Poultry is a major source of wholesome (low fat, high protein) nutrition around the world, in addition to being one of the most economical of livestock production species [1
]. Ensuring the health of laying hens and maintaining the economy of poultry products (eggs and meat) is crucial for global food security and sustainability. Two major upcoming shifts in U.S.-based poultry production models have emerged as significant challenges to be addressed within the next decade; firstly, the shift from caged to cage-free production systems prompted by consumer concerns about welfare, and secondly, the U.S. ban on antibiotic usage in livestock production [2
]. These shifts have brought both the rearing environment and gut health into sharp focus. Even as poultry gut microbiota has been well investigated for several years [3
], Kogut [7
] has argued, for example, that we still lack information on how the microbiota interacts with the host immune system, especially in the context of nutrition or environment. In this study, we assessed the role of dietary protein source (conventional versus soy-free), as well as rearing environment (caged versus cage-free) in shaping the gut microbiota community of laying chicken.
The gut microbiota in vertebrates is recognized as an essential functional site for metabolic and immune health. The gut microbiota is an assemblage comprising hundreds of microbial taxa and has a vital role in the feed metabolism [9
], stimulation of the immune system [11
], and competitive exclusion of pathogenic organisms [13
]. The digestive tract of chickens is short (total length of approximately 250 cm, 6–10× body length) compared to the digestive tracts of mammals (10–30 × body length), with a complete transit time of five to six hours. The shorter length of the chicken gastrointestinal tract allows for faster digestive processes and also microbiota composition that differ from other sections of the gastrointestinal track. The ceca, two blind sacs that serve as a site for fermentation and digestion, are considered crucial for avian health and immunity and are one of the best-studied foci for gut microbiota [16
]. It is well known that various dietary factors, including feed composition, antibiotic growth promoters and dietary supplements [10
] may alter the cecal microbiota composition. However, there is relatively limited knowledge about how such dietary factors interact with the rearing environment. For example, the bedding material or litter affects the gut microbiota composition [18
]. Understanding this interaction is vital for determining the extent to which diets can be used to modulate gut microbiota, and whether microbiota differences elicited by the rearing environment supersedes or is superseded by dietary changes. Furthermore, it is also necessary to determine if such differences are qualitative, quantitative, or both, in terms of the presence and abundance of microbial taxa.
1.1. Role of Environment in Microbiota Structure
The commercial poultry rearing environment has seen several changes over the past century. While historical (pre-1960’s) production systems relied on cage-free barns with outdoor access, streamlined commercial production systems since the 1960s [21
] have relied on battery cage production systems as the predominant model for egg production. However, changes in consumer awareness centered around animal welfare have induced major U.S. egg producers to shift to cage-free eggs by 2025. Cage-free and free-range housing typically offers more space per individual bird as well as access to the outdoors and exposure to natural lighting [22
]. These changes have posed various challenges to the egg producers, ranging from biosecurity to skeletal health, or gut health. One of the critical questions that deserves deeper understanding is the role of rearing environment on gut microbiota. A recent study by Cui et al. [23
], which characterized gut microbiota between caged and free-range hens, found differences in gut microbiota by age, but as the study depended on PCR-DGGE (Polymerase Chain Reaction - Denaturing Gradient Gel Electrophoresis) fingerprints, and not on sequence databases, it was not possible to assess taxon representation and abundance. Generally speaking, we would expect that the gut microbiota structure would differ between rearing environments, due to the differences in the litter or the soil with which the birds interact. For example, Torok et al. [20
] showed that broilers raised on different bedding materials (sawdust, shavings, paper, etc.) displayed different cecal microbiota communities, which in turn was suspected to influence growth and health. Therefore, it is necessary to understand the extent of differences in the gut microbiota of birds raised on the same diet, but in different housing systems.
1.2. Role of Dietary Protein Source
In the poultry industry, the cost of feed is the most expensive component of production, accounting for 60–70% of costs. In the U.S., most poultry diets utilize the lowest cost ingredients, typically corn and soy as the sources of energy and protein, respectively, to keep the cost of nutrition low [24
]. While soybean meal (SBM) represents over 40% of the poultry diet [25
], concerns about the anti-nutritive properties of SBM [26
], the high concentration of isoflavones on bird or human health [28
], have necessitated the search for alternative feed sources that may reduce or replace SBM in poultry diets. Climate change and increased demands from fisheries production are also expected to increase costs [29
]. In humans, it is well established that the source of dietary macronutrients (proteins, carbohydrates, lipids) significantly influences gut microbiota composition and function [31
]. While our knowledge of these phenomena in chicken lags behind that of our understanding in humans, poultry studies on diets with varying sources of medium-chain fatty acids [32
], or dietary fiber [33
] show that diets modulate intestinal microbiota in chicken. However, no published studies have simultaneously assessed the roles and interaction of diets and rearing environment on gut microbiome in chicken. Finally, antibiotic growth promoters (AGP) also influence the gut microbiome [34
], but with the ban on AGPs in the United States, it is necessary to characterize the extent to which diet and environment can be utilized as tools to modulate beneficial gut microbiota. Therefore, the motivation for the present study was to assess the differences in cecal microbiota arising from dietary protein source and housing environment.
In this study, we investigated the cecal microbiome of caged and cage-free Hy-line Brown laying hens reared on either a standard industrial soybean meal (SBM) based diet versus a diet that substituted cottonseed meal (CSM) for SBM. Our overall hypothesis was that the housing environment has a profound influence, compared to dietary protein source in altering the microbiota composition in chicken. Specifically, for the housing environment aspect, we expected that the cage-free environment would generate greater gut microbiota diversity irrespective of dietary protein source.
Sequencing of 16s amplicons generated a total of 261,093 raw reads, of which 220,891 reads remained after quality filtering, and were used for analysis using the Mothur protocol. After completion of all quality control steps (duplicate removal, chimera removal, removal of non-prokaryotic sequences), 219,188 sequences remained, with an average 9132 sequences per library (sample) and were used for operational taxonomic unit (OTU) classification and statistical analyses.
A total of 1497 unique OTUs were observed across the treatments, which were classified up to the genus level. Individual samples showed some variation in the number of taxa identified, ranging from 290 to 432 among libraries. Estimates of alpha diversity showed that overall, the cage-free groups (CFS and CFSF) had higher microbiota richness compared against caged treatments, irrespective of protein source (Figure 1
A). Dietary protein source was not a predictor of alpha diversity; although the median value was consistently higher for the soy-free group, the differences were not statistically significant (Figure 1
B). The microbiota communities in hens raised in cage-free environments showed higher alpha diversity (Figure 1
C), and except for Inverse Simpson index, the other indices (Chao1, ACE, Fisher) were statistically different between caged and cage-free environments (Kruskal-Wallis Chi-Square test, P < 0.005, df = 3). Pairwise comparisons showed that the alpha diversity estimates between CFS and CFSF treatments were not significantly different (Wilcoxon Sum Rank Test, P = 0.866).
Similarly, caged treatment groups (CS and CSF) were not different from each other (P > 0.2). All pairwise comparisons of cage-free groups against caged groups were statistically significant (Wilcoxon Sum Rank Test, P < 0.05, Benjamini-Hochberg adjusted for multiple comparisons). The CS treatment had the lowest alpha diversity of all experimental groups.
β-diversity, calculated using the permutational analysis of variance (PERMANOVA) using the diet x environment as grouping factor (CS, CSF, CFS, CFSF), was significantly different among the groups (F = 2.53, P = 0.001, df = 3). The weighted and unweighted UNIFRAC (Unique Fraction) analyses also confirmed these communities to be significantly different. These differences in communities were also observed using the Constrained Analysis of Principal Coordinates (CAP) ordination analysis, based on the Bray-Curtis distances (Figure 2
). Each diet x environment group clustered distinctly from each other. However, the ellipse encompassing the 95% normal multivariate distribution around the individual samples showed an overlap only between the CFS and CFSF groups.
When comparing across groups, we found that 697 OTUs were shared among the four groups. While the CS and CFSF did not have any unique taxa, the other two groups showed 29 (CSF) and 80 (CFS) OTUs found exclusively in those groups. However, there was considerable intra-group variation in the OTU distribution among individual samples, both at the family and the genus levels. Bacteroidaceae
were the most abundant families, whereas Bacteroides
was the most abundant taxon, followed by Lachnoclostridium
Next, we compared differential enrichment of taxa (linear discriminant analysis) using only the diet or environment as differentiating factors. We focused on these comparisons given that the alpha diversity indices did not show significant separation between the CFS and CFSF groups, or between the CS and CSF groups. While soy versus soy-free microbiota communities were not significantly different, we were interested in determining specific taxa that are differentially enriched. The Linear Discriminant Analysis (LDA) analysis showed 93 differentially enriched taxa between the soy and soy-free treatments (LDA > 2, P < 0.05), whereas the caged versus cage-free comparison showed 120 differentially enriched taxa. In the soy versus soy-free comparisons, 49 taxa were enriched in the soy diets, of which Faecalibacterium
(OTU0013), and Olsonella
were among the top enriched. In the soy-free treatment Bacteroides
, and Ruminoclostridum
were among the top enriched (Figure 4
A). In the caged versus cage-free comparison, the distribution of enriched taxa was more lopsided. The caged group showed 31 that were taxa significantly enriched (LDA > 2, P < 0.05), whereas the cage-free group had 91 taxa that were significantly enriched. In the caged group, Faecalibacterium
, and Bacteroides
(OTU0047) were among the top enriched, whereas in the cage-free group Ruminococcacceae
(OTU0039), and Pseudoflavonifractor
were the top enriched (Figure 4
Utilizing the differentially enriched taxa in hierarchical clustering analysis, we found that there was a high degree of intra-group variation in the soy-free versus soy comparison (Figure 5
A). However, the cage-free group clustered together, and also showed a more consistent pattern of taxon enrichment (Figure 5
B). This intra-group consistency was also observed in the caged treatments.