs, but these differences were found ium species in any region of the GI tract. The in the minor components (103 to 106) of the latter genus was a predominant member of the aerobic-facultative microflora. For example, microbial flora (10" to 10'0) of the dogs in groups Staphylococcus aureus, which was found in the 2 and 3. Additionally, group 3 dogs harbored ilea of all dogs, was not found in the ceca of any

200

DAVIS ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 5. Predominant microbes isolated from the colons of beagle dogs housed in conventional or locked environments I Predominant microbes isolated from:

Genera Group 1 dogs Bacteroides (3)b 1O>'0 Peptostreptococcus (1) 109 Peptococcus (1) 108 Clostridium (2) 109 Eubacterium (3) 109 Lactobacillus (1) 1010

Group 2 dogs Bacteroides (4) 1074 Peptostreptococcus (2) 109 Clostridium (2) 108 Eubacterium (1) 109 Lactobacillus (5) 10 Bifidobacterium (1) 109

Total anaerobes

8.6 x lOlod

1.9 x 108, 9.6 x 109

Aerobic and facultative

Streptococcus (3) 10910"0 Escherichia (1) 105 Proteus (1) 104

Streptococcus (4) 108 Escherichia (1) 106 Enterobacter (1) 106 Staphylococcus (1) 104

Total aerobes

6.0 x 1010

5.6 x 107, 1.5 x 108

Anaerobic and microaerophillic

Group 3 dogs Bacteroides (3) 101 Peptostreptococcus (1) 109 Peptococcus (1) 109 Clostridium (3) 107-8 Eubacterium (3) 109 Lactobacillus (3) 107 Bifidobacterium (1) 1010 Anaerobiospirillum (1) 105 3.4 x 10"', 4.6 x 1010

Streptococcus (5) 101 Escherichia (1) 105 Proteus (1) 105 Enterobacter (1) 105 Klebsiella (1) 105 Unidentified fungus (1) 103 3.0 x 108, 8.0 x 108

See footnote a, Table 2. Parentheses indicate the number of species found in the genus. Indicates the total population number(s) ofthe species within the genus per gram (dry weight) oftissue and contents. d Indicates the total number (per gram [dry weight] of tissue and contents) of all genera (two numbers indicate that two different samples were examined). a b

I

of the dogs and was only found in the colons of surfaces, but the bacteria occurred singly or in group 2 dogs, and a Veillonella species was only a small microcolony (Fig. 3). No population of isolated from the ceca of group 3 dogs (Table 4). segmented filamentous bacteria was observed One consistent finding was that Streptococcus to associate with the proximal small bowel epiwas usually the most numerous of the faculta- thelium in any of the dogs studied (Table 6). Dogs housed either in the conventional or tive and aerobic genera, whereas the most numerous of the anaerobic genera varied from locked environments possessed no bacterial populations that attached or layered to the dissample to sample. Although many genera were consistently iso- tal ileum. Distal ilea in three out of four dogs lated from ilea, ceca, and colons of these three from group 2, however, had segmented filamengroups of dogs, the species of these genera did tous organisms attached to their ileal epithelial not segregate into any discernible groups ac- cells (Table 6). The attachment sites and segcording to the dog housing conditons. Two spe- ments of the microorganisms were easily obcies, however, did occur in every dog examined: servable in SEM examination of the dogs in Streptococcus mitis and a Eubacterium species. group 2 (Fig. 4a). Both frozen and glutaraldeThis Eubacterium was unique because it pro- hyde-fixed specimens indicated that the segduced long filaments and spontaneously lysed mented filamentous microbes were located underneath a mucin layer (Fig. 4a and b). after 24 h of incubation. Microscopy: light and scanning. In all of the Ceca from the three dog groups showed variadog stomachs, the cardiac epithelium was usu- bility in the possession of a bacterial layer on ally covered with a thick layer of mucin (Fig. their cecal surfaces. Conventionally housed 1), and bacteria were often undetectable. To dogs showed almost no bacteria on the epithefind any bacteria on the stomach epithelium, lial surface; similarly, two dogs from each ofthe many fields had to be observed with specimens other two groups of dogs (2 and 3) housed under prepared for both light microscopy and SEM. different conditions showed either no bacteria Close examination of the stomach surface on the cecal epithelium or, at the most, a few showed that two of the nine dogs possessed a regions with layering bacteria. In those ceca bacterium that had unusually tight helical coils that lacked a layer of bacteria on a large porand bipolar flagella (Fig. 2). Unlike the murine tion of their cecal epithelium, a layer of mucin stomach, no bacterial population layered or at- was present. Mucin predominated in the ceca of tached to the stomach epithelium of the dog dogs housed conventionally (Fig. 5), which corresponded with a relatively lower (107 versus (Table 6). The proximal small bowel contained a sparse 108 to 9) bacterial count (Tables 2 and 5). Ceca bacterial population. Bacteria could occasion- with a bacterial layer (five out of nine ceca ally be found in the GI lumen adjacent to villus [Table 5]) showed a predominantly heteroge-

FIG. 1. SEM of the cardiac region of a dog stomach covered with a layer of mucin. Arrows show where mucin layer has been split (probably a drying artifact) to reveal the underlying epithelium. x45. FIG. 2. An unusual helical coiled bacterium with bipolar flagella (arrows) on a dog stomach. xl0,650. FIG. 3. Proximal small bowel villi from a dog that shows mucin strands (double arrows) and a single bacterium (arrow) adjacent to the villi. x275. FIG. 4. Segmented filamentous microbes, found only in dogs in transition from a locked to a conventional environment (group 2 dogs), attached directly to distal ileal epithelial cells. Double arrows (a) indicate the individual segments ofthese organisms and the single arrows show their attachment site. x920. (b) shows the mucin that covers the microbes in the distal ileum. x960. 201

202

DAVIS ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 6. Layering and attachment of microbes in the GI tracts of beagle dogs housed in conventional or locked environments a Environment

Microscopy

Stomach

Group 1

Light SEM Light SEM Light SEM

0/2b

Group 2 Group 3

0/2 0/4 0/4 0/3 0/3 0.9

Proximal ileum Distal ileum

0/2 0/2 0/4 0/4 0/3 0/3 0/9

0/2 0/2 3/4 3/4 0/3 0/3 3/9

Cecum

Colon

0/2 0/2 3/4 2/4 1/3 2/3 5/9

2/2 4/4 2/4 1/3 2/3 8/9

2/2

Total See footnote a, Table 2. bNumber of dogs with layering or attaching microorganisms per total number of dogs examined by either light microscopy or SEM. a

bacterial population on the cecal epithelium. The cecal layer, when it was observed, consisted mainly of gram-positive rods and cocci and some gram-negative (mainly rods) bacteria (Fig. 6 and 7). Layers of bacteria, composed primarily of gram-positive rods and cocci, were most prominent on colonic epithelium of the dogs from groups 2 and 3 (see Fig. 7 for similar results from the cecum). Dogs of group 1 and occasionally those of groups 2 and 3 (in some regions of the colon in dogs of the latter two groups) manifested a distinct microbial population that consisted of gram-negative spiral- and rod-shaped microbes. This layer, which was distinct from the layer shown in Fig. 7, of bacteria separated the bulk of the lumenal flora from the colonic epithelial cells (Fig. 8) and was often difficult to see unless the tissue-gram-stained sections were viewed with phase optics (cf. Fig. 9 with Fig. 8). SEM observation of the latter microbes, which were most numerous in the colons of group 1 dogs, indicated that the population was heterogenous with respect to the rods and spirals present (Fig. 10a and b). All of the dogs with either a predominantly gram-positive or a gram-negative layer in either the cecum or the colon possessed spiraland rod-shaped bacteria in their crypts of Lieberkuhn. Some crypts had only spiral- or rodshaped bacteria, whereas other crypts had both types present (Fig. 9). Dogs were also able to have bacteria in the crypts without a discernible layer of bacteria on their epithelial cells. The density of this microbial population associated with crypts varied from crypt to crypt in the cecum and colon of each animal. Some crypts possessed large numbers of bacteria (Fig. 11), whereas other crypts had few or none. Conventionally housed dogs (group 1) had easily detectable bacteria within their crypts; dogs in group 2 had only a few bacteria in their crypts, and only one of the three dogs from group 3 showed any bacteria in its crypts of Lieberkuhn. nous

DISCUSSION Almost all previous studies of the microbial flora of dogs have been limited to an elucidation of the fecal or nasopharyngeal flora (3, 22). Studies of the bacterial flora located in and on the surface of canine GI tissues usually reported only the total counts of genera cultivated on selective media; identification of species, especially of the anaerobic genera, was rarely attempted (3, 22). This study detailed the genera and species, as far as they could be identified with our methods, in the distal ileum, cecum, or colon of the dogs. It indicated that bacteria did not localize on the epithelial surface of dog stomach or proximal small bowel but could localize in the distal ileum, cecum, and colon. Total counts of anaerobes were about 1 to 2 logs higher than were the total counts of aerobes, except in the conventionally housed dogs, whose colon counts yielded approximately the same numbers (1010) of anaerobic and facultative bacteria. Such high numbers of aerobic and anaerobic bacteria have been reported previously in dog feces (3), although the anaerobes were usually about 1 log higher than were the aerobes (108 to 9 versus 109 to 10). We observed a wide variation in isolatable bacterial species from dog to dog regardless of housing conditions. However, an examination of the predominant genera isolated from all three groups of dogs showed only one case where a major difference in the number of genera was detected. This difference in genera was observed in the ceca of the conventionally housed group 1 dogs where only two genera, Bacteroides and Streptococcus, were detected. The latter observation was the most obvious part of a general trend that suggested that conventionally housed dogs possessed a less complex flora, in terms of genera and species present, than either of the other groups of dogs (cf. Tables 3, 4, and 5) that were housed in a locked environment. A study on the fecal flora of man in locked environments showed that his fecal flora does

M

,

-

-

FIG. 5. An example of the thick layer of mucin (m) that covers the dog cecum and overlays most ofthe cecal epithelium (e). Bacteria were difficult or impossible to locate in such ceca (in the mucin or on the epithelium) in contrast to other dog ceca that possessed a layer of bacteria (see Fig. 6). x95. FIG. 6. Light micrograph of a frozen section of the gram-positive layer of bacteria on a dog's colonic epithelium (e) stained with a tissue Gram stain. This type oflayer was also seen on the cecum. Arrows indicate some of the gram-negative bacteria that also could be observed in this layer. x1,320. FIG. 7. SEM of the gram-positive layer of bacteria found on a dog's cecal surface. Note the predominance of rods, cocci, and coccobacilli, and compare this figure with Fig. 10. Arrow indicates the epithelial cell surface. x3,500. FIG. 8. Light micrograph of a layer (shown by bars) that separates the colonic epithelium (lower left side) from the lumenal contents (right side). A weakly staining gram-negative (spiral- and rod-shaped) bacterial population inhabits this layer. The latter bacteria are best viewed with light microscopy at a high magnification and with phase optics (see Fig. 9). x195. FIG. 9. Light micrograph of a frozen section of a dog colon (tissue Gram stained) that shows spiral- and rod-shaped bacteria in the layer of mucin. x2,365. 203

204

DAVIS ET AL.

APPL. ENVIRON. MICROBIOL.

not markedly change with a short-term (about 2 months) confinement; however, subtle alterations can take place (13). Other studies on confinement suggest that the microbial and fungal flora increase after short- (16 days) or long-term (12 months) confinement (17, 25, 26). Our study suggests that in dogs housed in a locked environment for over 2 years, the colonic flora does not markedly change, but some other changes, such as an increase in the diversity of genera and species present, may occur in the cecal and ileal flora. The group 2 dogs had a flora intermediate in complexity between groups 1 and 3 (cf. Tables 3 through 5), which suggests that upon removal ofthe dogs from a locked environment, the flora begins to simplify and thus resembles the flora in conventionally housed animals. Most fungi were detected in those dogs kept in the locked environment (group 3). Results from our study should not be construed to mean that because organisms (both bacterial and fungal) were not cultured they were not present. Our results simply mean that the organisms were not detected by our culture techniques; they could be present in such low numbers (1 to 103) that many culture samples and highly selective media would be needed to detect them. Another source of error may be due to selection of strains to study on the basis of colony counts; bacterial species that have colonies visibly identical to colonies of strains isolated may belong to different species. Also, only 1 composite sample from each site in dogs of group 1 were studied. In addition, AII medium, although one of the best enriched nonselective media we have used for growing anaerobes, probably did not allow us to culture all of the bacteria present. For example, morphologically distinct microbes (spirals, types a, b, and c, Fig. 10) were never cultured. Thus, SEM results show that not all the bacteria present can, as yet, be cultured. Although no difference in the microflora was observed in the stomachs or proximal small bowels of dogs housed under different conditions, changes in the microflora were pronounced when their ileal, cecal, and colonic

FIG.

10.

Heterogenous population of spiral-

and

rod-shaped bacteria found adjacent to a dog's colonic epithelium (A). At least four morphologically dis-

tinct spiral-shaped bacteria can be observed: (a) short and thin spiral; (b) wavy outer-enveloped covered spiral (see also [B] for types a and b); (c) spiral (vibrio) with a single polar flagellum; (d) spiral with bipolar flagella. x3,000. (B) Enlargement of(A) that shows a and b organisms. x9,600. (C) Rods (e,f), coccobacilli (g,h), and cocci (i). Note the fine filaments attached to organism h. x9,600. FIG. 11. Colonic crypt of Lieberkuhn in a light micrograph showing a predominant population of thin, gram-negative rods. The "c"-shaped structures are goblet cells (G) that empty into the crypt. x3,015.

VOL. 34, 1977

surfaces were examined by light microscopy or SEM. For example, only group 2 dogs, in transition from the locked environment to the conventional, showed segmented filamentous organisms in their ilea. In addition, the bacterial population that comprised the epithelial cell layers differed; the conventionally housed dogs possessed a distinct gram-negative bacterial layer adjacent to the colonic epithelium, but the locked-environment dogs showed mainly a gram-positive bacterial layer. Furthermore, obvious differences between group 1 and groups 2 and 3 were observed at the frequency with which crypts of Lieberkuhn were populated by bacteria; confinement appeared to suppress crypt populations. Thus, differences in the layering and localization of the microbial flora were noted in the three groups of dogs, but the reasons for the changes are not clear. It is known that either antibiotics, diet, or environmental stress, or a combination of these factors, can alter the composition, attachment, and layering of the GI flora (9, 19, 24). Of these factors, only the housing environment was varied in our study, and flora changes were noted.

Although each of two dogs (group 2) was fed autoclaved and nonautoclaved Purina Dog Meal, their flora was not markedly changed from that of the other group 2 dogs. The observation that segmented filamentous organisms were found in group 2 dogs but did not occur in dogs housed under either conventional or locked environments is of interest but should be confirmed by further studies. Three samples (one for light microscopy and two for SEM) were taken from adjacent regions of each dog ileum, but such sampling may not detect the organisms. For example, it is known that the segmented filamentous microbes colonize nonuniformly the ilea of the murine species; that is, the segmented filamentous microbes may be found both proximal and distal to a region of the ileum that possesses no such organisms (8). Although it is very unlikely that we would miss by chance segmented filamentous microbes in five of the nine dogs (groups 1 and 3) and find them in three out of the four dogs of group 3, it is possible. To our knowledge, no other reports on the occurrence of these segmented filamentous bacteria in beagles or other dogs are available. This study indicates that specific regions of the GI tracts of higher mammals such as dogs can possess populations of bacteria that attach and localize, either in crypts of Lieberkuhn or in layers. Attachment of segmented filamentous microbes to epithelial cells paralleled that found in murine species (4, 8, 12), but the frequency of their occurrence in dogs was less. The

BACTERIA IN BEAGLE GI TRACTS

205

layering of bacteria in the canine GI tract was somewhat different from that described in murine species (20). In dogs, bacterial populations adjacent to the epithelial cells were much more sparse and contained markedly fewer fusiformshaped bacteria than those bacterial populations found in mice and rats. Additionally, another type of layer, composed mainly of grampositive organisms, was observed in the dogs. In mice, a predominance of gram-positive microbes adjacent to the large bowel epithelium strongly suggests that they have an altered microbial flora (19). The occurrence of bacteria (mainly spiral-shaped microbes) has been noted in the crypts of Lieberkuhn of rat ceca (7) and in the large intestines of randomly bred dogs (15). Our observations in dogs generally agree with the latter findings. If the most numerous bacterial genera and species from the GI tracts of dogs are compared with those found in the GI contents of man (W.E.C. Moore, M. Ryser, and L. V. Holdeman, Abstr. Annu. Meet. Am. Soc. Microbiol. 1975, DS7, p. 61), several differences and similarities in the type and number of genera present can be found. For example, the most numerous genera uniformly found in the dog ilea and large bowel are Bacteroides and Streptococcus (S. bovis and S. acidominimus). Although the latter are among the predominant genera, they are not the most numerous bacteria in the GI contents of man. Fusobacterium, a predominant genus found in man, was not as numerous in dogs (approximately 1010 `O 11 versus 107 to 8). The Clostridium species predominant in man are easily identified but those in the dog are not (Table 1). Genera that are numerous in both the GI tract of dogs and man are Lactobacillus, Bifidobacterium, Eubacterium, Bacteroides, and Peptostreptococcus. Localization of GI flora in man, in terms of either bacterial layer formation or bacterial populations in crypts of Lieberkuhn have not, to our knowledge, been reported. Our work with dogs suggests that other higher mammals, including man, may possess such localized bacterial populations. Our results and those of other investigators (3, 13, 17) do not support the concept that the microbial flora of mammals will undergo a dramatic change (simplification to one or two genera) if confined to a locked environment and fed sterile food and water for a long time period (16). Indeed, our results indicate that with long-term confinement, the flora becomes more complex, and its distribution along the GI tract changes. We hypothesize that microorganisms that are not predominate members of the microflora (transient microflora), but which are initially carried into the locked environment by

206

DAVIS ET AL.

any animals, will have to adapt to the environment to survive. In addition, the chance for reassociation with the animal in a locked environment is much greater than in a conventional environment. Thus, adaptation to the locked environment and frequent reassociation (i.e., ingestion) could select for those microbes that are not predominate members of the flora. Ultimately, successful competition could lead to an increase in the number and variety of microorganisms in the GI tracts of the animals within the locked environments. Alternatively, a hypothesis that the microenvironments of GI tracts in animals held in locked environments may change with time should also be considered. Such an alteration in the GI tract also could allow other less predominate species to proliferate. As indicated by our data, whether prolonged contact with such an increase in the numbers and variety of microorganisms could pose a threat to a host is simply not known. Nonetheless, the flora should be monitored carefully when humans are confined in close chambers for any considerable length of time. ACKNOWLEDGMENTS We would like to thank Bangio Wong and Jim Brown for their excellent technical assistance. This study was supported by Public Health Service Medical Microbiology and Immunology Training grant A10045104 from the National Institute of Allergy and Infectious Diseases, NASA grant NGR-50-002-191, and NIH RR 05427.

APPL. ENVIRON. MICROBIOL.

9.

10. 11.

12.

13.

14.

15.

16. 17.

18.

LITERATURE CITED 1. Aranki, A., and R. Freter. 1972. Use of anaerobic glove boxes for the cultivation of strictly anaerobic bacteria. Am. J. Clin. Nutr. 25:1329-1334. 2. Aranki, A., S. Syed, E. Kenney, and R. Freter. 1969. Isolation of anaerobic bacteria from human gingiva and mouse cecum by means of a simplified glove box procedure. Appl. Microbiol. 17:568-576. 2a.Balish, E., J. F. Brown, and T. D. Wilkins. 1977.

Transparent plastic incubator for the anaerobic glove box. Appl. Environ. Microbiol. 33:525-527. 3. Balish, E., C. Shih, C. E. Yale, and A. Mandel. 1974. Effect of a prolonged stay in a locked environment on the microbial flora in dogs. Aeros. Med. 45:1248-1254. 4. Davis, C. P. 1976. Preservation of gastrointestinal bacteria and their microenvironmental associations in 5.

6. 7.

8.

rats by freezing. Appl. Environ. Microbiol. 31:304312. Davis, C. P., D. Cleven, J. Brown, and E. Balish. 1976. Anaerobiospirillum, a new genus of spiral-shaped bacteria. Int. J. Syst. Bacteriol. 26:498-504. Davis, C. P., J. S. McAllister, and D. C. Savage. 1973. Microbial colonization of the intestinal epithelium in suckling mice. Infect. Immun. 7:666-672. Davis, C. P., D. Mulcahy, A. Takeuchi, and D. C. Savage. 1972. Location and description of spiralshaped miroorganisms in the normal rat cecum. Infect. Immun. 6:184-192. Davis, C. P., and D. C. Savage. 1974. Habitat, succes-

19. 20. 21. 22.

23.

24.

25. 26.

sion, attachment, and morphology of segmented, filamentous microbes indigenous to the murine gastrointestinal tract. Infect. Immun. 10:948-956. Davis, C. P., and D. C. Savage. 1976. Effect of penicillin on the succession, attachment, and morphology of segmented filamentous microbes in the murine small bowel. Infect. Immun. 13:180-188. Drasar, B. S., and M. J. Hill. 1974. Human intestinal flora. Academic Press Inc., New York. Erlandsen, S. L., and D. G. Chase. 1974. Morphological alterations in the microvillous border of villous epithelial cells produced by intestinal microorganisms. Am. J. Clin. Nutr. 27:1277-1286. Erlandsen, S. L., A. Thomas, and G. Wendelschafer. 1973. A simple technique for correlating SEM and TEM on biological tissue originally embedded in epoxy resin for TEM, p. 350-365. In 0. Johari and I. Covin (ed.), Scanning electron microscopy, 1973. ITT Research Institute, Chicago. Holdeman, L. V., I. J. Good, and W. E. C. Moore. 1976. Human fecal flora: variation in bacterial composition within individuals and a possible effect of emotional stress. Appl. Environ. Microbiol. 31:359-375. Holdeman, L. V., and W. E. C. Moore (ed.). 1972. Virginia Polytechnic Institute anaerobe laboratory manual. Virginia Polytechnic Institute, Blacksburg, Va. Leach, W. D., A. Lee, and R. P. Stubbs. 1973. Localization of bacteria in the gastrointestinal tract: a possible explanation of intestinal spirochaetosis. Infect. Immun. 7:961-972. Luckey, T. 0. 1966. Potential microbic shock in manned aerospace systems. Aerosp. Med. 37:1223-1228. Puleo, R. J., G. S. Oxborrow, N. D. Fields, C. M. Herring, and L. S. Smith. 1973. Microbiological profiles of four Apollo spacecraft. Appl. Microbiol. 26:838-845. Savage, D. C., and R. Blumershine. 1974. Surface-surface associations in microbial communities populating epithelial habitats in the murine gastrointestinal ecosystem: scanning electron microscopy. Infect. Immun. 10:240-250. Savage, D. C., and R. Dubos. 1968. Alterations in the mouse cecum and its flora produced by antibacterial drugs. J. Exp. Med. 128:97-110. Savage, D. C., R. Dubos, and R. W. Schaedler. 1968. The gastrointestinal epithelium and its autochthonous bacterial flora. J. Exp. Med. 127:67-75. Smith, H. W. 1965. Observations on the flora of the alimentary tract of animals and factors affecting its composition. J. Pathol. Bacteriol. 89:95-122. Smith, H. W., and W. E. Crabb. 1961. The fecal bacterial flora of animals and man: its development in the young. J. Pathol. Bacteriol. 82:53-66. Suegara, N., M. Morotomi, T. Watanabe, Y. Kawai, and M. Mutai. 1975. Behavior of microflora in the rat stomach: adhesion of lactobacilli to the keratinized epithelial cells of the rat stomach in vitro. Infect. Immum. 12:173-179. Tannock, G. W., and D. C. Savage. 1974. Influences of dietary and environmental stress on microbial populations in the murine gastrointestinal tract. Infect. Immun. 9:591-598. Taylor, G. R., M. R. Henney, and W. I. Ellis. 1973. Changes in the fungal autoflora of Apollo astronauts. Appl. Microbiol. 26:804-813. Zaloguyey, S. N., T. G. Utkina, and M. M. Shin Kareva. 1971. The microflora of human integument during prolonged confinement. Life Sci. Space Res. 9:5559.