Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Nutrition

Coprophagous behavior of rabbit pups affects implantation of cecal microbiota and health status1

 

This article in JAS

  1. Vol. 92 No. 2, p. 652-665
     
    Received: Feb 23, 2013
    Accepted: Dec 13, 2013
    Published: November 24, 2014


    2 Corresponding author(s): Sylvie.Combes@toulouse.inra.fr
 View
 Download
 Share

doi:10.2527/jas.2013-6394
  1. S. Combes 2,
  2. T. Gidenne*†‡,
  3. L. Cauquil*†‡,
  4. O. Bouchez§# and
  5. L. Fortun-Lamothe*†‡
  1. INRA, UMR1289 Tissus Animaux Nutrition Digestion Ecosystème et Métabolisme, F-31326 Castanet-Tolosan, France
    Université de Toulouse INPT ENSAT, UMR1289 Tissus Animaux Nutrition Digestion Ecosystème et Métabolisme, F-31326 Castanet-Tolosan, France
    Université de Toulouse INPT ENVT, UMR1289 Tissus Animaux Nutrition Digestion Ecosystème et Métabolisme, F-31076 Toulouse, France
    INRA, UMR444 Laboratoire de Génétique Cellulaire, INRA Auzeville, F-31326 Castanet-Tolosan, France
    GeT-PlaGe, Genotoul, INRA Auzeville, F-31326 Castanet-Tolosan, France

Abstract

During the first few weeks after delivery, female rabbits excrete fecal pellets, which are ingested by their pups. We hypothesized that maternal excretion of hard fecal pellets and the coprophagous behavior of their pups were involved in cecal microbiota implantation. Four groups were compared: in 1 group (FM), pups had free access to maternal fecal pellets; in a second group, ingestion of feces was prevented (NF); and in 2 additional groups, pups had access only to fecal pellets excreted by foreign females receiving either no antibiotic (FF) or tiamulin and tetracycline (FFab). A total of 109 litters in 3 batches were used to quantify excretion and ingestion of feces and mortality. Bacterial composition was assessed by 454 pyrosequencing of the V3 to V4 region of 16S RNA genes and fermentative measurements in 128 rabbits of 1 batch at age 14, 35, 49, and 80 d with 8 rabbits per group for each age with 2 rabbits per litter. The number of fecal pellets excreted by does from 2 to 20 d after delivery ranged widely, but was similar among groups (16.1 ± 12.6 fecal pellets/doe). The excretion peaked during the first 6 d after delivery. Foreign fecal ingestion (FF and FFab groups) was 3 times greater (P < 0.001) than ingestion of maternal feces (9.9 ± 7.8). Ingestion of feces in the FF group was greater than in the FFab groups (35.6 ± 9.3 vs. 29.5 ± 9.7; P < 0.05). Compared with the FM group, ingestion of feces in the FF and FFab groups began later (6 to 7 d vs. 2 to 3 d after birth) and peaked at 14 to 17 d (4.0 ± 1.8 hard fecal pellets·litter−1·d−1) and 13 to 15 d (3.5 ± 1.7 hard fecal pellets litter−1 d−1), respectively. During the 36 to 49 d period, the FF and NF groups exhibited the least (2.8%) and greatest (9.5%) mortality, respectively (P = 0.03). At age 14 d, the cecal bacterial community was dominated by Bacteroidetes phyla (63.3 ± 15.1%), Bacteroidaceae family (36.0 ± 18.8%), and Bacteriodes genus (36.0 ± 2.3%). With increasing age, Firmicutes phyla, Lachnospiraceae, and Ruminococcaceae families became the dominant taxa (92.0 ± 4.7, 44.0 ± 13.7, 37.9 ± 11.6% at age 80 d, respectively). Impairment of fecal ingestion delayed this ecological succession, with greater and lower relative abundance of Bacteroidaceae and Ruminococcaceae, respectively, than in the other 3 groups at age 35 d (P < 0.10). In conclusion, although excretion of hard fecal pellets by does ranged widely, the coprophagous behavior of their pups affected the implantation of cecal bacterial microbiota.



INTRODUCTION

At birth, the gastrointestinal (GI) tract of a young mammal is almost sterile. The GI tract is gradually colonized through exposure to a meta-community of microorganisms, of which only a few species that can proliferate in the GI tract will become established (Curtis and Sloan, 2004). Maternal microbiota (vaginal, fecal, oral, and dermal), together with the immediate environment, are key determinants of the initial colonization of the GI of the pups. Colonization of the GI tract is orchestrated in an ecological succession of species that lead to a final mature ecosystem. Besides a function in hydrolysis and fermentation of nutrients providing VFA available to the body, microbiota provide key roles in GI tract development, angiogenesis, and maturation of the immune system, and acts as a barrier against pathogens (Berg, 1996; Hooper et al., 2012). In rabbits, the microbiota is necessary for development of the primary antibody repertoire (Knight and Crane, 1994; Knight and Winstead, 1997).

Rabbit females give birth to relative immature (hairless and blind) pups in a nest after a 31 d gestation. Both in rabbitries and in the natural environment, before parturition, the doe builds her nest using abdominal fur and surrounding plant materials. Unlike most mammal species bearing altricial young, the doe displays only limited mother-young interactions (Coureaud et al., 2010). After birth, the doe visit the nest to suckle her young only once per day (1.12 per 24 h; Hoy and Selzer, 2002) and for a short time (3 to 5 min; Zarrow et al., 1965; González-Mariscal, 2007). After nursing, the doe leaves the nest for 24 h, and in commercial rabbitries, spends the rest of the day in the cage. At nursing, maternal fecal excretion in the nest has been observed, and coprophagous behavior of the pups has been described and partly quantified (Moncomble et al., 2004; Kovács et al., 2006).

Kovács et al. (2006) demonstrated that Bacteroides colonization of the cecum during the first 10 d of life occurred at a reduced rate in pups that had no access to the doe’s feces. Thus, maternal excretion of hard fecal pellets and the subsequent ingestion of doe’s feces by pups in the nest might be involved in the implantation of the GI ecosystem in offspring. Moreover, fecal ingestion may offer an efficient tool to modulate early microbiota establishment to improve health status of the young rabbit. The objectives of this study were to quantify both maternal fecal pellet excretion and pups’ ingestion of feces, and to compare the effect on cecal microbiota implantation of natural feces ingestion behavior with impairment of feces ingestion, and ingestion of foreign feces from does with and without antibiotic medication.


MATERIALS AND METHODS

All animal housing and handling procedures complied with the guidelines for animal research of the French Ministry of Agriculture (Anonymous, 1988).

Animals

In all, 109 does and their progeny (PS Hyplus 19 × PS Hyplus 39, commercial hybrids; Hypharm, Roussay, France) were born and reared at the PECTOUL Experimental Unit, (INRA, Castanet-Tolosan, France). Does were managed in 3 batches, which corresponded to 3 different periods with 37, 38, and 34 does, respectively, in the 3 batches. Average doe BW and doe mean parity were 4,353 ± 377 g and 3.9 ± 1.1, respectively. Does, submitted to a conventional 42 d productive rhythm, were kept in wire cages (width: 61 × length: 68 × height: 35 cm) containing a nest box for pups (width: 39 × length: 27 × height: 35 cm). Cages were maintained in a ventilated breeding unit and under 12 h light (0700 to 1900 h). The nest box had a door that could be closed to control the access of the mother to the nest. Both mothers and growing rabbits had free access to fresh water and were fed ad libitum a pelleted feed manufactured at the PECTOUL Experimental Unit (INRA). The feed was formulated to meet both doe and growing rabbit nutritional requirements (Table 1). The diets contained neither coccidiostatic nor antibiotics. At birth (d 0), litters were adjusted to 8 to 10 pups with no cross-fostering (9.8 ± 0.6). To facilitate control of doe fecal excretion and pup’s hard fecal pellet ingestion in the nest, initial bedding material (wood shavings and maternal fur) was removed and replaced by a layer of wood shavings and a layer of carded cotton. Suckling was controlled by allowing does access to pups until d 20 for 20 to 30 min in the morning. This was the time necessary to open and close all nest boxes in the rabbitry. Within batches, litters were randomly assigned to 1 of 4 treatment groups. In all the groups, fecal pellets excreted by the mothers in nest boxes were counted immediately after suckling and the cotton layer was removed and replaced with a new layer. In the control group with access to feces from mother (FM), maternal fecal pellets were counted and left in nest boxes. In this group, maternal hard fecal pellets were allowed to accumulate in the nest box depending on maternal feces excretion and pups’ consumption. In the group with no access to feces (NF), maternal hard fecal pellets were removed immediately after counting. In the group with access to feces from foreign does without antibiotic treatment (FF) and the group with access to feces from foreign does supplemented with antibiotics (FFab), maternal fecal pellets were removed after counting and replaced with foreign feces (5, 7, and 9 feces from 2 to 13, 14 to 17, and 18 to 20 d, respectively). The objective was to enhance and prolong the possibility for pups to consume feces. Foreign fecal pellets were collected 1 mo before experimentation from 5 nonlactating, nonpregnant females from the same husbandry and fed the same experimental diet. In a preliminary experiment, pups ingested hard fecal pellets from foreign females independently of their physiological state (Gidenne et al., 2013). After a 15-d adaptation to the diet, hard fecal pellets were collected for 5 d and frozen. Foreign females were then medicated in drinking water with tetracycline (50 mg/kg BW) and tiamulin (10 mg/kg BW). The medication corresponded to commonly used antibiotics and dosages applied for digestive and respiratory diseases of does. After a 15-d adaptation to the medication, hard fecal pellets were collected for 5 d and frozen. Each morning before nursing, the number of fecal pellets in the nest box of FF, FFab, and FM females was recorded from d 2 until 20. Weaning occurred at age 35 d, but offspring had access to doe’s solid feed before weaning. After weaning, litters were kept in their birth cages. Maternal excretion of feces in the nest, pups’ ingestion of feces, and pup mortality were recorded daily from age 2 to 70 d for each of the 3 batches. Progeny from batch 1 was sampled for bacterial taxonomic profile and fermentative measurements determination. In batches 2 and 3, only 6 pups per litter were kept in the same wire cages from weaning to age 70 d to comply with density recommendations. Other pups were removed from the experiment.


View Full Table | Close Full ViewTable 1.

Ingredients and chemical composition of the experimental diets fed to rabbits

 
Item Content
Ingredient, % as-fed
    Barley 14.5
    Wheat 15.0
    Soybean meal 12.0
    Sunflower meal 11.0
    Beet pulp 12.0
    Dehydrated alfalfa 33.0
    Carbonate calcium 0.8
    Dicalcium phosphate 0.6
    Mineral-vitamin premix1 0.5
    NaCl 0.5
    L-Lys/HCl, 98% 0.1
Chemical composition
    DM, % 91.4
    Ash, % DM 11.4
    CP, % DM 19.1
    Crude fat, % DM 1.5
    NDF, % DM 33.6
    ADF, % DM 20.1
    ADL, % DM 6.4
1Contents (per kilogram of premix): vitamin A (retinol acetate): 8000 IU; vitamin D3 (cholecalciferol): 600 IU; vitamin E (dl-alpha-tocopheryl acetate): 5 mg; Cu (copper sulfate): 11 mg; Zn (zinc oxide): 30 mg; Mn (manganous oxide): 2 mg; I (potassium iodide): 0.20 mg; Co (cobalt carbonate): 0.20 mg; and Se (sodium selenite): 0.05 mg.

Cecal Sampling

At age 14, 35, 49, and 80 d, 2 pups per litter in batch 1 (or 1 if the number of pups was not sufficient) were sacrificed after an injection (20 mg Nesdonal·kg−1, Merial, Lyon, France; and 0.3 mL T61·kg−1, Intervet, Beaucouzé, France) so that a total of 285 rabbits were sampled (71, 78, 75, and 61 rabbits for NF, FF, FFab and FM groups, respectively). The stomach was isolated and a picture of its contents was taken to confirm the presence of ingested hard fecal pellets. The cecum was isolated and weighed, and samples of the cecal content were collected and stored at −80°C to study the establishment of bacterial community using 16S RNA gene 454 pyrosequencing. The pH (Unitrode with Pt 1000; Metrohm, Herisau, Switzerland) was recorded and 2 samples (1 g each) of fresh cecal content were diluted in storage solutions, 1 in HgCl2 (2 mL, 2% wt/vol) and 1 in H2SO4 (3 mL, 2% wt/vol), for further analysis of VFA and NH3, respectively. Quantifications of VFA were performed by automated GC (Chrompack CP 9000; Chrompack B.V., Middelburg, the Netherlands) according to Playne (1985). The NH3 concentrations were determined using a colorimetric method by a continuous flow analyzer (SAN11; Skalar, Norcross, GA) as described previously (Verdouw et al., 1977). Dry matter was determined in cecal samples by drying at 103°C for 24 h.

To check for the presence of antibiotics, concentrations of tetracycline, and tiamulin were measured using HPLC coupled to an ion trap mass spectrometer (LCQ Deca XP Max; Thermo Scientific, Waltham, MA). The antibiotic assay was performed on 1) 1 sample of hard fecal pellets from 1 biological mother, 2) 1 sample of a composite of hard fecal pellets from foreign does before and 1 after antibiotic treatment, and 3) 4 samples of cecal content of 4 pups (1 per treatment group) at age 14 d. The levels of tetracycline and tiamulin in pooled fecal samples from foreign does after antibiotic treatment reached 38.2 µg/g feces and 0.97 µg/g feces for tetraclycine and tiamulin, respectively. Tetracycline and tiamulin were not detected in feces from foreign does before antibiotic treatment, from the biological mother, or in cecal content of 14-d old pups.

DNA Extraction and Preparations of Sequencing PCR Amplicons

The DNA extraction and sequencing of PCR amplicons were performed on 128 rabbits from batch 1. Samples included 8 rabbits per group for each age with 2 rabbits per litter. To maximize the difference between NF and FF and FFab groups, DNA extractions were performed on litters that exhibited the greatest total feces consumption. Photo-analysis of the stomach content showed hard fecal pellet consumption in 85% of the 14-d old pups in FF and FFab groups and 31% for the FM group. Total DNA from about 0.2 g of cecal sample was extracted and purified (QIAamps DNA Stool Mini kit; Qiagen Ltd, West Sussex, UK) according to the manufacturer instructions after mechanical lyses with a previous bead-beating step (FastPrep Instrumen; MP Biomedicals, Illkirch, France). The quality and quantity of DNA extracts were checked using a spectrophotometer (ND-1000; NanoDrop Technologies, Wilmington, DE).

Amplicons from the V3 to V4 regions of 16S rRNA genes (460 bp on E. coli) were amplified with bacterial forward 343F (TACGGRAGGCAGCAG; Liu et al., 2007) and reverse 784R (TACCAGGGTATCTAATCCT; Andersson et al., 2008) primers as previously described (Zened et al., 2013). Each primer had a barcode sequence of 11 nucleotides on 5’ that was unique for each sample. The PCR assays were performed in a total volume of 50 μL containing 1X PCR buffer, 200 µM of dNTP, 0.5U Isis DNA polymerase (MP Biomedicals), 0.5 μM of forward and reverse primers, and 0.5 to 3 ng of DNA template. The amplification program consisted of an initial denaturation step at 94°C for 2 min; 32 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and elongation at 72°C for 30 s, and a final extension step at 72°C for 7 min. The PCR assays of products were purified (QIAquick PCR Purification Kit; Qiagen) followed by DNA yield quantification and quality estimation (ND-1000 Spectrophotometer; NanoDrop Technologies). The size of the PCR products was confirmed by gel electrophoresis. The purified PCR products were quantified (Quant-iT PicoGreen ds DNA Assay kit; Invitogen, Saint-Aubin, France) on sequence detection system (ABI Prism 7900HT; Life Technologies-Invitrogen-A-BIOSYSTEM, Villebon-sur-Yvette, France) and then mixed in equimolar amounts to a final DNA concentration of 500 ng/μL for each library. The amplicons were pyrosequenced using titanium chemistry, with a titanium sequencer (454 FLX 454, Life Sciences; Roche Company, Branford, CT) at the Genomic and Transcriptomic Platform (INRA). Each forward primer was appended with the titanium sequencing adaptor A, and the reverse primer carried the titanium adaptor B at the 5’ end.

Pyrotag Handling and Analysis

A total of 1,202,455 16S rDNA sequences were sorted based on their respective barcodes representing the 128 collected cecal samples. Sequences were sequentially filtered using a Python script developed (Bioinformatics Platform, INRA,Toulouse, France), first removing sequences with a short sequencing length (less than 150 nt; 60,293 sequences removed), those with at least 1 ambiguous base (78,648 sequences removed), or with a long homopolymer (greater than 9; 78 sequences), those that did not match the proximal PCR primer sequences (with 2 mismatches allowed; 14,001 sequences removed), and finally those having both primers but with a length shorter than 350 pb (5112 sequences removed). A total of 1,011,323 sequences were retained corresponding to 8158 ± 4146 sequences per sample. The average read length was 408 ± 70 nucleotides after trimming the barcodes.

Taxonomic Classification

After cleaning, sequences were analyzed using a MOTHUR software (Schloss et al., 2009). Reads were aligned over a SILVA alignment database provided by a MOTHUR software (14,956 sequences corresponding to the unique sequences in the SSU Ref Database (Pruesse et al., 2007) and an alignment quality was calculated using the SILVA secondary structure map file (17,947 badly aligned sequences were removed). After calculating a pairwise distance between aligned sequences, they were clustered into operational taxonomic units (OTU; cutoff 0.03 using furthest neighbor clustering). Taxonomic assignment was performed using Ribosomal Database Project (RDP) classifier algorithm (Cole et al., 2009) with a bootstrap of 60%. Finally, all unclassified bacteria at the phylum level were considered as chimera and were removed from statistical analysis (21,460 sequences). Richness (e.g., the number of species present in the ecosystem) was estimated using abundance coverage-based estimator (ACE), Chao1 richness estimators, and diversity using Shannon indice were calculated on the 128 pyrosequenced cecal samples using the relevant MOTHUR modules with an OTU definition at a similarity cutoff of 97% (Schloss et al., 2009).

Statistical Analysis

All statistical analyses were performed using R version 2.14.2 (R Development Core Team, 2008). Variables were analyzed by a linear mixed model, including age, and treatment group as fixed effects, and litter as a random effect. For doe excretion and pup consumption of hard fecal pellets, batch was also considered as a random effect. To stabilize variance, a square root transformation was applied to bacterial community taxonomic data. Chi-squared analysis was used to compare mortality rate of the 4 treatments. Because ingestion of feces was impaired in the NF group (ingestion zero), this group was removed from the analysis for the ingestion feces variable. A principal component analysis (PCA) was performed on the taxonomic profiles at the family level to describe the internal structure of the cecal bacterial community. The between-group differences of the cecal bacterial community profiles were calculated using an analysis of similarity (ANOSIM), which results in a P-value yielding the significance of similarity and an R-ANOSIM value yielding the level of similarity (Ramette, 2007). Using pairwise Euclidean distance calculation, bacterial community stability was evaluated between 2 consecutive age groups using the principle of moving window analysis.


RESULTS

Doe Excretion and Pup Consumption of Hard Fecal Pellets

Between 2 and 20 d after birth, the total number of hard fecal pellets excreted by does in their nest during milking was similar among treatments (16.1 ± 12.6 hard fecal pellets per doe), but was extremely variable among does. Five out of 109 does had no fecal excretion, while 7 does excreted more than 40 hard fecal pellets in their nest. The hard fecal pellets excreted in the nest peaked during the first 6 d after parturition, and then decreased regularly (Fig. 1). Twelve days after birth only 13% of the does (14 out of 109) still excreted hard fecal pellets in their nests. Fecal excretion in the nests was not observed 19 d after parturition.

Figure 1.
Figure 1.

Average number of fecal pellets excreted by does in the nest during suckling from 2 to 20 d after parturition. a-eMeans without a common superscript differ (P < 0.05).

 

The number of fecal pellets ingested by pups from 2 to 20 d averaged 10 in the FM group (control group), but varied within litter (9.9 ± 7.8). The fecal pellet intake was 3 times greater in FF and FFab groups than in control group (P < 0.001), and was 5% greater in the FF than in the FFab group (35.6 ± 9.3 vs. 29.5 ± 9.7; P < 0.05). An interaction between age and group (P < 0.001) indicated that the ingestion pattern differed among treatments (Fig. 2). Despite the presence of foreign feces (5 in the nest), the FF and FFab pups started fecal consumption later than those in the FM group. The FF and FFab maximum consumption levels were reached in 14 to 17 d (4.0 ± 1.8 feces·litter−1·d−1) and 13 to 15 d (3.5 ± 1.7 feces·litter−1·d−1). In the FM group, a peak in maximum fecal pellet ingestion was not detected, because no effect of age occurred in this group. From d 11 until the end of the observation period, FM pups exhibited the lowest feces ingestion despite the presence of fecal pellets from their mother in the nest. The FF pups exhibited a greater consumption than FFab only at 17 d (3.8 ± 1.7 vs. 2.7 ± 1.6; P < 0.05) and 19 d (2.5 ± 1.4 vs. 1.6 ± 1.2; P < 0.05).

Figure 2.
Figure 2.

Average number (line) of fecal pellets ingested by pups in the nest from 2 to 20 d after birth of rabbits having free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab). Bars indicate the number of feces available in the nest for pups.

 

Mortality, Growth, and Cecal Fermentative Characteristics

Mortality rate recorded in the 3 batches was different between treatments for the 36 to 49 d period. The FF group exhibited the lowest mortality rate and NF group the greatest (Table 2). Moreover estimation of mortality over the entire period, taking account of the pups euthanized in batch 1, showed a 12 point difference between those in later groups. Assignment of litters among the different treatments led to a greater BW of pups at 2 d in the FF group. Body weight at 14, 35, 49, and 80 d was not affected by treatments (250 ± 44, 887 ± 144, 1,699 ± 217, and 3,134 ± 272 g, respectively), but was influenced by BW at 2 d (P < 0.001). Conversely, the full cecal relative weight of FF and FFab pups was lower at 35 d than those of the FM and NF pups (7.5 ± 1.8, 7.3 ± 1.6, 9.2 ± 1.5, and 9.7 ± 1.9%, respectively; P < 0.05). The expected increase in VFA level and butyrate proportion with age was observed (Supplemental Table S1; see online version at journalofanimalscience.org). Treatment effects were detected on pH and total VFA levels (P < 0.02). The greatest pH associated with the lowest VFA level was observed in FFab pups at age 14 d as opposed to FM pups (P < 0.05). At age 80 d, NF pups exhibited a 30% lower VFA level than FM pups, which had the greatest level (106 mmol; P < 0.05).


View Full Table | Close Full ViewTable 2.

Percent mortality throughout the experiment of rabbit pups (%)1,2

 
Item d 1 to 14 d 14 to 35 d 35 to 49 d 49 to 80 d 1 to 803
FM 3.8 (237) 2.4 (212) 5.8 (191) 4.5 (110) 15.5
NF 5.4 (279) 3.2 (246) 9.5 (220) 6.7 (120) 22.8
FF 4.8 (270) 1.3 (238) 2.8 (215) 0.8 (132) 9.3
FFab 7.8 (282) 3.7 (240) 4.7 (211) 3.9 (127) 13.2
P-value4 0.253 0.383 0.035 0.109 0.003
1Had free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab).
2Number of live rabbits at the beginning of the period is given in parentheses.
3Mortality on whole period took account of the sacrificed animal.
4Mortality differences among groups were test using Pearson’s Chi-squared test.

Taxonomical Profiles, Richness, and Diversity of Cecal Bacterial Community

For all the 128 samples, 9 bacterial phyla were found in the cecum, but 4 represented almost the entire microbiota abundance (more than 99.9% of total sequences; Table 3) and were affected by age but not by group. Bacteroidetes phylum was dominant at 14 d with a relative abundance of 63% and then decreased sharply until 49 d to represent about 10% (P < 0.05). Concomitantly, Firmicutes phylum rose to be dominant by age 35 d (65% of abundance; P < 0.05). Accordingly, the Bacteroidetes to Firmicutes ratio increased from 0.56 at 14 d to 35.5 at 80 d (P < 0.001). Proteobacteria phylum represented around 7% at 14 d and decreased to an abundance 10 times lower at 80 d (P < 0.001), while Actinobacteria phylum exhibited the lowest abundance at 35 d (0.7%) and the greatest at 49 d (1.5%; P < 0.05). Other phyla ( < 0.1% of the total bacterial community), such as Fibrobacteres, Fusobacteria, Synergistetes, Candidate division TM7, and Verrucomicrobia, were not consistently present in all the different cecal samples. Based on the Ribosomal Database Project classifier algorithm, 89.9 and 55.2% of sequences were assigned to family and genus levels, respectively. On the whole sample set, 71 families and 178 genera were detected. The predominant bacterial families belonged to Clostridia and Bacteroidia classes (Lachnospiraceae, Ruminococcaceae and Bacteroidaceae, Rikenellaceae, and Porphyromonadaceae; Table 4). Regarding Proteobacteria phylum, the major families were Enterobacteriaceae and Desulfovibrionaceae. Actinobacteria were mostly represented by the Coriobacterineae family. Bacteroides was the most abundant genus and the only one identified from the Bacteroidaceae family (Table 5). Ruminococcaceae and Lachnospiraceae each contained 4 genera, among which the 2 most abundant corresponded to incertae sedis and uncultured groups.


View Full Table | Close Full ViewTable 3.

Influence of age on the bacterial core phylum relative abundance (%) in the cecum of rabbits1

 
Age, d
P-value
Item 14 35 49 80 SEM Age Group2 Age × group Litter3
Actinobacteria 0.69bc 0.45c 1.52a 0.75b 0.07 0.006 0.13 0.35 0.99
Bacteroidetes 63.35a 32.93b 9.09c 6.68c 2.45 < 0.001 0.16 0.87 0.29
Firmicutes 29.04c 65.26b 87.69a 91.97a 2.59 < 0.001 0.34 0.44 0.24
Proteobacteria 6.91a 1.34b 1.70b 0.60c 0.37 < 0.001 0.08 0.81 0.99
a-cWithin a row, means without a common superscript differ (P < 0.05).
1Based on 8 rabbits per group for each age with 2 rabbits per litter.
2Group: had free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab).
3Random effect.

View Full Table | Close Full ViewTable 4.

Influence of age on the bacterial core family relative abundance (%) in the cecum of rabbits1

 
Age, d
P-value
Family 14 35 49 80 SEM Age Group2 Age × group Litter3
Lachnospiraceae 19.83b 37.00a 45.50a 43.97a 1.47 < 0.001 0.019 0.405 0.525
Ruminococcaceae 6.86b 23.04a 34.11a 37.95a 1.48 < 0.001 0.053 0.428 0.054
Bacteroidaceae 36.03a 23.22a 3.21b 2.38b 1.82 < 0.001 0.080 0.954 0.001
Rikenellaceae 6.77a 2.97b 2.57b 1.42b 0.41 < 0.001 0.864 0.804 0.240
Porphyromonadaceae 9.43a 2.37b 0.71b 0.76b 0.47 < 0.001 0.307 0.618 0.302
Enterobacteriaceae 3.63a 0.02b 0.03b 0.05b 0.28 < 0.001 0.952 0.887 0.050
Desulfovibrionaceae 1.75 0.96 1.18 0.27 0.14 0.013 0.118 0.720 1.000
Coriobacterineae 0.67bc 0.44c 1.46a 0.73b 0.07 0.007 0.059 0.331 1.000
Family XIII Incertae Sedis 1.02a 0.17b 0.21b 0.15b 0.05 < 0.001 0.118 0.843 1.000
S24–7 0.02b 0.30b 0.26b 0.44a 0.04 < 0.001 0.399 0.541 0.026
Campylobacteraceae 0.98a 0.01b 0.01b 0.01b 0.06 < 0.001 0.197 0.762 1.000
Prevotellaceae 0.00b 0.14ab 0.48a 0.20ab 0.05 0.009 0.956 0.555 < 0.001
Anaeroplasmataceae 0.03b 0.10ab 0.38ab 0.26a 0.06 0.001 0.424 0.130 0.426
Erysipelotrichaceae 0.01 0.01 0.11 0.51 0.03 < 0.001 0.879 0.210 0.223
BSV13 0.43a 0.15b 0.03c 0.02c 0.03 < 0.001 0.634 0.141 0.307
a-cWithin a row, means without a common superscript differ (P < 0.05).
1Based on 8 rabbits per group for each age with 2 rabbits per litter.
2Group: had free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab).
3Random effect.

View Full Table | Close Full ViewTable 5.

Influence of age on the bacterial core genus relative abundance (%) in the caecum of rabbits1

 
Age, d
P-value
Family genus 14 35 49 80 SEM Age Group2 Age × group Litter3
Bacteroidaceae Bacteroides 36.03a 23.23a 3.21b 2.38b 1.82 < 0.001 0.080 0.953 0.010
Porphyromonadaceae Barnesiella 6.40a 1.28b 0.38b 0.44b 0.40 < 0.001 0.892 0.187 1.000
P. Parabacteroides 1.98a 0.81b 0.28c 0.26bc 0.13 < 0.001 0.057 0.346 0.068
Rikenellaceae Alistipes 5.45a 2.67b 2.48b 1.17b 0.34 < 0.001 0.818 0.826 0.549
R. RC9 0.72 0.12 0.00 0.19 0.17 0.917 0.588 0.378 < 0.001
Lachnospiraceae Blautia 1.51 2.85 3.22 1.14 0.19 0.827 0.279 0.197 1.000
L. Coprococcus 0.13b 0.25ab 0.29a 0.52a 0.03 < 0.001 0.140 0.159 0.874
L. Incertae_Sedis 3.30 5.21 7.59 3.96 0.44 0.105 0.016 0.183 0.152
L. Moryella 0.45b 1.19ab 2.47ab 3.81a 0.31 < 0.001 0.448 0.171 < 0.001
L. uncultured 0.37b 6.68a 5.39a 7.46a 0.44 < 0.001 0.127 0.888 1.000
Ruminococcaceae Hydrogenoanaerobacterium 0.05c 0.66ab 1.27a 0.22b 0.10 < 0.001 0.398 0.220 0.038
R. Incertae_Sedis 0.73a 5.52b 8.25b 6.80b 0.40 < 0.001 0.218 0.940 1.000
R. Ruminococcus 0.01c 1.11b 2.06a 3.17a 0.16 < 0.001 0.168 0.333 1.000
R. Subdoligranulum 0.23b 1.90a 2.21a 4.23a 0.36 < 0.001 0.529 0.084 0.035
R. uncultured 0.25c 2.40b 6.49a 8.16a 0.39 < 0.001 0.426 0.814 0.021
Erysipelotrichaceae Allobaculum 0.00c 0.00c 0.10b 0.51a 0.03 < 0.001 0.505 0.278 1.000
Desulfovibrionaceae Desulfovibrio 0.83 0.49 0.61 0.10 0.08 0.285 0.227 0.318 1.000
Campylobacteraceae Campylobacter 0.98a 0.01b 0.01b 0.01b 0.07 < 0.001 0.197 0.762 1.000
Enterobacteriaceae Escherichia 0.81a 0.00b 0.01b 0.01b 0.10 < 0.001 0.874 0.141 0.755
E. Klebsiella 0.78 0.02 0.01 0.04 0.07 < 0.001 0.838 0.959 0.753
a-cWithin a row, means without a common superscript differ (P < 0.05).
1Based on 8 rabbits per group for each age with 2 rabbits per litter.
2Group: had free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab).
3Random effect.

A PCA was performed on the 128 taxonomic profiles. The first 3 principal components (PC) explained 50.6% of the total variation (Fig. 3 and 4). The first PC was negatively correlated to Bacteroidaceae and Porphyromonadaceae, and positively correlated to Ruminococcaceae. This first PC was strongly related to rabbit age. The bacterial communities sampled in the youngest rabbits were characterized by high relative abundance of Bacteroidaceae and Porphyromonadaceae, and were opposed to the oldest rabbits, which exhibited a high level of Ruminococcaceae. The second PC showed opposite effect on the Enterobacteriaceae and Prevotellaceae families, which were positively and negatively correlated with this second PC, respectively (Fig. 3). The first 2 PC allowed a separation of the taxonomic profile for 14 d-rabbits. The variability among this group was mainly related to the level of Enterobacteriaceae. Finally, the third PC was related to relative abundance of the S24–7 family, and allowed a separation of the bacterial community of the oldest rabbits according to their level of Lachnospiraceae, Ruminococcaceae, and S24–7 (Fig. 4). In contrast to the other groups, NF 35-d-old rabbits were on the left side of the plot near the 14-d-old rabbits, and their bacterial community still exhibited a high level of Bacteroidaceae and a low level of Ruminococcaceae. In 3 other groups, only 2 (FF) or 3 (FFab and FM) pups out of 8 were on the left side of the plot (Fig. 3 and 4). This observation was confirmed by an analysis of similarity (Supplemental Table S2 ; see online version at journalofanimalscience.org), which showed that at 35 d only the NF bacterial community could be separated, but overlapped the bacterial community of the other 3 groups (R-ANOSIM = 0.45 and P = 0.003, R-ANOSIM = 0.18 and P = 0.035, and R-ANOSIM = 0.17 and P = 0.06 in FF, FFab, and FM groups, respectively). Also, NF taxonomic profiles at 49 and 80 d were separated but overlapped only in the NF group (R-ANOSIM = 0.36 and P < 0.001), while they were similar between these 2 ages in the other 3 groups (R-ANOSIM < 0.12 and P = 0.068).

Figure 3.
Figure 3.

Principal component analysis (PCA) projection on the first (PC1) and the second (PC2) PC of bacterial community taxonomic profile at the family level from 454 pyrosequencing run of rabbit cecal samples. The 4 plots illustrate the same analysis performed on the whole data set (n = 128 rabbits). The marks relate to the 4 experimental groups of rabbits having free access to mother feces in the nest (+, FM, control group), whose fecal ingestion in the nest was impaired (○, NF), or that had only access to feces from foreign does receiving either no antibiotics (▲, FF) or medicated with tiamulin and tetracyclin (■, FFab), and numbers indicate the age of rabbits at sacrifice. The numbers shown in parentheses correspond to percentage of explained variance of the PC.

 
Figure 4.
Figure 4.

Principal component analysis (PCA) projection on the first (PC1) and the third (PC3) PC of bacterial community taxonomic profile at the family level from 454 pyrosequencing run of rabbit cecal samples The 4 plots illustrates the same analysis performed on the whole data set (n = 128 rabbits). The marks relate to the 4 experimental groups of rabbits having free access to mother feces in the nest (+, FM, control group), whose fecal ingestion in the nest was impaired (○, NF), or that had only access to feces from foreign does receiving either no antibiotics (▲, FF) or medicated with tiamulin and tetracyclin (■, FFab), and numbers indicate the age of rabbits at sacrifice. The number shown in parentheses correspond to percentage of explained variance of the PC.

 

Using moving analysis window, the dynamics of the bacterial community was assessed by measuring distances between 2 consecutive ages for each group (Supplemental Table S3; see online version at journalofanimalscience.org). The bacterial community was more stable as the distance calculated on the taxonomic profile between 2 consecutive ages was low. Between 14 and 35 d, the smallest distance was calculated for the NF group (P < 0.05), reflecting a slower evolution of the bacterial community compared with the other 3 groups. By contrast, between 49 and 80 d the smallest distance was observed in the FF and FFab groups (P < 0.05). Between 35 d and 49 d the smallest distance was observed in the FF groups.

The separate analysis of each family’s relative abundance revealed 3 different patterns for their abundance time course (Table 4). In the first pattern, the relative abundance increased with age. This pattern was observed in the 2 most abundant families (Lachnospiraceae and Ruminococcaceae), which peaked by 35 d, but also the S24–7, Anaeroplasmataceae and Erysipelotrichaceae families, whose maximal abundances were observed only at 80 d. In the second pattern, the relative abundance peaked at 14 d and decreased thereafter. This pattern clustered Bacteroidaceae and BSV13 families, whose minimum level was reached at 49 d, Rikenellaceae, Porphyromonadaceae, Enterobacteriaceae, Family_XIII_incertae sedis, and Campylobacteraceae families, whose minimum level was reached by 35 d. Finally, the fluctuation of the relative abundance of Prevotellaceae and Coriobacterineae families belonged to a third pattern, with no net change over time. Eight out of the 20 abundant genera exhibited an age-related increase, but 6 out of the 20 abundant genera showed an age-related decrease (Table 5). Bacteroides accounted for more than one-third of total bacterial community at age 14 d, but decreased sharply after weaning. By contrast, the frequencies of detection of uncultured Lachnospiraceae, incertae sedis, and uncultured Ruminococcaceae represented less than 1% at age 14 d and became the most abundant genera at age 80 d (6 to 8% of total bacterial community). Concerning treatment effects, the FFab group exhibited the lowest level of Lachnospiraceae at age 14 and 49 d (P < 0.10 and 0.05, respectively), and the greatest level of Ruminococcaceae at 49 d (P < 0.10; Fig. 5A and B). The NF group exhibited the lowest level of Ruminococcaceae and the greatest level of Bacteroidaceae at 35 d (Fig. 5B and 5C), consistent with the position of this group on the PCA plot (Fig. 4). Finally, the ratio of Lachnospiraceae to Ruminococcaceae was greater in the NF group than in the FFab group (Fig. 5D). At the genera level, a group effect was observed only in incertae sedis Lachnospiraceae (P < 0.05), Bacteroides and Barnesiella (P < 0.1). Richness e.g., the number of species in the ecosystem, evaluated by ACE and Chao1 estimators exhibited a threefold increase between d 14 and d 80 (Table 6); the diversity evaluated by the Shannon index increased (+ 1.43 points). No effect of treatments could be observed on these measurements.

Figure 5.
Figure 5.

Relative abundance of Lachnospiraceae (A), Ruminococcaceae (B), Bacteroidaceae (C) family and Lachnospiraceae Ruminococcaceae ratio (D) in bacterial cecal community of rabbits having free access to mother feces in the nest (+, FM, control group), whose fecal ingestion in the nest was impaired (○, NF), or that had only access to feces from foreign does receiving either no antibiotics (▲, FF) or medicated with tiamulin and tetracyclin (●, FFab) at 14, 35, 49, and 80 d of age (values are mean ± SE). a,bWithin an age group, means without a common superscript differ (P < 0.05). A, B Within an age group, means without a common superscript differ (P < 0.10).

 

View Full Table | Close Full ViewTable 6.

Influence of age on bacterial richness and diversity estimators in the cecum of rabbits1

 
Age, d
P-value
Item 14 35 49 80 SEM Age Group2 Age × group Litter3
No. of sequences 7,162 7,720 6,997 8,480 330 0.348 0.362 0.12 0.471
No. OTU4 1,334c 2,197b 2,437b 3,329a 114 < 0.001 0.505 0.044 0.999
Chao15 2,324c 4,536b 5,290b 7,625a 273 < 0.001 0.320 0.062 0.996
ACE5 3,090c 6,808b 8,252b 12,279a 452 < 0.001 0.167 0.067 0.993
Shannon indice6 5.68c 6.48b 6.73b 7.11a 0.06 < 0.001 0.378 0.300 0.320
a-cWithin a row, means without a common superscript differ (P < 0.05).
1Based on 8 rabbits per group for each age with 2 rabbits per litter.
2Group: had free access to mother feces in the nest (FM, control group), whose fecal ingestion in the nest was impaired (NF), or that had only access to feces from foreign does receiving either no antibiotics (FF) or medicated with tiamulin and tetracyclin (FFab).
3Random effect.
4Operational taxonomic unit.
5Chao1 and ACE (abundance coverage-based estimator) are richness estimators of ecosystems.
6Shannon indice is a diversity indice of ecosystem.


DISCUSSION

Maternal Fecal Excretion in the Nest and Pup Coprophagous Behavior

The rabbit is a herbivorous lagomorph that ingests its cecotrophs (directly from the anus; Gidenne and Lebas, 2006), but coprophagous behavior is not observed in weaned rabbit. Maternal feces excretion behavior in the nest and coprophagous behavior of the pups has been recently described (Moncomble et al., 2004; Kovács et al., 2006), but never quantified from 2 to 20 d. Moncomble et al. (2004) observed an average excretion of 2 hard fecal pellets per day during the first 13 d after parturition on 10 litters. Similarly, in this study, we observed on 109 litters that the does excreted an average of 16 hard fecal pellets between 2 and 20 d after birth, but this excretion varied among does. Nursing stimulates secretion of pituitary hormones, prolactin, and oxytocin (Fuchs et al., 1984). In addition to a role in contraction of myoepithelial cells to facilitate milk secretion and uterine smooth muscle contraction during parturition (Gimpl and Fahrenholz, 2001; Lollivier et al., 2006), recent studies indicate that oxytocin may also affect GI motility (Li et al., 2007). Therefore, excretion of feces in the nest might result from oxytocin secretion during suckling. The observed variability among does might result from a differential level of oxytocin production or sensitivity.

As in the previous study conducted by Kovács et al. (2006), the coprophagous behavior of the pups was confirmed by the disappearance of maternal fecal pellets from the nest and by the presence of solid fecal particles in the gastric content of the pups. Ingestion of maternal feces were observed as early as 2 d after birth and averaged about 1 fecal pellet per pup between 2 and 20 d after birth. In rabbits, this coprophagous behavior might correspond to an adaptive response of pups to the parsimonious interaction of the doe with their litter and would, thus, promote the colonization of the GI and affect cecal microbiota implantation as suggested by Kovács et al. (2006). Preliminary studies in our laboratory showed that pups also accepted hard fecal pellet from foreign does in the same or different physiological states (Gidenne et al., 2013). When hard fecal pellets from foreign does were provided in the nest, the feces ingestion by pups was 3 times greater than in natural conditions. This difference could not be ascribed to a lack of fecal pellets in the nest. The greater consumption might be explained by greater palatability of foreign feces. The foreign fecal pellets were frozen immediately after their excretion and thawed just before distribution. Fresh thawed feces, thus, seemed more attractive than feces excreted 10 d earlier. This is supported by the reduced consumption of foreign feces when the mother was supplemented with antibiotics (FFab group vs. FF groups). Fecal odor may depend on fermentation processes and microbiota composition, which is altered during antibiotic supplementation (Willing et al., 2011). Impairment of ingestion of feces had no effect on growth, but a lower proportion of cecum contents were observed in FF and FFab groups.

Ecological Succession of Bacterial Taxa in Rabbit Cecum

In rabbits, as in other mammals, microbiota colonization is orchestrated in an ecological succession of species (Gouet and Fonty, 1973, 1979; Kovács et al., 2006, Combes et al., 2011). At 14 d, the cecal bacterial community was dominated by Bacteroidetes phyla and Bacteriodes genus but variability was also related to Enterobacteriaceae. Increasing the densities of cecal Enterobacteriaceae in 2-wk old rats had lasting effects on intestinal inflammation and altered barrier properties (Fåk et al., 2008). However, the frequencies of detection of this latter family were almost 100 times lower at the 3 ages studied. With age, Firmicutes phyla became the overwhelmingly dominant phyla, as previously observed using Sanger cloned based technique (Abecia et al., 2005; Monteils et al., 2008). Inside the Firmicutes phyla, the dominance was shared among Lachnospiraceae and Ruminococcaceae families that gather anaerobic and morphologically diverse taxa (Euzéby, 1997), but the dominant genera at age 80 d, correspond to non defined genera, underlining the poor knowledge of the taxonomic diversity of the rabbit cecal ecosystem. However, the specificity of the composition of the cecal ecosystem established was confirmed using the culture technique, namely the reduced occurrence of Lactobacillus, Streptococcus, and Escherichia (Gouet and Fonty, 1973). As expected, fermentative activity took place progressively with a decrease in NH3, an increase in total VFA content of cecum, and a typical inversion of the proportion of propionate and butyrate (Padilha et al., 1995; Combes et al., 2011).

Newborn Rabbit Coprophagous Behavior Affected Cecal Microbiota Implantation

Ingestion of maternal hard fecal pellets in the nest affected the implantation of the pup cecal microbiota. The bacterial community in pups that could not practice this coprophagous behavior displayed a delayed ecological succession of species in the cecum. By contrast, enhanced and prolonged ingestion of hard fecal pellet accelerated the microbiota implantation in the cecum. The maturation delay observed at 35 d in pups that had no access to maternal feces was partly explained by a reduced abundance of Ruminococcaceae and a greater abundance for Bacteroidaceae than in the other 3 groups. At weaning, the microbiota profiles of pups with no access to maternal fecal pellets were closed to a 14-d-old rabbit taxonomic profile. Our assumption agreed with culture-based results of Kovács et al. (2006), who observed that Bacteroides cecal colonization during the first 10 d of life occurred at a reduced rate in pups that had no access to maternal feces. However this delay in microbiota implantation could not be linked to modification of fermentative chracteristics. Moreover, the differences observed at age 80 d in NF and FM pups, were mainly attributed to 2 rabbits belonging to the same litter in the NF group.

The FFab group was used to validate that fecal ingestion could offer an efficient, easy-to-use leverage to control rabbit microbiota implantation. The ingestion of fecal pellets from foreign females treated with a broad-spectrum antibiotic dosage of tiamulin and tetracycline slightly affected the taxonomic profile of the cecal bacterial community, mainly through a reduction of the relative abundance of Lachnospiraceae. Therefore, eating feces containing antibiotics might help to select a particular bacterial population, at least at age 14 d. However, further studies are needed to validate this method to control the GI microbiota.

Ingestion of Feces in the Nest Reduced Mortality Rate

Finally, an interesting result of this study was the link between fecal ingestion, microbiota maturation, and health status of rabbits after weaning. The rabbit pup mortality was reduced in the group with the greatest maternal fecal pellet consumption, but was greater in pups with no access to maternal fecal pellets. Mortality was associated with a delayed implantation of cecal microbiota. In rabbits, digestive disorders and mortality occur mainly around weaning when the cecal microbiota composition is still highly variable and the diet changes from mixed milk and solid feed to only solid feed. Conversely, at age 70 d, the cecal microbiota composition is more stable (Combes et al., 2011) and rabbits are less subject to digestive disorders. One way to overcome mortality around the weaning might be to accelerate the implantation process to reach an adult climax community before weaning. This suggestion is supported by the observation that pups with no access to fecal pellets also reached the latest adult pattern of GI microbiota.

In conclusion, hard fecal excretion by the doe into the nest during lactation is a variable behavior. The coprophagous behavior of the pups toward these maternal fecal pellets was modified and enhanced when foreign fecal pellets were provided. Impairment of coprophagy delayed the cecal bacterial community implantation and enhanced mortality in the pups. Conversely, bacterial community implantation and health status were improved when hard fecal pellet ingestion was enhanced. These results provide new methodologies for control of rabbit pup microbiota with the objective of improving health.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.