Microbial Ecology

, Volume 60, Issue 3, pp 691–702

Characterization of Housing-Related Spontaneous Variations of Gut Microbiota and Expression of Toll-Like Receptors 2 and 4 in Rats

Authors

  • Evangelina Terán-Ventura
    • Department of Cell Biology, Physiology and ImmunologyUniversitat Autònoma de Barcelona (UAB)
  • Mercè Roca
    • Centre de Recerca en Sanitat Animal (CRESA)
  • Maria Teresa Martin
    • Department of Cell Biology, Physiology and ImmunologyUniversitat Autònoma de Barcelona (UAB)
    • Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd)Instituto de Salud Carlos III
  • Maria Lourdes Abarca
    • Department of Animal Health and AnatomyUniversitat Autònoma de Barcelona (UAB)
  • Vicente Martinez
    • Department of Cell Biology, Physiology and ImmunologyUniversitat Autònoma de Barcelona (UAB)
    • Department of Cell Biology, Physiology and ImmunologyUniversitat Autònoma de Barcelona (UAB)
    • Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd)Instituto de Salud Carlos III
    • Edificio V, Unidad de FisiologíaUniversitat Autònoma de Barcelona
Environmental Microbiology

DOI: 10.1007/s00248-010-9737-z

Cite this article as:
Terán-Ventura, E., Roca, M., Martin, M.T. et al. Microb Ecol (2010) 60: 691. doi:10.1007/s00248-010-9737-z
  • 214 Views

Abstract

Gut microbiota has been suggested as a key component of gut homeostasis, affecting immune responses within the gut. We determined changes in intestinal commensal bacteria and expression of toll-like receptors (TLR) 2 and 4 in rats bred under microbiologically controlled conditions (barrier), under standard conditions (conventional), and in barrier animals adapted to standard conditions (barrier/conventional). Cecal microbiota was analyzed by plate culture, and fluorescence in situ hybridization and microbial profiles were assessed by terminal restriction fragment length polymorphism. Cecal expression of TLR-2 and TLR-4 was determined by reverse transcription polymerase chain reaction (PCR). Total number of cecal bacteria was similar in the three groups. However, the barrier group showed a higher number of strict anaerobic bacteria (Bacteroides spp. and Clostridium spp.) while Bifidobacterium spp. were scarce. Re-housing the barrier-bred rats into conventional conditions led to a microbiota with intermediate characteristics between the barrier and conventional groups. Richness of the cecal microbial ecosystem was similar in the three groups, although a relative time-dependent variation, with highest homogeneity in the barrier group, was observed. Expression levels of TLR-2 and TLR-4 had no clear correlation with the microbiota. These results show that the relative composition of the cecal microbiota in rats varies spontaneously with changes in the environmental conditions, with minor impact in the expression of TLR-2 and TLR-4. These observations might be important in the understanding of variability in animal responses, particularly to immune-related stimuli, when assessed in the context of the environmental/microbiological conditions.

Introduction

The intestinal tract harbors a complex bacterial ecosystem, which has not yet been fully characterized [4]. It contains numerous species, most of them anaerobes, and the concentration can be as great as 1 × 1011 bacteria/g of feces [25]. The presence of these commensal bacteria appears to be of crucial importance in the pathogenesis of several gastrointestinal alterations, such as inflammatory bowel disease (IBD) or irritable bowel syndrome (IBS) [8, 37, 47]. In general, lactic acid bacteria are regarded as beneficial members modulating gut immune responses; on the other hand Bacteroides, Clostridium, and Enterobacteriaceae have been associated with the exacerbation of intestinal inflammatory responses in humans and experimental models [25, 30, 34, 37].

Gut microbiota represents a dynamic system in continuous interaction with the host. Mammalian toll-like receptors (TLRs) play a central role in the recognition of gut microbiota and the initiation of immune responses to microbial pathogens. Among TLRs, TLR-2 and TLR-4 are considered key players in the recognition of gut microbiota. TLR-2 recognizes a variety of microbial components from Gram-positive bacteria, whereas TLR-4 recognizes LPS (lipopolysaccharide) of Gram-negative bacteria [6, 40]. It is not fully understood how TLRs distinguish between commensal and pathogen bacteria [29]. However, an increase of TLRs expression has been observed after exposure to commensal bacteria, a crucial response in the maintenance of gut homeostasis and in the prevention of gastrointestinal pathophysiological alterations [29].

Highly hygienic conditions induce changes in the composition of commensal intestinal microbiota, leading to the appearance of abnormal immune responses, including an altered recognition of the commensal microbiota, which seems to characterize IBD [15, 26]. For instance, an increase of IBD incidence in industrialized countries has been related to the improvement in standards of hygiene and the consequent alterations of the microbiota. This agrees with the presence of changes in TLR-2 and TLR-4 expression, as described both in IBD patients and rodent models of gut inflammation [12, 36, 38]. Although systematic studies have not been performed, this suggests that spontaneous variations in gut microbiota might influence the susceptibility to and course of intestinal inflammation [17]. In the present study we characterized, for the first time, the indigenous intestinal (cecal) microbiota in rats born and bred under barrier conditions, with a highly controlled environment and a well-characterized inoculated microbiota at the time of formation of the colony, and in rats with the same genetic background but born and bred in standard (conventional) conditions. Furthermore, we also characterized spontaneous changes of gut microbiota present in rats born and bred under barrier conditions but adapted for a period of 3 weeks, during adulthood, to conventional conditions. Finally, we assessed changes in TLR-2 and TLR-4 expression, as two of the main components of the microbiota recognition system, in the same animals, aiming to establish a correlation between TLR expression and gut microbiota.

Methods

Animals

Twelve 6-week-old and 12 9-week-old male OFA Sprague–Dawley rats bred and maintained in a barrier protected area, with all materials, water, food, and bedding sterilized before entering the barrier, were provided by Charles River Laboratories (Lyon, France). The original microbiota used to associate the barrier colony is summarized in Table 1, as provided by the breeder. Transport of barrier-bred animals was made in filtered boxes to guarantee maintenance of their microbiological status.
Table 1

Original microbiota (starter culture) used to associate the OFA Sprague–Dawley barrier rat colony

Original implantation bacteria

Bacteroides distasonis

Lactobacillus acidophilus

Lactobacillus salivaris

Schaedler fusiform-shaped bacterium

3 strains of CRL fusiform-shaped bacterium

CRL mouse spirochete

Escherichia coli (non-hemolytic)

Streptococcus faecalis (group D, Enterococcus spp.)

Data obtained from Charles River Laboratories, France

In addition, 16 9-week-old male OFA Sprague–Dawley rats bred in conventional conditions in the Animal Facility of the Universitat Autònoma de Barcelona were used. This conventional colony was established in 1994 from OFA Sprague–Dawley rats from Charles River Laboratories and has been appropriately cross-bred in order to maintain genetic stability. When in conventional conditions, water, food, and bedding were given to animals as facilitated by the commercial provider, without any further treatment. All animals maintained at the university’s animal facility were housed in standard plastic cages with stainless steel grid roofs in an environmentally controlled room (20–21°C, 40–70% humidity, 12 h light/dark cycle), and received a commercial pellet diet (15.4%, protein; 2.9% fat and 3.9% fiber; SASE, Panlab S.L., Barcelona, Spain) and tap water ad libitum. Cages and wood shavings used for bedding (Ultrasorb, Panlab S.L., Barcelona, Spain) were changed once a week. All animals were 9 weeks old at the time of testing.

All procedures were approved by the Ethical Committee of the Universitat Autònoma de Barcelona.

Experimental Groups and Time Course of the Studies

The following experimental groups were defined (n = 12–16 for each): (1) 9-week-old barrier-bred rats tested immediately on arrival (barrier group), (2) 9-week-old rats bred under conventional conditions (conventional group), and (3) 6-week-old barrier rats maintained under conventional conditions at the animal facility of the Universitat Autònoma de Barcelona for 3 weeks (barrier/conventional group). All procedures were performed between June 2005 and October 2008. In each experiment, at least two of the three experimental groups were included, and all experimental groups were repeated at least three times during the course of the studies. For the sake of clarity, and taking into account the time course of the studies, for some of the results presented, the experimental groups have been regrouped as two main subgroups, 2005–2006 and 2007-2008.

Sample Collection

Rats were euthanized by CO2 inhalation. Immediately, a medial laparotomy was performed, the cecum was localized and exteriorized, and samples of cecal content (about 0.5 g) and cecal tissue were obtained under sterile conditions and immediately frozen with liquid nitrogen. All samples were stored at −20°C (cecal content) or −80°C (tissue) until analysis.

Enumeration of Cultivable Bacteria Using Plate Culture

Immediately upon collection, samples of cecal content were placed in preweighed sterile tubes filled with CO2 gas. A suitable volume of fluid thyoglicolate medium (FTM) to obtain 10−1 dilution was added, and the tubes were vortexed for 1 min to homogenize the sample. Thereafter, serial dilutions (from 10−3 to 10−7) of the homogenates were prepared, and 0.1 ml of each dilution was spread on the surface of plates containing the following agar media: MRS agar (Oxoid, Madrid, Spain) for Lactobacillus spp. [25], Schaedler Neomicine/Vancomicine plus 5% sheep blood medium (SNV + 5% sheep blood medium) (Biomeriux, Madrid, Spain) for the Gram-negative anaerobes Bacteroides and Prevotella [25], membrane Clostridium perfringens (mCP) agar (Oxoid, Madrid, Spain) for Clostridium spp. [46], and McConkey agar (Biomerieux, Madrid, Spain) for coliforms [28]. MRS and SNV + 5% sheep blood agar plates were incubated at 37°C in an anaerobic bag (Biomerieux, Madrid, Spain). Plates for Clostridia were incubated at 44 ± 1°C in an anaerobic bag, and MacConkey agar plates were incubated aerobically at 37°C. All cultures were incubated for 48 h. Plate cultures were performed only for the experimental groups tested within the 2005–2006 subgroup.

After incubation, representative colonies of each selective medium were identified to genus level by standard bacteriological methods. Representative colonies developed on SNV + 5% sheep blood agar and mCP agar were subcultured on blood agar plates and incubated aerobically at 37°C to assure they were obligate anaerobes. In all cases, counts of colonies were expressed as colony-forming units (CFU) per gram of intestinal content.

Enumeration of Bacteria Using Fluorescence In Situ Hybridization

For fluorescence in situ hybridization (FISH), oligonucleotide probes consisted of a single strain DNA covalently linked with Cy3 at the 5′-end. Probes used were: EUB 338 (5′GCTGCCTCCCGTAGGAGT3′) to total bacteria [1]; NON 338 (5′ACATCCTACGGGAGGC3′) to non-bacteria (negative control) [1]; BAC 303 (5′CAATGTGGGGGACCTT3′) to Bacteroides spp. [33]; EREC 482 (5′GCTTCTTAGTCAGGTACCG3′) to Clostridium cluster XIVa [33]; LAB 158 (5′GGTATTAGCACCTGTTTCCA3′) to Lactobacillus spp. and Enterococcus spp. [33]; ENT-D (5′TGCTCTCGCGAGGTCGCTTCTCTT3′) to Enterobacteria [27]; and BIF 164 (5′CATCCGGCATTACCACCC3′) to Bifidobacterium spp. [42]. All probes were obtained from Tib MolBiol (Mannheim, Germany).

Frozen cecal contents were thawed, and 4.5 ml of Millipore filtered PBS and three to five glassbeads (diameter 3 mm) were added, and the mixture was homogenized on a vortex for 3 min. The suspension obtained was then centrifuged for 1 min at 700×g. From the supernatant, a 1-ml aliquot was collected and fixed in 3 ml of freshly prepared 4% paraformaldehyde in PBS. After overnight fixing at 4°C, samples were divided in small aliquots of 0.4 ml and stored at −20°C until use.

At the time of analysis, aliquots of fixed samples were diluted in PBS until appropriate concentrations of cells, according to preliminary experiments: 1:1,600 for the EUB338 probe, 1:160 or 1:400 for the BAC303 probe, 1:160 or 1:400 for the EREC482 probe, 1:80 or 1:160 for the LAB158 probe, 1:40 or 1:80 for the ENT-D probe, and 1:40 or 1:80 for the BIF-164 probe. Ten-well (8 mm diameter) gelatine-covered slides were used. Subsequently, 5 or 10 μl of the proper dilution of sample was placed in each well, air-dried at room temperature, and fixed for 10 min with ethanol. For hybridization, probe solutions were further diluted in 50°C preheated hybridization buffer (20 mM Tris–HCl, 0.9 M NaCl, 0.1% sodium dodecyl sulfate pH 7.2), with 20% formamide for the LAB 158 and NON 338 probes, to a final concentration of 5 ng/μl. The hybridization mixture was added to the slides for an overnight incubation in a dark moist chamber (except for the BAC 303 probe, which was hybridized for only 3 h, and for the BAC 303 and LAB 158 probes, which were incubated overnight at 47°C). Samples to be hybridized with the LAB 158 probe were pre-treated with lysozime for 1 h at 37°C prior to the hybridization process. After incubation, slides were rinsed in preheated washing buffer (20 mM Tris–HCl, 0.9 M NaCl, pH 7.2, 180 mM NaCl for the LAB 158 and NON 338 probe) for 30 min at 50°C. Thereafter, the slides were briefly rinsed with milli-Q water, air-dried, and mounted with Vectashield (Vector Laboratories, Peterborough, UK).

Slides were viewed under oil immersion, using a Nikon Fi 60 epifluorescence microscope equipped with a filter for Cy3. Twenty-five randomly selected fields were counted for each sample (in duplicate).

Terminal Restriction Fragment Length Polymorphism: Procedure and Data Analysis

Terminal restriction fragment length polymorphism (t-RFLP) analysis of bacterial community was performed following the procedures described by Hojberg et al. [18]. Briefly, a 1,497-pb fragment of the 16S rDNA gene was amplified using a 6-carboxy-fluorescein-labeled forward and reverse primers (S-D-Bact-0008-a-S-20: 5′-6-FAM-AGAGTTTGATCMTGGCTCAG-3′; and PH1552: 5′AAGGAGGTGATCCAGCCGCA-3′, respectively) against the first 20 bases of the 16S RNA sequence. Duplicate PCR were performed for each sample. Fluorescent-labeled PCR products were purified on QIAquick PCR purification kit columns (Qiagen, West Sussex, UK) and eluted in a final volume of 30 μl of Milli-Q water. Then, the resultant PCR product was subjected to a restriction with HhaI (20,000 U/μl) (Biolabs Inc., New England, USA). Fluorescent-labeled terminal restriction fragments (TRF) were analyzed by capillary electrophoresis on an automatic sequence analyzer (ABI 3100 Genetic Analyzer, PE Biosystems, Warrington, UK) in Gene-Scan mode with a 25-U detection threshold. Determination of the TRF sizes in the range 50–700 bp were performed with the size standard GS-1000-ROX (PE Biosystems).

Data obtained consisted of size (base pairs) and peak area for each TRF. To standardize the DNA loaded on the capillary, the sum of all TRF peak areas in the pattern was used to normalize the peak detection threshold in each sample. Following the method described by Kitts [21], a new threshold value was obtained by multiplying a pattern’s relative DNA ratio (the ratio of total peak area in the pattern to the total area in the sample with the smallest total peak area) by 323 area units (the area of the smallest peak at the 25 detection threshold in the sample with the smallest total peak area). For each sample, peaks with a lower area were deleted from the data set. Thereafter, a new total area was obtained by the sum of all the remaining peak areas in each pattern.

Biodiversity (also known as richness) was considered as the number of peaks in each sample after standardization. For pair-wise comparisons of the profiles, a Dice coefficient was calculated, and dendograms were constructed using the Fingerprinting II software (Informatix, Bio-Rad, CA, USA) and an unweighted pair-group method with averaging algorithm. To deduce the potential bacterial composition of the samples, in silico restrictions for the major rat gut bacteria with the primers and the enzyme used were obtained by using the analysis function TAP-tRFLP from the Ribosomal Database Project II software. Results are presented as potential compatible bacterial species. Note also that direct attribution of species to individual peaks is not unequivocally possible unless fingerprinting is complemented with sequence analysis of clone libraries. Analysis of electropherograms was used for the visual comparison of compatible TRF with different bacteria for the three experimental groups.

Semi-quantitative Reverse Transcription Polymerase Chain Reaction

Total RNA was extracted from cecal tissue samples using RNAwiz (Ambion, Madison, WI, USA) and treated with DNA-free (Ambion) for 30 min at 37°C. cDNA was synthesized from 5 μg total RNA in a reaction mixture of 50 μl containing 0.5 μg of oligo 18 (dT) primer (Ambion), 2 nM dNTP (Ecogen, Barcelona, Spain), and 10 units of Moloney Murine Leukemia Virus (MMLV) reverse transcriptase (Ambion). The resulting cDNA was amplified in a total volume of 50 μl with 1 unit of taqDNA, 1 mM dNTP, and 0.5 μM of primers. Primers used were (1) TLR2: left: 5′-CTGACCTCTCTCAACGAACT-3′; right: 5′-CGCTGAGGTCTAAGAACTCT-3′ (NCBL reference sequence NM-198769); (2) TLR4: left: 5′-CATAGCAGATGTTCCTAGGC-3′; right 5′-GGAGTCTGTAGAGTGTGTCA-3′ (NCBL reference sequence NM-19178); and (3) GAPDH: left: 5′-ATGAGCCCTTCCACGATGCC-3′; right: 5′-CCGCCCCTTCCGCTGATGCC-3′ (NCBL reference sequence NM-017008) all from (Proligo-Sigma, Madrid, Spain). The PCR amplification protocol was as follows: 35 (GAPDH) or 40 (TLR-2 and TLR-4) cycles with 1 min (GAPDH) or 30 s (TLR-2 and TLR-4) of denaturation at 95°C, 1 min (GAPDH) or 30 s (TLR-2 and TLR-4) of annealing at 50°C, and 1 min of extension at 72°C on a thermal cycler (Techno Cambridge Ltd.). Amplified products were electrophoresed on 2% agarose gel in TAE buffer, stained with ethidium bromide, photographed under ultraviolet light, and quantified using image-analyzing software (Quantity-One, Bio-Rad laboratories). For semiquantification, the ratio of the optical density of each PCR product and GAPDH was determined.

Statistical Analysis

Data are expressed as mean ± SEM. Statistical analysis was performed using a one-way analysis of variance (one-way ANOVA), followed when necessary, by a Student–Newman–Keuls multiple comparisons test. Results were considered statistically significant when P < 0.05.

Results

Characterization of Intestinal (Cecal) Microbiota by FISH

Total bacteria, determined by FISH as EUB338-positive cells, between 1 × 1010 and 8 × 1010 cell/ml and was similar in all experimental groups, regardless of the time of testing (Table 2). In the barrier group, overall characteristics of the microbiota were in agreement with the microbiological composition of the inoculus used originally to colonize the intestine of these animals (Fig. 1 and Tables 1 and 2).
Table 2

Bacterial counts in the different experimental groups, as determined by FISH

 

Barrier group (×108 cells/ml)

Barrier/conventional group (×108 cells/ml)

Conventional group (×108 cells/ml)

Total bacteria

352 ± 191

421 ± 221

423 ± 179

Enterobacteriaceae

0.1 ± 0.2

0.6 ± 1.7

0.07 ± 0.09

Bacteroides spp.

25.1 ± 7.5

16.6 ± 8.7a

13.7 ± 5.7b

Clostridium coccoides–Eubacterium rectale (Clostridium cluster XIVa)

82.3 ± 38.6

69.6 ± 43.4

44.5 ± 27.6a

Lactobacillus and Enterococcus spp.

1.2 ± 0.8

4.8 ± 4.6

5.4 ± 7.7

Bifidobacterium spp.

0.01 ± 0.02

1.5 ± 2.5

3.1 ± 4.7a

Data are mean ± SEM, n = 12–16 per group (see “Methods” for details of the counting process)

aP < 0.05 vs. barrier group (ANOVA)

bP < 0.05 vs. barrier/conventional group (ANOVA)

https://static-content.springer.com/image/art%3A10.1007%2Fs00248-010-9737-z/MediaObjects/248_2010_9737_Fig1_HTML.gif
Figure 1

Relative distribution of cecal microbiota, as quantified by FISH, in the different experimental groups. Data represent the percent composition of the main bacterial groups present in the gut microbiota (Bacteroides spp., Bifidobacterium spp., Clostridium, Enterobacteria, Lactobacillus spp., and Enterococcus spp.) quantified using FISH techniques and represent mean values from 12–16 animals per group. Percent composition of the microbiota was calculated taking as 100% the bacterial counts obtained by FISH with the EUB 338 probe. For exact cell counts, see Table 2. BAC, Bacteroides spp.; ENT, Enterobacteria; CLO, Clostridium cluster XIVa; LAB, Lactobacillus spp. and Enterococcus spp.; BIF, Bifidobacterium spp.

Among Gram-negative bacteria, the counts of Enterobacteriaceae (ENT-D probe) were scarce (in most cases less than 1% of the microbiota quantified) and of similar magnitude in all experimental groups (Fig. 1 and Table 2). In contrast, Bacteroides spp. (BAC 303 probe) were relatively abundant (by 19% to 26% of the microbiota quantified) and slightly, but significantly higher in the barrier group when compared with the conventional and the barrier/conventional groups (P < 0.05; Fig. 1 and Table 2). These inter-groups variations were observed regardless the time of testing.

Gram-positive bacteria represented 80% of the FISH-quantified microbiota. In particular, Clostridium coccoides–Eubacterium rectale group (Clostridium cluster XIVa, EREC 482 probe) accounted for the largest bacterial population in all experimental groups (Fig. 1), with higher counts in the barrier group when compared with the conventional group (P < 0.05), and the barrier/conventional group showing an intermediate situation (Table 2). Counts for Lactobacillus spp. and Enterococcus spp. (LAB 158 probe) were similar in the conventional and barrier/conventional groups with lower counts in the barrier group, although statistical significance was not achieved (Table 2). As it relates to Bifidobacterium spp. (BIF164 probe), counts were high in cecal samples from the conventional group and scarce in the barrier group (P < 0.05), while the barrier/conventional group showed a transition towards the conventional conditions (Fig. 1, Table 2).

These results were largely confirmed using classical plate culture techniques, although this was performed only for the experimental subgroup 2005–2006. Results obtained showed that the number of CFU/g of cecal content for Bacteroides spp. was significantly higher in the barrier group when compared with the conventional and the barrier/conventional groups (Table 3). Similarly, Enterobacteriaceae had a trend towards higher counts in the barrier group (Table 3), while Lactobacillus spp. was relatively reduced in the barrier group, although statistical significance was not reached (Table 3). Interestingly, and conversely to that observed with FISH, Clostridium spp. population was higher in the barrier/conventional than in the barrier or the conventional group (Table 3). These observations served as a validation of the FISH techniques, and therefore, plate cultures were not further pursued.
Table 3

Bacterial counts in the different experimental groups, as determined by culture

 

Barrier group (×108 CFU/g of cecal content)

Barrier/conventional group (×108 CFU/g of cecal content)

Conventional group (×108 CFU/g of cecal content)

Enterobacteriaceae

0.1 ± 0.06

0.05 ± 0.03

0.03 ± 0.01

Bacteroides spp.

3.1 ± 1.2a

3.4 ± 0.4a

0.8 ± 0.1

Clostridium spp.

2.5 ± 1.1b

20.0 ± 4.5

8 ± 1.7b

Lactobacillus spp.

11.0 ± 4.8

33.0 ± 9.0c

22.0 ± 3.3c

Data are mean ± SEM, n = 4–9 per group. Plate culture was only performed for experimental groups sampled between 2005 and 2006, as confirmed by the FISH technique

aP < 0.05 vs. conventional group

bP < 0.05 vs. barrier/conventional group

cP = 0.08 vs. barrier group

Ecological Characterization of the Intestinal Microbiota: t-RFLP Analysis

The similarity indexes of the t-RFLP profiles, illustrated in the form of a dendogram, of the cecal microbiota in the different experimental groups are shown in Fig. 2a. Overall, and despite the time of testing, the dendogram obtained shows a relatively high homogeneity in the microbiota of the barrier group when compared with the conventional or the barrier/conventional groups, which showed less homogeneity. Nevertheless, the overall biodiversity of the microbiota was similar in the three experimental groups, with an average of t-RFs (taken as a measure of biodiversity) that varied from 22 to 45 among the different experimental groups, and a tendency for higher diversity in the conventional group (P = 0.054, ANOVA; Fig. 2b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-010-9737-z/MediaObjects/248_2010_9737_Fig2_HTML.gif
Figure 2

a Dendogram illustrating the clustering of the different experimental groups according to the t-RFLP banding patterns obtained from the analysis of the cecum microbiota. Each line represents an animal, and the code to the right identifies the experimental group and the time at which the samples were obtained (Conv conventional, Barr/Conv barrier/conventional, 05–06 samples obtained between 2005 and 2006, 07–08 samples obtained between 2007 and 2008). Notice how, although with some variability over time, the animals of the different experimental groups have a tendency to cluster together. The dendogram distances represent percentage of similarity. bBar graph showing the overall biodiversity among the different experimental groups, as determined by t-RFLP (see “Methods” for details)

Table 4 summarizes the main bacterial groups, as detected by t-RFLP analysis, with differential presence in the three experimental groups. In general, the t-RFLP analysis was concordant with the FISH data, at least for those bacterial groups identified simultaneously by FISH and by t-RFLP analysis. For instance, the general prevalence of Clostridium spp. was higher in the barrier group vs. the conventional and barrier/conventional groups (Fig. 1 and Table 3). Similarly, the prevalence of Bacteroides fragilis detected by t-RFLP was also higher in the barrier group vs. the conventional and the barrier/conventional groups, as detected by FISH (Fig. 1 and Table 3). Interestingly, in many cases, the barrier/conventional group showed an intermediate situation between the barrier and the conventional groups with respect to bacterial prevalence. In general, non-identified bacterial groups (classified as “unidentified bacterium” and “uncultured bacterium” in the t-RFLP analysis) showed a higher prevalence in the barrier group than in the conventional or barrier/conventional groups.
Table 4

Theoretical restriction 5′-fragment (tRF) size predicted for the major rat gut bacteria and prevalence in the different experimental groups

Compatible bacterial group

tRF size

Frequencya

Barrier (n = 10)b

Barrier/conventional (n = 12)

Conventional (n = 16)

Unidentified

54–55

1 (10)

10 (83)

12 (75)

Bacillus spp./Lactococcus lactis ssp.

61–62

0 (0)

7 (58)

10 (62)

Rhodoplanes spp.

65

4 (40)

6 (50)

3 (19)

Uncultured rumen bacterium/Leptotrichia spp.

71

5 (50)

2 (17)

3 (19)

Photorhabdus sp.

74–75

6 (60)

6 (50)

0 (0)

Uncultured bacterium

77–78

8 (80)

6 (50)

9 (56)

Uncultured bacterium

79

0 (0)

6 (50)

7 (44)

Erythrobacter spp./uncultured bacterium

82–83

3 (30)

1 (8)

9 (56)

Desulfovibrio defluvii/Roseiflexus spp.

86–87

4 (40)

6 (50)

0 (0)

Flavobacterium psychrophilum

88–89

7 (70)

0 (0)

2 (12)

Flavobacteriaceae bacterium/uncultured rumen bacterium/Desulfovibrio spp.

93–94

7 (70)

4 (33)

0 (0)

Desulfovibrio profundus/uncultured bacterium

95

6 (60)

7 (58)

3 (19)

Uncultured bacterium

96

8 (80)

6 (50)

6 (38)

Desulfococcus oleovorans/Desulfomonile limimaris/Helicobacter pylori

97–98

7 (70)

11 (92)

9 (56)

Helicobacter pylori/uncultured rumen bacterium

99

2 (20)

8 (67)

8 (50)

Bacteroides spp./uncultured rumen bacterium

100

1 (10)

6 (50)

1 (6)

Bacteroides fragilis/uncultured rumen bacterium/Prevotella ruminicola

101–102

9 (90)

9 (75)

0 (0)

Uncultured rumen bacterium

103–104

6 (60)

7 (58)

5 (31)

Desulfitobacterium hafniense

107–108

9 (90)

7 (58)

0 (0)

Thiobacillus spp.

110–111

8 (80)

4 (33)

0 (0)

Unidentified

113

6 (60)

7 (58)

2 (12)

Unidentified

115

7 (70)

4 (33)

1 (6)

Unidentified

117

6 (60)

2 (17)

3 (19)

Unidentified

123–124

3 (30)

5 (42)

0 (0)

Streptomyces rimosus subsp. rimosus

125

3 (30)

3 (25)

0 (0)

Uncultured rumen bacterium

126

6 (60)

0 (0)

0 (0)

Unidentified

127–129

7 (70)

0 (0)

4 (25)

Uncultured rumen bacterium

134–135

4 (40)

6 (50)

2 (12)

Unidentified

136

4 (40)

0 (0)

2 (12)

Uncultured rumen bacterium

137

1 (10)

4 (33)

0 (0)

Unidentified

138

3 (30)

0 (0)

0 (0)

Microbacterium spp.

144–145

0 (0)

7 (58)

5 (31)

Unidentified

148–149

5 (50)

6 (50)

0 (0)

Unidentified

156

4 (40)

6 (50)

1 (6)

Unidentified

165–167

10 (100)

9 (75)

0 (0)

Unidentified

171

3 (30)

0 (0)

0 (0)

Micrococcus sp./Acetobacter pasteurianus

174–175

1 (10)

11 (92)

3 (19)

Uncultured rumen bacterium

181–182

1 (10)

10 (83)

1 (6)

Uncultured rumen bacterium

184–185

7 (70)

8 (67)

3 (19)

Uncultured rumen bacterium/Clostridium spp./Butyrivibrio fibrisolvens

189–190

9 (90)

4 (33)

2 (12)

Psychrobacter spp./uncultured bacterium/Francisella spp.

194–195

4 (40)

5 (42)

1 (6)

Clostridium spp.

196–197

3 (30)

8 (67)

7 (44)

Unidentified

199

6 (60)

0 (0)

0 (0)

Clostridium rectum/uncultured bacterium/Mycobacterium spp.

201

6 (60)

2 (17)

2 (12)

Uncultured rumen bacterium

203–204

4 (40)

6 (50)

0 (0)

Uncultured rumen bacterium

205–206

4 (40)

1 (8)

1 (6)

Clostridium spp.

231–232

0 (0)

6 (50)

4 (25)

Clostridium perfringens

234–235

1 (10)

5 (42)

3 (19)

Bacillus

239–240

1 (10)

4 (33)

1 (6)

Bacillus subtilis subsp. subtilis/Bacillus licheniformis/Bacillus spp.

241–242

4 (40)

0 (0)

1 (6)

Uncultured bacterium

251–252

4 (40)

1 (8)

1 (6)

Simkania navegensis

271–272

1 (10)

4 (33)

5 (31)

Unidentified

391–392

4 (40)

0 (0)

2 (12)

Unidentified

402–403

5 (50)

0 (0)

0 (0)

Unidentified

422

3 (30)

0 (0)

0 (0)

Unidentified

488

6 (60)

0 (0)

0 (0)

Uncultured bacterium

500

4 (40)

0 (0)

0 (0)

aData represent the number of animals within each group presenting the bacterial group predicted by the corresponding tRF size and the incidence, in percentage (between brackets)

bBecause of technical problems, only samples from 10 out of the 12 animals included in this group were analyzed

Expression of TLR-2 and TLR-4

TLR-2 and TLR-4 transcripts were clearly identifiable, with a variable relative intensity, in all tissue samples analyzed. A relatively high intra-group variability was observed in the expression of both receptors. Despite this, animals bred and maintained in barrier conditions had a clear tendency to over-express TLR-4 when compared with the conventional group (P = 0.057), while the barrier/conventional group showed intermediate levels of expression (Fig. 3b). Except for five out of 12 animals in the barrier/conventional group and four out of 16 animals in the conventional group, expression levels of TLR-4 were very low in these groups. No differences among groups were observed for TLR-2 expression; in most cases, expression levels were also low, particularly in the conventional group (Fig. 3a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-010-9737-z/MediaObjects/248_2010_9737_Fig3_HTML.gif
Figure 3

Relative expression of TLR-2 (a) and TLR-4 (b) in cecal tissues of the three experimental groups. Each open symbol represents an individual animal, and the black symbol to the right corresponds to the mean ± SEM of the group. +P = 0.057 vs. barrier group (ANOVA)

Discussion

The present results show that the relative composition of the cecal microbiota in rats varies spontaneously with changes in the environmental conditions (highly controlled environment, barrier, vs. standard conditions, conventional), although the total bacterial count seems to be unaffected. Furthermore, changes in microbiota seem to be characteristic of the environmental conditions considered with only relatively minor variations over time, at least for the main groups of bacteria analyzed here. However, these changes in bacterial microbiota seem to have a rather minor impact in the bacterial recognition systems present within the gut, at least as it relates to the expression of TLR-2 and TLR-4.

Although several authors have attempted to define the gut ecosystem and the number of bacteria present in the large intestine [4, 9], the difficult analysis of intestinal microbiota is far from being completed. Fortunately, the recent application of molecular techniques to this field, such as the use of FISH with oligonucleotide probes or t-RFLP analysis, has proven to be very useful [2, 9, 10, 42]. In our study, the results obtained with the EUB 338 probe, designed to visualize total bacteria, give cecal cell counts similar to those previously reported [9]. Interestingly, the total cell count was similar in animals maintained in highly controlled microbiological conditions (barrier group), animals kept in standard conditions (conventional group), or barrier animals adapted to standard conditions (barrier/conventional group). This might suggest that, in otherwise normal conditions, a highly controlled hygienic environment does not imply a lower number of bacteria colonizing the gut, but a selection in the bacterial types forming the commensal microbiota. Moreover, this might suggest that intestinal microbiota develops quantitatively up to a number that could be regarded as normal, at this point changes in microbiota might be generated through qualitative (characteristic and/or proportions of the bacterial groups) rather than quantitative variations. Overall, maintenance in barrier conditions seems to favor the settlement of strict anaerobic bacteria, particularly Bacteroides spp. and clostridia. However, less restrictive hygienic conditions, such as a conventional housing, favor an increase of lactic acid bacteria. This switch in bacterial microbiota was particularly evident for Bifidobacterium, very scarce in animals bred, and maintained in barrier conditions while relatively abundant in conventional conditions (by 5% of the microbiota quantified by FISH).

With these general characteristics, gut microbiota should be regarded as a relatively dynamic system that varies over time, space, and environment [5]. In our studies, apart from the environment-related changes described above, time-related changes were also evidenced when a relatively large proportion of cecal microbiota was analyzed by t-RFLP. As observed in the dendogram derived from the t-RFLP analysis (see Fig. 2), the animals included in the study tended to cluster according to both environment and time of testing, thus indicating the existence of time-related variations in the microbiota associated to the same environment. Nevertheless, for the bacterial groups characterized by FISH, a clear environment-related constancy was observed over time.

The present work is limited to the characterization of Enterobacteria, Lactobacillus spp., Bacteroides spp., Clostridium spp., and Bifidobacterium spp. In agreement with previous reports [9], and regardless the experimental group considered, Clostridium spp. were dominant, followed by Bacteroides spp., while Enterobacteria, Lactobacillus spp., and Bifidobacterium spp. were less numerous. This was especially evident for the barrier and the conventional groups. Interestingly, barrier-bred animals housed in conventional conditions for 3 weeks showed, in general, an intermediate state in their microbiota, thus suggesting a spontaneous adaptation towards the microbiological characteristics of the new environment. Overall, the FISH-quantified bacteria represented 12% to 18% of the total cecal bacteria, as determined by FISH when using the probe EUB 338, a result in accordance with that previously reported in rats [9] as well as in humans [23]. Although this might appear to be a relatively small percentage of the total cecal microbiota, these bacterial groups might have important physiological and pathophysiological significance within the gut. They have been identified as key components in determining the intestinal immune response [13, 47], have been implicated in different gastrointestinal physiopathological alterations (such as IBD or IBS), and/or have a potential interest as pre/probiotics [7, 8, 14, 22, 24, 32, 35, 39, 47]. Therefore, the present observations might have implications at an experimental level, indicating that the gut microbiota might be an important factor to be taken into account when performing physiological, pathophysiological, and/or pharmacological studies within the gastrointestinal system. This agrees with a growing body of evidence implicating the commensal microbiota as an experimental factor influencing the outcome of both digestive and extra-digestive studies in animals [8, 31, 41, 43, 44]. For instance, in our conditions, although not thoroughly assessed, none of the animals showed any indication of gut inflammation at the time of euthanasia. Nevertheless, animals bred and maintained in barrier conditions, with a controlled gut microbiota, had a predominant commensal microbiota that has been associated with IBD; therefore, one can speculate that these animals, although not showing spontaneous inflammation, might be more susceptible to intestinal inflammation than the other groups. This agrees with data showing that susceptibility to intestinal inflammation increases when gut microbiota has a limited development or is absent [20, 43, 45].

Overall, our observations might have a relationship with the “hygiene hypothesis” [15, 26]. As previously mentioned, qualitative characteristics of the cecal microbiota were different in animals bred and maintained under barrier conditions, and therefore with high hygienic standards, and in animals maintained under normal hygienic conditions (conventional group). These differences are also emphasized by the fact that barrier animals that adapted to conventional conditions for 3 weeks (barrier/conventional group) show a switch in their intestinal microbiota towards the characteristics of the conventional group. As also mentioned, this can have important implications in the responses to immunological stimuli, for instance gut inflammatory responses, as initially suggested by the hygiene hypothesis. It is important to mention that other housing/environmental factors not directly assessed here (such as food and water supply) as well as any genetic drift experienced over generations by the conventional colony might contribute to the observed between-group differences in gut commensal microbiota.

Immune responses to intestinal bacteria are mediated, at least partially, by the interaction of bacterial wall components with TLRs [6, 40]. Among the TLRs, TLR-2 and TLR-4 have been reported as the main partners interacting with the gut microbiota [38, 40]. Most of the studies performed so far investigating the interplay between TLRs and bacteria have been done either in vitro by exposing cells to bacterial wall components or, if in vivo, by exposing germ-free animals to specific bacterial species or to bacterial components [19, 40]. According to these studies, Gram-positive bacteria are recognized by TLR-2, while TLR-4 recognizes mainly Gram-negative bacteria. Nevertheless, both receptors can interact with multiple microorganisms and are likely to be subjected to complex, multifactorial, regulatory mechanisms [3, 11, 16]. Results obtained suggest a relative heterogeneity in the intra-group expression of TLR-2 and TLR-4, without clear distinctive patterns of expression between experimental groups. Nevertheless, animals maintained in conventional conditions showed a clear tendency to have lower expression levels of both TLR-2 and TLR-4 than, particularly, animals maintained in barrier conditions. This correlates with the higher number of Gram-negative bacteria, particularly Bacteroides spp., detected in the barrier group. However, Gram-positive and Gram-negative non-identified bacterial groups (classified as unidentified bacterium and uncultured bacterium in the t-RFLP analysis) might have a significant influence in the expression levels of both TLR-2 and TLR-4, respectively.

In summary, the present results show that the relative composition of the cecal microbiota in rats varies spontaneously with changes in the environmental conditions. These observations might have an impact in the understanding of variability in animal responses, particularly to immune-related stimuli, when assessed in the context of the environmental/microbiological conditions and support the so-called hygiene hypothesis. The present data warrant further studies assessing spontaneous variations in other components of the intestinal microbiota and other bacterial recognition systems as well as specific studies designed to assess environment/microbiological conditions-related differences in gut immune response, particularly as it relates to intestinal inflammation, of interest to understand the role of intestinal microbiota in gut pathophysiology.

Acknowledgments

Our gratitude to Charles River Laboratories, France, for providing information on the microbiological status of CRL animals. We would also like to thank Blanca Chavez, Jenifer Toskano, and Rosa Torres for technical assistance with molecular techniques and Andrew Hudson for editorial revision of the manuscript. This study was financed with grants 2005SGR-00255 and 2009SGR708 from the Generalitat de Catalunya, PRP2005-16 from Universitat Autonoma de Barcelona, BFU2007-62794 from Ministerio de Educación y Ciencia, and BFU2009-08229 from Ministerio de Ciencia e Innovación.

Copyright information

© Springer Science+Business Media, LLC 2010