A carvacrol-based product reduces Campylobacter jejuni load and alters microbiota composition in the caeca of chickens
Funding information: The PhD fellowship was supported by the French National Association of Research and Technology (ANRT).
Abstract
Aim
This study was conducted to test the ability of a carvacrol-based formulation (Phodé, France) to decrease the C. jejuni caecal load in inoculated broiler chickens and to study the impact of the C. jejuni inoculation alone or combined with the product, on the caecal microbiota.
Methods and Results
On day 1, chickens were either fed a control feed or the same diet supplemented with a carvacrol-based product. On day 21, the carvacrol-supplemented chickens and half of the non-supplemented chickens were inoculated with C. jejuni (108 CFU). Quantitative PCR was used to quantify C. jejuni in chicken caecal samples and 16S rRNA gene sequencing was carried out at 25, 31 and 35 days of age. A significant decrease of 1.4 log of the C. jejuni caecal load was observed in 35-day-old chickens supplemented with the product, compared to the inoculated and unsupplemented group (p < 0.05). The inoculation with C. jejuni significantly increased the population richness, Shannon and Simpson diversity and altered beta-diversity. Compared to the control group, the C. jejuni inoculation causes significant changes in the microbiota. The carvacrol-based product associated with C. jejuni inoculation increased the diversity and strongly modified the structure of the microbial community. Functional analysis by 16S rRNA gene-based predictions further revealed that the product up-regulated the pathways involved in the antimicrobial synthesis, which could explain its shaping effect on the caecal microbiota.
Conclusions
Our study confirmed the impairment of the caecal bacterial community after inoculation and demonstrated the ability of the product to reduce the C. jejuni load in chickens. Further investigations are needed to better understand the mode of action of this product to promote the installation of a beneficial microbiota to its host.
Significance and Impact of the Study
Results suggested that this product could be promising to control C. jejuni contamination of broilers.
INTRODUCTION
Zoonoses are diseases transmitted from animals to humans by direct or indirect contact or through food consumption. In the European Union, Campylobacter spp. is the most important zoonotic bacteria involved in foodborne gastrointestinal diseases with more than 246,000 cases of campylobacteriosis in 2018 (European Food Safety Authority and European Centre for Disease Prevention and Control, 2019). Chicken meat contamination occurs mainly during evisceration at the slaughterhouse (Herman et al., 2003). The European Food Safety Authority (EFSA) estimated that 37.5% of fresh broiler meat were contaminated in the EU (European Food Safety Authority and European Centre for Disease Prevention and Control, 2019).
Campylobacter spp. are Gram-negative, microaerophilic, spirally curved bacteria that mainly live as commensal organisms in the gastrointestinal tract of birds (Silva et al., 2011; Vandamme et al., 2015). Chickens are natural hosts for Campylobacter species and especially C. jejuni (Williams et al., 2016). C. jejuni colonizes primarily the caeca, but also the large intestine and the cloaca reaching 106–109 CFU/g of contents (Dhillon et al., 2006; Meunier et al., 2016). Although C. jejuni is considered as a commensal, studies have shown that this bacterium can cross the epithelial intestinal barrier and disseminate to internal organs (e.g. liver and spleen) (Cox et al., 2009; Meade et al., 2009). Several researchers suggested an alteration in the intestinal physiology of chickens (modification of intestinal villi and cryptae), as well as an activation of the immune system (Awad et al., 2015; Humphrey et al., 2014; Lamb-Rosteski et al., 2008). However, the impact of C. jejuni colonization after inoculation or exposure with infected birds on the structure of the microbiota remains unclear and might be modulated by the genetic origin of the chickens (Connerton et al., 2018; Thibodeau et al., 2015). Unfortunately, previously cited studies limited their scope on C jejuni impairment of microbiota taxonomy and structure diversity. The study of the functional pathways of caecal microbiota together with its composition and diversity could provide a better understanding of the interactions of C. jejuni within the caecal microbial community and therefore health consequences.
In 2013, Romero-Barrios et al. proposed a mathematical model suggesting that a reduction of the C. jejuni load by two to three log at the farm level could reduce human infections by, respectively, 76%–90% (Romero-Barrios et al., 2013). The use of essential oil as bactericidal ingredients in chicken feed might be a promising strategy to reach that goal. Indeed, the antibacterial properties of essential oil compounds have been widely described against a large range of pathogenic bacteria (Du et al., 2015). The positive effect of carvacrol (2-methyl-5-[1-methylethyl]-phenol), a monoterpenic phenol component of the essential oils isolated from Origanum vulgare (oregano) and Thymus vulgare (thyme), against C. jejuni has been demonstrated in vitro (Anderson et al., 2009; Kollanoor Johny et al., 2010) and in vivo (Allaoua et al., 2018; Arsi et al., 2014; Szott et al., 2020). Due to their lipophilic properties, essential oil compounds such as carvacrol, easily cross the biological membranes (Zotti et al., 2013), resulting in an early absorption in the stomach and proximal small intestine (Michiels et al., 2008). Fixing carvacrol and thymol on silica was not sufficient to bring them to the distal part of the digestive tract of chickens, where C. jejuni grows (Du et al., 2015). An appropriate formulation is thus necessary to release the carvacrol in the caeca to inhibit C. jejuni development. We previously demonstrate in vitro the antimicrobial activity of a carvacrol-based formulation against C. jejuni and evidenced a delay in the release of this formulated carvacrol into the caeca (Allaoua et al., 2018). However, the effectiveness of this formulation to reduce the C. jejuni load in vivo remains to be established. Furthermore, previous studies reported the effect of oregano essential oil or thymol mixed with carvacrol (Bortoluzzi et al., 2017; Ruan et al., 2021; Zhu et al., 2019) on the chicken caecal microbiota, but to our knowledge, none mentioned the effect of carvacrol either alone, pure or formulated.
Thus, the aim of this study was to (i) test the ability of the carvacrol-based formulation to decrease the C. jejuni caecal load on broiler chickens after inoculation and (ii) study the impact of the inoculation with C. jejuni alone or combined with the carvacrol-based product, on the caecal microbiota structure and predicted metabolic functions.
MATERIALS AND METHODS
Birds and experimental design
The in vivo assay was conducted in accordance with the current guidelines of the directive 2010/63/EU of the European Parliament and of the Council, in the facilities of the UE-1277 Plateforme d’Infectiologie Experimentale (PFIE, INRAE, 2021. Infectiology of the farm, model and wild animal facility, Centre Val de Loire, Nouzilly, France). All the experimental procedures were approved by the Loire Valley ethical review board (CEEA VdL, committee number 19, n°02110.02). The experimental procedure is detailed in Figure 1.
One hundred male ROSS 308 broiler chickens of 1 day old were used and housed in the breeding rooms underregulated environment: temperature from 32°C for 1-day-old chicks to 25°C for 35-day-old chickens and relative humidity at 60%. The environment of the animals was enriched by placing resting mats and suspended metal plates.
On the first day of the experiment (D1), the absence of enteropathogenic Escherichia, Salmonella spp., Campylobacter spp. and Clostridium perfringens was checked by enumeration in the caecal content of 10 chickens with, respectively, BromoCresol Purple medium, Xylose-Lysine-Deoxycholate agar medium, Charcoal Cefoperazone Deoxycholate Agar and Glucose Yeast medium. Chickens were allocated to two groups: 50 chickens received a non-supplemented soybean meal-wheat-corn-based diet (Table 1) and the 40 others received the same commercial feed supplemented with 2.5 kg/ton of feed (0.25%) of a product formulated and provided by Phodé (France). The product (Phodé, France) is composed of 1%–2% of carvacrol (active ingredient), 3%–5% of surfactants (solubilizing agents), 40%–45% of monoglycerides (stabilizing agents), 4%–7% of water and 15%–25% cellulose (adsorbent carrier) and 20%–30% of silica (caking inhibitor). A previous pharmacokinetic study demonstrated that the carvacrol-based product was not absorbed in the upper part of the digestive tract but was released in the caeca and large intestine (Allaoua et al., 2018). On D14, the health status of three animals/group were checked for the absence of enteropathogenic Escherichia, Salmonella spp., Campylobacter spp. and Clostridium perfringens by enumeration in the caecal content.
| Ingredient (%) | |
|---|---|
| Corn | 28.01 |
| Wheat | 30 |
| Soybean meal | 34.2 |
| Soybean oil | 4 |
| Calcium carbonate | 0.94 |
| Dicalcium phosphate | 1.88 |
| Salt | 0.4 |
| Methionin | 0.17 |
| Vitamin + mineral + anticoccidial premix | 0.4 |
| Total | |
| Calculated nutritional composition (%) | |
| Crude protein | 21.34 |
| Lysin | 1.02 |
| Methionin | 0.43 |
| Fat | 5.96 |
| Calcium | 0.97 |
| Phosphore | 0.71 |
| AME (kcal/kg) | 2936 |
- Abbreviation: AME, apparent metabolizable energy.
The absence of enteropathogenic Escherichia, Salmonella spp., Campylobacter spp. and Clostridium perfringens was also checked by enumeration in 400 g of feed on the first day of the test. The feed intake and the bodyweight measurements were carried out at the beginning of the test (mean body weight on D1: 0.051 ± 0.004 kg) and on 21, 25, 31 and 35-day-old chickens. On D21, the chickens with the non-supplemented feed were separated into two groups, a group of non-inoculated and non-treated chickens (Control group n = 24) and one group with inoculated and non-supplemented chickens (Jejuni group n = 23). The chickens of the Jejuni and Jejuni + Product groups were orally inoculated with 1.108 CFU (equivalent dose to that found in the caeca of the broiler chickens) of C. jejuni per chicken (CJ-MDR1, a multidrug-resistant wild-type strain of C. jejuni, ANSES, France) on D21. Chickens were inoculated at 21 days because it is known that maternal antibodies stay in the chick’s blood circulation for 14–21 days post-hatch (King et al., 2010). It is only at this stage that the chicks are considered immunologically mature. Euthanasia were carried out using CO2 or by cervical dislocation depending on the age of the animals. Necropsies were performed on seven chickens of each group (n = 7) on D25, 31 and 35 (4, 10 and 14 days post-inoculation). During those autopsies, both caeca of each chicken were recovered, frozen in liquid nitrogen and then stored at −80°C. The remaining chickens were euthanized at the end of the procedure.
qPCR evaluation of the C. jejuni caecal load
Caeca were emptied and rinsed with 1X Phosphate Buffered Saline (PBS) solution prepared from 10X PBS (Invitrogen, Fisher Scientific). Then, DNA was extracted from 200 mg of caecal content, using the QiAmp Fast DNA Stool Mini Kit (Qiagen) according to the protocol of Josefsen et al. and the supplier’s instructions (Josefsen et al., 2015). The expression of the N-benzoylglycine amidohydrolase (hippuricase, hipO) gene has been used to differentiate strains of C. jejuni and C. coli. In order to specifically quantify C. jejuni, the primers and probe designed by LaGier et al. (2004) were chosen to amplify the hipO gene. The two primers were Cj-F1 TGCTAGTGAGGTTGCAAAAGAATT and Cj-R1 TCATTTCGCAAAAAATCCAAA, and the probe was Cj-probe [6FAM]ACGATGATTAAATTCACAATTTTTTTCGCCAAA[TAM]. QPCR reactions were performed on Get-TQ platform (Toulouse, France) on a 96-well plate on a StepOne Plus thermocycler (Thermo Fisher Scientific). The qPCR assays were carried out in a 12 μl volume using the TaqMan Fast Advanced Mastermix (Thermo Fisher Scientific, Life technologies). Each sample was treated individually, they were never pooled (n = 7/group/time). The qPCR reaction contained 1X TaqMan Fast Advanced Mastermix, the two primers (10 μmol L−1), the probe (250 μmol L−1) and 2 μl of template DNA. Thermal cycling conditions were as follows: one cycle at 95°C for 20 min, followed by 40 cycles at 95°C for 1 min and 60°C for 20 min.
A standard curve was first made using 1/10 serial dilutions of a 109 CFU/ml C. jejuni stock solution. After centrifugation, the pellets were collected to obtain a similar matrix as the samples. DNA was extracted as described above and quantified using qPCR.
16 rDNA amplicon sequencing
On each individual sample, the V3 and V4 hypervariable regions of the 16S rRNA gene were amplified with indexed and adaptor-linked universal primers (343F: 5’CTTTCCCTACACGACG CTCTTCCGATCTACGGRAGGCAGCAG and 784R: 5’GGAGTTCAGACGTGT GCTCTTCCGATCTTACCAGGGTATCTAATCCT). The amplicons were sequenced on Illumina MiSeq at the GeT PLAGE platform. Paired-end reads of 250 bp were obtained. The raw sequences were analysed using the Galaxy-supported FROGS pipeline (Escudié et al., 2018) to process the 3,041,045 16S ribosomal RNA gene amplicon sequences obtained. Amplicons without any ambiguous base, with a length between 400 and 500 nucleotides and matching with V3 and V4 proximal PCR primer sequences, were kept for clustering. Reads were clustered into OTUs (Operational Taxonomic Units) using the iterative growth process SWARM (Mahé et al., 2014). Chimaera was detected using VSEARCH (Rognes et al., 2016) and then discarded. The remaining OTUs were filtered to keep OTUs present at least in 2 samples. The taxonomic affiliation of the OTUs (Operational Taxonomic Units) was performed with the BLAST algorithm against the Silva 138 database (Quast et al., 2013). The mean number of reads per sample was 34,907 (min: 20,675 to max: 61,415). A phyloseq object (McMurdie & Holmes, 2013) containing the OTUs table (OTU abundance for each sample with taxonomic classification) as well as sample information was built for subsequent analysis. Diversity indexes (α-diversity), including Shannon and Inverse Simpson, were calculated with the phyloseq R package after the rarefaction of the OTU table at 20,675 sequences. PICRUSt2 analysis was used to predict the caecal microbial community functions with the unrarefied OTU abundance table as input (Douglas et al., 2020). Following PICRUSt2 authors’ recommendations, OTUs with a weighted Nearest Sequenced Taxon Index (NSTI) score higher than 0.15 were discarded due to low prediction quality. Relative predicted abundance of MetaCyc pathways and prediction based on Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa & Goto, 2000) were normalized by dividing the abundance of each pathway by the sum of all pathway abundances per sample.
Statistical analysis
Data analyses were conducted using the R software (release 4.0.3, The R foundation for statistical computing). Root square transformed taxonomic relative abundances at phylum, family and genus levels, diversity indexes (number of observed OTU, Shannon and InvSimpson indexes), log-transformed qPCR data and body weight and feed intake for each time point were analysed with ANOVA. Age, treatment and their interaction were considered as fixed effects. P values were adjusted using the Hochberg methods. When the main effect was significant, post hoc comparisons with a Tukey adjustment for multiple comparisons were made (R package emmeans). To compare the bacterial community structures and MetaCyc pathways resulting matric across all samples, a non-Metric Distance Scaling (nMDS) analysis was conducted with a Bray–Curtis dissimilarity index on the total sum scaling normalized matrix. A PERMANOVA (vegan package) was then used to test whether composition among groups was similar or not.
Data access
All raw sequences were deposited in the NCBI Sequence Read Archive (accession no. PRJNA736580).
RESULTS
Body weight of animals
All animals were in good health and no group showed body weight or feed intake reduction. Body weight of chickens is presented in Figure 2.
Efficacy of C. jejuni inoculation
Chickens were inoculated with 108 CFU of C. jejuni. In order to assess C. jejuni load in the control and inoculated chickens, a qPCR experiment using primer designed for specific amplification of hipO gene was performed. As expected, chickens from the Control group were exempt from C. jejuni throughout the test (Figure 3). There was no significant effect of the sampling age, but a significant impact on the group and a significant interaction between day and group (p < 0.05 after Bonferroni correction). Indeed, the carvacrol-based product significantly decreased by 1.4 log (p < 0.05) the C. jejuni caecal load after 35 days while no significant difference was observed at 25 and 31 days. On 31-day-old chickens, the carvacrol-based formulation induced a non-significant decrease of 0.77 log of the targeted bacteria load. The analysis of the relative abundance of the OTU affiliated to Campylobacter by sequencing the V3–V4 region of the 16S rRNA gene leads to the same conclusions (Figure 9). Of note, this latter genus and the Campylobacteraceae family were the unique taxa of the Campylobacterota phylum.
Diversity and structure dynamics of the caecal bacterial community
The α- and β-diversity of caecal microbiota
To evaluate the effect of C. jejuni inoculation and carvacrol-based product on caecal bacterial communities, we first assessed α-diversity using the richness index, evaluated as the number of observed OTUs, and Shannon and InvSimpson indices (Figure 4). As expected, in the Control group, the diversity indices were stable across the time of sampling except for a slight decrease of the Shannon index between D31 and D35. A significant interaction between age and group was observed. Compared with the Control group, the inoculation with C. jejuni significantly increased the population richness at 31 and 35 days and its diversity at 35 days according to the two indices.
Supplementation with carvacrol-based product in the feed of the inoculated animals increased the Shannon diversity of the caecal bacterial community from 25 days on compared to the inoculated group (Jejuni group). The same results were observed for InvSimpson, except at 35 days of age, where the differences between the inoculated animals receiving or not carvacrol-based product were not significant anymore. Finally, adding the carvacrol-based product to the feed appeared to stabilize the diversity with age in inoculated animals since no effect of age was observed for this group.
As a next step, we analysed the consequence of C. jejuni inoculation and the effect of the addition of the carvacrol-based product in the feed in inoculated animals on beta-diversity of the caecal bacterial community. Besides the age-related effect, the non-metric multidimensional scaling (nMDS) plot indicated that the community structure strongly differed between groups as highlighted by the clear separation of the three groups (Figure 5) and the results of the multilevel pairwise adonis (Figure S2). This separation was effective from 25 days of age. An age-related change was observed in the three groups (Figure 5) and the results of the multilevel pairwise adonis (Figure S2). This separation was effective from 25 days of age. An age-related change was observed in the three groups. The treatment with the carvacrol-based product in C. jejuni inoculated chickens led to the greatest change in community structure compared to the two other groups.
Core bacterial community
To assess the effect of our treatments on the core bacterial community, an Euler diagram (Larsson et al., 2020) was displayed at each age, to represent shared OTUs in each experimental group based on the OTUs present in at least 50% of the 63 individuals, that is representative of the core bacterial community (Figure 6). The addition of the carvacrol-based product in the feed of the inoculated animals lead to a specific microbial signature, as indicated by the majority of unique OTUs (81, 107 and 109 OTUs out of the 416 OTUs at, respectively, 25, 31 and 35 days of age).
Relative abundances of the dominant phyla
To better characterize the microbial signature of treatments on the chicken caecal community, we analysed the taxa relative abundance at the phylum, family and genus levels.
A total of 10 different phyla could be identified in the chicken caeca but only 3 had a relative abundance above 0.5%. As expected, Campylobacterota phylum was observed only in the two groups inoculated with C. jejuni (Jejuni and Jejuni + Product groups). The major phyla in the caecal content of the chickens were the Firmicutes and the Proteobacteria, whatever the age or the treatment (respectively between 70% and 95% and between 3% and 29%, Figure 7, Table S1). The age-related increase in Firmicutes at the expense of Proteobacteria in the Control group was not observed in the chickens inoculated with C. jejuni. The addition of the carvacrol-based product seemed to restore this age-related dynamic at least for the Firmicutes phylum.
Differentially abundant caecal bacterial families
There were 79 families present in the chicken caecal ecosystem of which 5 were classified as ‘unknown family’. However, only 13 families had a relative abundance of over 0.5% in at least one experimental group and were further considered for statistical analysis (Paës et al., 2020) (Figure 8a, Table S1). Compared to the Control group, the inoculation significantly halved the relative abundance of Enterobacteriaceae on 25 days chickens and doubled that of Lachnospiraceae and Ruminococcaceae at 35 days of age while the establishment of the Monoglobaceae family was observed from D31 (Figure 8b). The addition of the carvacrol-based product in the feed of the inoculated animals further decreased the Enterobacteriaceae abundance on D35. Treatment with the carvacrol-based product modified the Oscillospiraceae to Ruminococcaceae balance. In the Control and Jejuni groups, the Oscillospiraceae were present in greater quantity, whereas in the group that received the product the Ruminococcaceae predominated (Figure 8b). Interestingly, upon treatment with the carvacrol-based product, we observed also the establishment of the Butyricicoccaceae family in the inoculated animals with an abundance of up to 2%.
Differentially abundant caecal bacterial genera
16S RNA gene sequences analyses revealed a total of 206 genera in the chicken caecal ecosystem of which 8 had multi-affiliation and 38 were classified as unknown genus is given a BLAST identity threshold of 97%. Thirty-five identified genera had a relative abundance of over 0.5% in at least one experimental group and were further considered for statistical analysis (Paës et al., 2020) (Figure 9, Table S1). Compared to the control group, the inoculation decreased the Escherichia-Shigella proportion at 25 days of age, those of Rombustia at 25 and 31 days of age, and the Clostridioides proportion at 31 days. Paludicola genus was increased with inoculation from D25, Agatobacter genus was increased on D31 while Lactobacillus, Selimonas and Eisenbergiella were increased from D31. The supplementation of feed with the carvacrol-based product hindered the increase of Agatobacter and Eisenbergiella observed in the non-treated inoculated animals and further decreased the Escherichia-Shigella proportion. Moreover, compared to the two other groups, C. jejuni inoculation associated with formulated carvacrol supplementation led to a decrease in the relative abundance of Klebsiella and Salmonella and Oscillobacter. Conversely, the development of [Ruminococcus] torques group, Faecalibacterium, Subdoligranulum, Butyricicoccus, Blautia, Shuttleworthia, Tyzzerella and Lachnospiraceae GCA-900066575 group was favoured in the caecal microbiota by the supplementation of inoculated chickens with the carvacrol-based product.
Predicted functions of the caecal bacterial community
To evaluate whether the shift in the microbiota composition following the C. jejuni inoculation alone or coupled with the addition of the carvacrol-based product would affect the microbiota functions, we identified the predicted MetaCyc and KEGG pathways using PICRUSt2 (Douglas et al., 2020). The nMDS plot of the 376 identified pathways (Figure 10) highlighted a separation between groups. According to PERMANOVA pairwise comparisons (Figure S2), the separation between groups was effective from 25 days of age. A total of 171 out of the 376 MetaCyc pathways, revealed a significant effect on the group (p-adjusted <0.05) while 53 exhibited a significant age effect (Table S1). To get a general vision of metabolic potential shift according to the group, we aggregated MetaCyc pathways to predict level3 KEGG pathways. A total of 165 out of the 257 level3 KEGG pathways revealed significant differences (p-adjusted <0.05) with the group while only 17 exhibited a significant age effect (Table S1). Focusing on significant KEGG level3 pathways involved in degradation and biosynthesis pathways, we could observe two kinds of the response of microbiota metabolic potential to C. jejuni inoculation (Figure 11 and Table S1). The first metabolic response occurred only rapidly after inoculation (from the fourth day onwards, i.e. D25). The predicted relative abundance of genes involved in lysine and zeatin biosynthesis were increased, whereas those in penicillin and cephalosporin, carotenoid, flavone and flavonol biosynthesis, atrazine and fluorobenzoate degradation were decreased. A second profile corresponding to a later response was observed from 10 to 15 days after inoculation (D31–D35). It included the increase of predicted relative abundance of genes involved in the biosynthesis of folate, streptomycin, amino acids (phenylalanine, tyrosine and tryptophan) and the decrease of those involved in the degradation of other amino acids (lysine, valine, leucine and isoleucine), of monoterpen degradation (geraniol, limonene and pinene), of aromatic compound utilization (benzoate, aminobenzoate, ethylbenzene, nitrotoluene, chlorocyclohexane and chlorobenzene degradation) and carbon fixation pathways.
Interestingly, supplementation with the carvacrol-based product in animals inoculated with the C. jejuni led to the greatest changes in predicted pathway structure compared to the two other groups. The increased pathways involved antimicrobial synthesis (tetracycline, butirosin, neomycin, streptomycin, penicillin and cephalosporin, bacterial toxins), bacterial structural compound syntheses such as peptidoglycan, fatty acids and lysine biosynthesis. The down-regulated pathways involved amino acid degradation (lysine, valine, leucine and isoleucine), aromatic compound utilization (benzoate, aminobenzoate, ethylbenzene, nitrotoluene, chlorocyclohexane and chlorobenzene degradation), monoterpen degradation (geraniol, limonene and pinene) and carbon fixation pathways. Altogether, the difference in microbiota composition according to C. jejuni inoculation alone or in broiler chickens treated with the carvacrol-based product was associated with major modifications of the predicted metabolic functions.
DISCUSSION
C. jejuni is the leading cause of bacterial zoonosis responsible for gastroenteritis in humans. Contamination occurs mainly through the consumption of contaminated chicken meat by contact with digestive content at the time of evisceration. The aim of this study was first to evaluate the ability of a carvacrol-based product to decrease the C. jejuni caecal load in inoculated broiler chickens and second to study the impact of the C. jejuni inoculation alone or combined with the carvacrol-based product on the caecal microbiota. Although we did not test the carvacrol-based product alone, we demonstrated that the product was effective in reducing the C. jejuni load in the caeca of inoculated animals by 1.4 log after 35 days of treatment. These beneficial effects of the carvacrol-based product were associated with a shift of the gut microbiota composition and predicted metabolic function potentially leading to beneficial effects on the chicken host.
The chicken caecal microbiota
In chickens, the microbiota colonizes the gut intestinal tract from the hatch and it stabilizes around 3 weeks (Diaz Carrasco et al., 2019). In accordance with previous studies on chicken caecal microbiota, we confirmed the predominance of Firmicutes, while the Bacteroidota and Proteobacteria were present in lower abundances (Awad et al., 2016; Han et al., 2016; Hankel et al., 2019; Kumar et al., 2018; Wei et al., 2013). The most represented families in the control group were Lachnospiraceae, Enterobacteriaceae, Peptostreptococcaceae and an unknown family of the Clostridiales vadin BB60 group. According to the literature, the proportions of these families are subject to variations due to differences in diet, breeding environment and genetics.
Effect of the C. jejuni inoculation on the caecal microbiota
In accordance with previous studies (Sofka et al., 2015; Thibodeau et al., 2015), the C. jejuni inoculation resulted in an increase in the richness and α-diversity while β-diversity was altered in the caecal microbiota 4 days after inoculation. The increase of diversity was related to a disruption of taxonomic balance. Indeed, although C. jejuni are naturally hosted in the caeca of broiler chickens (Newell, 2002), their massive introduction through inoculation led to a decrease in the abundance of Firmicutes, as previously observed by Sofka et al. (2015). In the present study, we further observed important modifications within the Fimicutes phylum with an establishment of Monoglobaceae family member and dominance of the Lachnospiraceae at the expense of members of the Clostridia vadimBB60 group unknown family under inoculation. According to Awad et al. (2016), the positive relationship between the presence of Campylobacter and the Lachnospiraceae proportion may be explained by cross-feeding properties since Campylobacter are known to use the organic acid produced by Lachnospiraceae as an energy source. Sakaridis et al. (2018) found that caecal samples with higher Campylobacter loads had higher Enterobacteriaceae proportions, and on the contrary Sofka et al. (2015) and Connerton et al. (2018) showed a decrease in the Enterobacteriaceae count in Campylobacter-positive chickens. It seems that the effect of the presence of Campylobacter on Enterobacteriaceae proportion depends on the genera that are most commonly found in the study. Indeed, when Escherichia-Shigella is the main genus, the proportion of Enterobacteriaceae decreases at the same time as the number of Campylobacter increases in the chicken microbiota samples (as we observed at 25 and 31 days) (Connerton et al., 2018; Sofka et al., 2015). We hypothesize that the genera Campylobacter and Escherichia-Shigella might belong to the same niche and therefore must be in competition. Fifteen days after inoculation, an increase of Lactobacillus, Paludicola, Sellimonas and Eisenbergiella proportion was still observed. According to the literature, although Campylobacter inoculation resulted in a taxonomic rearrangement of the composition of the microbiota, no clear pattern is yet identifiable (Connerton et al., 2018; Kaakoush et al., 2014; Thibodeau et al., 2015). Besides altering the composition of caecal microbiota, we observed that C. jejuni inoculation led to either an early or late effect on the metabolic potential of the caecal bacterial community. To our knowledge, no data are available on C. jejuni impact on microbiota functionality. These predicted gene relative abundance evolution with time may reflect an adaptation of microbiota to the installation of the exogenous C. jejuni.
Effect of the carvacrol-based product on the caecal microbiota of inoculated animals
Our carvacrol-based product added to the feed was effective in reducing the load of C. jejuni in chicken ceca by 1.4 log. This is consistent with our previous pharmacokinetic study, demonstrating that the carvacrol-based product was not absorbed in the upper part of the digestive tract while it was released in the caeca and large intestine, the main sites for C. jejuni development (Allaoua et al., 2018). In addition to the beneficial effect on reducing C. jejuni abundance, the product decreased the abundance of potentially pathogenic bacteria such as Escherichia-Shigella, Salmonella and Klebsiella. An effective action of Origanum vulgare essential oil and carvacrol against enteropathogens like Salmonella and Klebsiella were demonstrated in vitro (Fournomiti et al., 2015; Mellencamp et al., 2011) and in vivo with a supplemented feed (Bauer et al., 2019). Carvacrol has been shown to increase the sensitivity of Salmonella enterica serovar Typhimurium to antibiotics (Kollanoor Johny et al., 2010). We demonstrated here the same effects when the carvacrol-based product was administered through the feed. Additionally, in chickens inoculated with C. jejuni, the carvacrol-based treatment restored the proportion of Firmicutes relative abundance and peculiarly the relative abundance of its genera Eisenbergiella and Paludicola to levels similar to those found in the control group.
Interestingly, the addition of the carvacrol-based product in the feed of the animals inoculated with C jejuni increased the diversity and strongly altered the structure of the bacterial community, leading to a specific microbial signature with a majority of OTUs found exclusively in this group. The differences in microbiota structure can be explained by (i) the exclusive presence in this group of Faecalibacterium, Shuttleworthia, Subdoligranulum and Tyzzerella and (ii) the increase of Blautia and Butyricicoccus and the reduced proportion of Oscillobacter compared to the inoculated chicken that did not receive the carvacrol-based product. Faecalibacterium and Subdoligranulum genera are well known to be butyrate producers which appear to be deleterious for Campylobacter growth (Ahsan et al., 2016; Cresci et al., 2017). Indeed, C. jejuni is able to sense the microbiota-derived butyrate via the BumSR to-component signal transduction system (Goodman et al., 2020). A bactericidal effect of butyrate on C. jejuni has been demonstrated in vitro (Van Deun et al., 2008) while a negative correlation between caecal butyrate content and C. jejuni load has been reported in broiler in vivo (Hankel et al., 2019). Altogether this could explain the negative correlation observed in our study between Campylobacter and Faecalibacterium/Subdoligranulum proportion. Additionally, Subdoligranulum also produces lactic, succinic and acetic acids that appear to have a bacteriostatic effect on enteropathogens (Bjerrum et al., 2006). Faecalibacterium and Shuttleworthia are also considered beneficial bacteria because they were associated with a body weight increase of chickens (Lee et al., 2017).
In our study, the observed shift in the composition of the caecal microbiota induced by the carvacrol-based product in chickens inoculated with C. jejuni was associated with important alteration of gut microbial predicted metabolism and functions. It could partly be explained by the microbiota functional enrichment in pathways involved in the biosynthesis of antimicrobial synthesis and bacteriocin (i.e. bacterial toxins) as previously observed in suckling piglets supplemented with benzoic acid and essential oils (Zhai et al., 2020). Besides support of gut barrier integrity, immune activation and nutrient competition, secretion of antimicrobial products allow the gut microbiota with colonization resistance (Ducarmon et al., 2019). Altogether, these mechanisms might explain both the resistance to C. jejuni development, the associated decrease of enteropathogens (Sassone-Corsi et al., 2016) and the microbiota structure specificity induced by the carvacrol-based product in chickens inoculated with C. jejuni. In inoculated chickens, the carvacrol-based product down-regulated the predicted degradation pathways of lysine, valine, leucine and isoleucine while up-regulated biosynthesis of lysine, phenylalanine, tyrosine and tryptophan. In accordance, Li et al. observed in piglets group supplemented with a mix of carvacrol and thymol, an increased concentration of several amino acids such as valine, isoleucine and alanine (Li et al., 2018). Knowing that the amino acid and protein composition and content was similar between the three groups (same diet), these results suggest that the carvacrol-based treatment in inoculated chickens could contribute to microbiota amino acid sparing in favour of protein biosynthesis. The monoterpen (geraniol, limonene and pinene) degradation pathways were down-regulated in inoculated and carvacrol-based product treated animals. Taking into account the bacteriostatic and bactericidal activities of carvacrol resulting from its monoterpenic structure (Suntres et al., 2015), our results might suggest tolerance of the selected bacterial community to essential oil compounds.
We can conclude that the carvacrol-based product provided to chickens from hatching is efficient to limit the development of C. jejuni after inoculation. Furthermore, our specific formulation allows the carvacrol to reach the caeca where it might select beneficial bacteria and promote the establishment of a beneficial bacterial community. Besides the direct bactericidal action of carvacrol, predicted metabolism and function analysis based on 16S rRNA gene abundance has allowed us to propose an indirect pathway of action of the carvacrol-based product via the selection of an adapted bacterial population producing antimicrobial substances that confer resistance to colonization. Further studies are needed to confirm the first results observed with PICRUSt2 on predicted pathways through whole microbiota genome shotgun sequencing and thorough analysis with and without Campylobacter colonization. In particular, this would provide a better understanding of the ability of certain microbiota to cope with the presence of carvacrol and identify new antimicrobial molecules active against C. jejuni and derived from chicken microbiota.
ACKNOWLEDGEMENTS
The authors are grateful to the members of the scientific and animal staff of the Plateforme d’Infectiologie Expérimentale (PFIE, INRAE, 2021. Infectiology of farm, model and wild animal’s facility, (INRAE, 2021) https://doi.org/10.15454/1.5572352821559333e12), UE-1277 PFIE, INRAE Centre Val de Loire, Nouzilly, France, especially to the study manager Mickaël Riou and the zootechnicians in charge of this project: Sylvain Breton, Alexis Pléau and Guillaume Martin. The authors are grateful to the Genotoul bioinformatics platform Toulouse Midi-Pyrenees and the Sigenae group for providing computing and storage resources thanks to Galaxy instance http://sigenae-workbench.toulouse.inra.fr.
CONFLICT OF INTEREST
This work received financial support from Phode. Elsa Bonnafé, Michel Treilhou and Sylvie Combes received investigator-initiated research support from Phode.
