1. Introduction
Epizootic rabbit enteropathy (ERE) is a digestive disease that has negatively affected international rabbit production since the 1990s [
1], with morbidities of up to 90% and mortalities of 80% [
2]. ERE generally affects rabbits aged 3–7 weeks, reducing their daily food intake [
3]. The earliest reports of ERE in Mexico were from late 2001 and early 2002 [
4], and currently, ERE represents 32% of the digestive disease cases involved in rabbit production in Mexico, with the loss numbers of weaned rabbits being around 12–20% [
5]. Although the etiology of ERE has not been fully described and identified so far, the most common treatment of ERE involves using oral antibiotics [
6], indicating a probable bacterial origin. In keeping with the assumption that ERE has a bacterial basis, we have shown that several bacterial phyla are associated with this disease [
7], although no single organism was identified as being the cause. However, since there is a trend towards avoiding antibiotics in rabbits’ diets, because there is a fear that overuse could help to promote resistant bacterial strains [
8], this means that, in large outbreaks, the extensive use of the treatment could be problematic in the longer term. In addition, the inclusion of antibiotics in the food can lead to changes in the diversity of the normal gastrointestinal microbiota, causing a microbial imbalance, which can lead to dysbiosis [
9]. The effects of antibiotics have already been investigated in rabbits, with different levels of effects being reported for different antibiotics [
10,
11], with some showing a clear impact on the microbial population. ERE is also characterized by a marked imbalance in the microbial community of the caecum [
2,
7], which, in turn, changes the profile of the metabolites produced and absorbed in the caecum [
12]. Combining these pieces of information led to the hypothesis that an increase in the neutral detergent fiber (NDF) within the diet modulates the microbial communities and the metabolites they produce. This then has the potential to reduce the effects of ERE in rabbits during the fattening stage of their production for meat. In most nutritional studies with herbivores, the change in the dietary fiber content is assumed to impact mainly on the gut microbial community [
13], although this, in turn, can influence the metabolites available for use by the host rabbit and can have additional benefits for the host, such as influencing the host’s anti-inflammatory signaling pathways [
14].
While this increase in dietary fiber shows some promise as a means of impacting on the ERE levels, the underpinning microbiology behind it is less clearly understood. Therefore, the objective of this study was to compare the effect of low- (32% NDF) and high (36% NDF)-fiber diets, in conjunction with rabbits either having previously shown, but having now recovered from, ERE versus those that have never shown any clinical signs of ERE, to determine the impact on the microbial community and the metabolites they produce. This was performed by a combination of assessing short-chain fatty acid (SCFA) concentrations produced as a means of examining the metabolites arising from the digestion of food and the microbiota to evaluate the organisms involved in performing these roles.
3. Discussion
A recent review of the factors associated with ERE and how to treat it [
16] included the use of elevated fiber levels as one mechanism of treatment. This approach constituted the underpinning experimental design of this work. In addition to examining the fiber levels, the presence or absence of ERE at the outset of the experiment impacted on the productive parameters evaluated in this experiment. This effect of ERE is in agreement with a previous study [
17], where it was reported that rabbits that survive ERE show lower weights compared to healthy rabbits in the same system. In the case of food consumption, groups with initial signs of ERE had a lower intake than healthy rabbits (data not shown). Those groups that were healthy at the outset, T1 and T2, had values in keeping with those reported previously for rabbits on diets with lower (≤30%) NDF [
18], which are associated with a shorter retention time in the cecum and, in turn, an increase in food intake [
19]. By way of contrast, the current work looks at the differences in the gut microbiota and associated metabolome in rabbits that have recovered from ERE versus those that have never suffered from it. As such, it is not an investigation of the causal agents of ERE but, rather, how closely rabbits respond and return to normality after an infection, together with the roles that different dietary fiber levels can promote. For this reason, we had an interest in examining both hard feces that are defecated and lost and also caecotrophs that are re-ingested, as the post-ERE recovery is poorly understood in both.
Rabbits that had ERE and were used during this trial had no more than level 2 symptoms. The reason for this was to maximize the likelihood that rabbits would recover from ERE and then survive to the end point of the experiment (70 day old), as those showing severe symptoms (level 3) are much less likely to survive, and our interests lay in examining the similarities between recovered rabbits and those that had never had symptoms. The effect that the diets had on diarrhea and mucus in feces is consistent with that reported previously [
20], although the previous work involved a reduction of NDF from 30% to 25%. This produced an increase in the biodiversity of the microbiota in the ileum but reduced biodiversity in the caecum. In addition to this, a decrease in the presence of diarrhea (from 31% to 0.8%) was reported at day 20 in rabbits with ERE supplemented with 33.6% of NDF [
21]. This was associated with a change in the caecotrophic ecosystem, leading to a higher production of volatile fatty acids [
21] because of a change in the microbial profile of the caecum.
The concentrations of ammonia observed in groups T1 and T2 are similar to those obtained previously [
22] with clinically healthy rabbits. These values are associated with a high absorption of ammonia from the intestinal mucosa of the rabbit [
12]. Here, the ERE groups (T3 and T4) showed higher values than those reported in clinically healthy rabbits, although these differences were not significantly different. Differences might have been expected under these circumstances, as this could be associated with two different factors triggered by the ERE. First, there could have been damage to the intestinal mucosa, where the apoptosis of epithelial cells has been observed, together with atrophy, fusion, and destruction of the intestinal villi, and a reduced infiltration of inflammatory cells [
23]. This resulted in the poorer absorption of digestive products into the bloodstream. Secondly, changes in the intestinal microbiota can also have an impact [
2,
24], with an increase in the abundance of bacteria of the genus
Clostridium causing lesions in the intestinal mucosa due to the toxins generated [
25], as well as an increase in ureolytic activity [
21,
25].
The indices for the diversity values observed are in keeping with a previous experimental work, where changes in the concentration of NDF led to changes in clinically healthy rabbits [
4]. Regarding the changes in the richness of the microbiota, these agree with that reported in previous studies [
2,
24,
26] for the ERE group that received the diet with 36% NDF. However, the genera affected in our study differ from those reported in previous studies [
2,
24]. The use of 36% NDF makes it is possible to manipulate the caecotrophic microbiota of rabbits with clinical ERE. This, in turn, may reflect a decrease in proteolytic activity and an increase in the hemicellulolytic communities of rabbits that are positive for ERE.
It has already been documented that the rabbit gut microbial community is different from that of other domesticated herbivores such as ruminants or horses [
27,
28], in addition to differences seen between different areas of the digestive tract of the rabbit [
29]. In addition, previous next-generation sequencing approaches have shown differences between caecotrophic samples from four virulent (i.e., affected by ERE) versus two non-virulent rabbits [
30]. In the current work, in terms of the population diversity of caecotrophic samples, the highest value of the Shannon Index (
Table 3) was seen in the T3 group (rabbits that had originally had ERE and were fed on the low-fiber diet). This is particularly interesting, as the Shannon index places a high emphasis on species richness [
27] but without the same emphasis on evenness that is seen in the Simpson index. This would suggest that there is still a very diverse population present but that the spread of the species has been affected. In keeping with this hypothesis, it would be anticipated that T3 would probably also have a higher Chao1 value, which measures numbers in the population. This is indeed the case (
Table 3). These differences manifested at the level of a dietary consideration (
p = 0.022 and
p = 0.020 for Shannon and Chao1, respectively), ERE status (
p = 0.004 and
p = 0.013 for Shannon and Chao1, respectively), and the dietary*disease interaction (
p = 0.013 and
p = 0.024 for Shannon and Chao1, respectively).
It is also interesting to note that, when the Abundance-based Coverage Estimators (ACE) index is used, there is a difference between the samples from the rabbits that previously had ERE relative to those that had never had it. The ACE index is used to measure the number of rare OTUs detected [
31], and the data in
Table 3 suggest that there are differences in rare OTUs on the basis of whether a rabbit has had ERE or not. Moreover, there were no differences detected with the Simpson index, which looks at the abundance of dominant species [
32], therefore suggesting that there has been no significant change in the number of abundant species. Thus, in some respects, the post-ERE microbial population is still different from that seen in rabbits that have never had the condition. Interestingly, the fecal samples did not show a similar pattern (
Table 4), with higher Simpson Indices in the T1 and T2 samples (both groups have rabbits that have never suffered from ERE), indicating certain species are more dominant than others in this group relative to those that were seen in the rabbits that had originally suffered from ERE (
p = 0.004). This pattern was also seen for rabbits with the higher fiber content in their diet (
p = 0.002), with an increase in species dominance. However, this was not observed for a dietary*disease interaction (
p = 0.279). Interestingly, the Chao1 value was also higher for the rabbits that never suffered from ERE (
p = 0.033), and this is a measure of the number of species within the population [
33]. This would suggest that, although there is a general level of dominance of the species, there are also a larger number of species present, possibly with some approaching trace levels. The largest Chao1 value is in T2 (rabbits that had never had ERE and were fed a 36% NDF diet), although this does not allow a significant difference to be seen in terms of dietary*disease interactions (
p = 0.062). However, a significant difference in terms of this interaction was seen using the Abundance-based Coverage Estimator (
p = 0.025), which is a measure of the detection of species above a certain minimum baseline [
33]. This is despite not showing significant differences for either the diet (
p = 0.092) or disease status (
p = 0.082). However, no significant differences were observed for the Shannon Indices, which consider species richness [
27].
Taken together, these results suggest differences between caecotrophic and fecal samples in terms of how they continue to be affected by the host rabbit having had ERE. In the caecotroph samples from rabbits that had previously had ERE and were fed on a lower fiber diet, increased species richness (Shannon Index) and diversity (Chao1) values without an increase in the dominance of certain species was shown. Conversely, the fecal samples showed an ERE effect when using the Simpson Index and a similar pattern for the Chao1 values. In addition, the higher fiber diet from rabbits that had never had ERE showed a higher level of low-abundance OTUs (Abundance-based Coverage Estimator) in fecal samples. In terms of identifying an individual species associated with the condition, as with other works, instead, we identified more than one organism that has been affected significantly in terms of its abundance even after the rabbit has recovered from ERE symptoms. What is clear though is that the differences detected in the sequence data from the caecotrophic samples do not translate to the equivalent information in fecal samples (
Table 5). In this context, it is evident that there was no large-scale change to the organisms involved in causing ERE to persist at the level of the phylum. Clearly, any differences must be at a lower taxonomic level. It remains impossible to pinpoint a single organism as the causal agent, although one of the most interesting is probably
Akkermansia muciniphila, which has a high abundance in the ERE positive groups in previous works. This is a microorganism that has been also associated with rabbits fed on lower fiber and that developed ERE [
34], and we have previously reported this as an important microorganism related with ERE in Mexico [
7]. It is a member of the phylum Verrucomicrobia, a phylum that showed no obvious patterns in terms of the abundance between the various samples collected in terms of caecotroph versus fecal, post-recovery from ERE versus never having been infected, and the two levels of fiber in the diet (
Table 5). Moreover,
Akkermansia was the only genus observed within the sequences identified as belonging to the phylum Verrucomicrobia (
Table S1).
Another organism of interest from our previous work [
7],
Cloacibacillus porcorum (which is a member of the phylum Synergestes) was not detected in the current work as the only genus detected here from the phylum Synergestes was Synergestes (
Table S1). Members of this phylum were present at very low levels in all fecal samples (irrespective of their ERE status) and caecotroph samples that recovered from ERE but not in those that never had ERE. However, specifically for
Cloacibacillus porcorum, its absence in the current work may be an indication that it has been lost following recovery from ERE in those rabbits that previously had it.
The final group of organisms that we highlighted as a potential causal agent for ERE were members of the genus Clostridium. Although the Firmicutes were present as one of the most abundant phyla (in some cases, the most abundant phylum) in all the samples, this was due primarily to the high abundance of members of the Class Ruminococcaceae. In terms of members of the genus Clostridium, these were only present at very low levels in the fecal samples and in the caecotrophs collected from previously ERE rabbits on the low-fiber diet. Furthermore, due to the wide range of species of bacteria within the genus Clostridium, these cannot be assumed to be the same species as were identified in the previous work.
Therefore, none of the organisms we highlighted in our previous work [
7] as being potential causal agents for ERE could be detected in the samples here, even for rabbits that had recovered from ERE. This means that, within the limits of detection here, those organisms were not present in rabbits that had recovered from ERE.
In the wider context, our data identify similar levels of Bacteroides but lower levels of Firmicutes than were seen in the previous work for both fecal and cecal samples, e.g., [
35], although these authors used a much lower fiber content in the diet (18.7%) than either of our sets of conditions. However, it is worth noting that, in our previous work [
7], we observed a lower abundance of Firmicutes than has been published previously, which we speculated may be, at least in part, be due to the removal of a PCR step prior to sequencing, as the PCR stages can introduce a bias in terms of amplification. The values we see here for Firmicutes in the fecal samples are in keeping with those we reported previously. However, the caecotroph values are lower than that, although similar to those seen in wild rabbits [
29]. Due to the differences detected in terms of Firmicutes relative to the other work, this makes any inferences to the Firmicutes:Bacteroides ratio or the impact SCFA production relative to the Firmicute numbers difficult to achieve.
Interestingly, members of the phylum Tenericutes were only observed in the cecum of rabbits that had never suffered from ERE and were absent from all fecal samples. The detection of Tenericutes in cecal samples of rabbits, with an absence in fecal samples, has been reported previously, although Tenericutes were found in fecal samples from guinea pigs in the same work [
29]. This differs from the observations in another work [
35], which detected Tenericutes in both fecal and cecal samples from rabbits at 8.17% and 7.48%, respectively.
To date, the roles of many of the bacteria identified in the digestive tract of rabbits remain unclear. This is true both in terms of their function in healthy rabbits and also in those that are either currently suffering from ERE or have previously suffered from ERE, as well as in response to different levels of dietary fiber. For example, members of the Class Synergistia (specifically the genus
Synergistes) were detected at low levels in fecal samples in groups of rabbits but only in cecal samples from rabbits that had previously suffered from ERE, irrespective of their dietary fiber content (
Table S1). This is interesting in the context of a previous work that reported that this group of organisms increased in abundance in the cecum of rabbits that were exposed to zearalenone, a mycotoxin that is known to affect the structure of the cecal microbiota [
36]. Thus, this could be speculated to be an example of organisms that may act as stress indicators, although this needs further investigation.