Introduction

Infections caused by gastrointestinal nematodes (GIN) are one of the most important diseases in grazing ruminants in temperate regions of the world1,2,3. The economic importance of these infections is related to reduced weight gain, milk and meat production and reproduction; it was estimated that GIN infection could have cost to European Union-28 €2.1 bn in 20184,5. However, in order to determine if microbiota influences GIN infection control, in this study we have described the composition and diversity of the microbiome in Churra breed sheep classified as resistant or susceptible to the infection by the GIN Teladorsagia circumcincta. T. circumcincta is the most prevalent nematode species present not only in the area where ewes were selected for the present study but also in many other temperate areas of the world6,7,8. For that, the microbiome presents in gastric mucosa and gastric content from 12 ewes experimentally infected with T. circumcincta have been sequenced and characterized. Besides, prevalence have suffer un increase since 1990 in the region this study was carried on9. GIN infections have been regularly controlled with anthelmintic drugs, mainly with bezimidazoles due to their good quality-cost relationship, combined with pasture management8,10. But, these drugs were incorrectly applied along the years -overused, misused, or applied incorrectly- causing the appearance of anthelmintic resistance (AR) and favoring the survival of parasites with resistant genotype11. Since in 1960s the first report about benzimidazole resistance was published, many farms has been informed as resistant; besides, last year farms with multi-resistant parasites are becoming increasing common12,13,14,15. This situation has generated the need to change the approach for GIN control with new strategies such as the development of vaccines against helminths, biological control with anthelmintic active compounds from plants and fungus, pasture management or the development of breeding programs for worm resistance11,16,17,18.

Under this circumstance, the interest in selecting animals with resistant phenotype to GIN infection, especially sheep, has increased within last years19 defined resistance as the ability of a host to suppress the establishment and/or subsequent development of a roundworm infection. Therefore, according to this concept, authors described resistant phenotype as those host that present low worm burden despite grazing pastures contaminated with infective larva. The most frequent method to measure the parasite burden is counting the number of worm eggs in faeces, being expressed as the faecal egg counts (FEC)20.

Besides, it has been shown that there are sheep breeds that are more resistant to GINs infection than other breeds, such as Canaria Hair, Barbados Blackbelly and Red Maasai, but at the same time within a particular breed there are individuals that are more resistant than others, such as within the Churra breed19,21,22,23. Genetic resistance to GIN have been associated with a protective immune response that is mediated by humoral (titre of antibodies, mainly IgA and IgE) and cellular response (eosinophils, T cells, globule leukocytes or mast cells)23,24,25,26. However, these studies have not considered an important component of the host that is in direct contact with gastrointestinal parasites and therefore could influence the worm burden, the gastrointestinal microbiota.

All mammals are colonized by wide diversity of microorganisms that live in symbiosis in the mucosal surfaces of the host 27,28. Microbiota produces a beneficial relationship host-bacteria, providing nutrients, metabolizing them and defending against colonization by pathogens29,30. However, occasionally harmful organisms, such as GIN, colonize the gastrointestinal tract provoking damages whose severity depends to the burden of the infection, the localization or the parasite specie, among others31. As GIN and bacteria share the habitat, it is expected that interact among them and with the host. Although it is known that parasitic infections are associated with profound changes in the structure and function of the host gut microbiome, in veterinary medicine this knowledge is limited and most of the studies have been conducted in human and rodent models27,31,32,33.

Understanding the composition of microbiome in presence of GIN is fundamental to develop a better knowledge about the pathogenesis of the infection. However, in order to determine if microbiota influences GIN infection control, in this study we have described the composition and diversity of the microbiome in Churra breed sheep classified as resistant or susceptible to the infection by the GIN T. circumcincta. For that, the microbiome present in gastric mucosa and gastric content from 12 ewes experimentally infected with T. circumcincta have been sequenced and characterized.

Results

Total bacteria DNA

The number of copies in gastric mucosa was 100-1300 and between 840,000–9,000,000 for gastric content without significant differences between groups.

Taxonomic profile analysis and microbiota

The sequencing of the V3-V4 region of the 16S rRNA gene performed for the 12 sheep gastric mucosa samples generated an average of 1,169,865,8 raw reads, while the gastric content generated a total of 231,472.6 raw. After removing host genome contamination, we retained an average of 413,853.9 sequences for abomasal mucosa and 174,316.8 sequences for gastric content. Retained reads were used for the subsequent analyses. Then the DADA2 analysis performed for the 12 ewes analyzed in this work identified 11,217 ASVs for abomasal mucosa and 8,514 ASVs for gastric content. The sampling depth was set to 1500.

In total, 23 phyla were identified in both types of samples, 1 from Archaea domain and 22 from Bacteria domain. The most abundant phyla in gastric mucosa, representing approximately 96% of the total, were Bacteroidetes (48.23%), Firmicutes (29.60%), Actinobacteria (9.27%), Verrucomicrobia (2.72%), Proteobacteria (2.50%), Fibrobacteres (2.26%), and Spirochaetes (2.20%). No significant differences were found in phyla abundances between resistant and susceptible groups in these gastric mucosa samples. While, in gastric content the most abundant phyla, which accounted approximately 94%,were Bacteroidetes (43.15%), Firmicutes (28.60%), Actinobacteria (14.13%), Fibrobacteres (2.89%), Spirochaetes (2.35%), Kiritimatiellaeota (2.06%) and Proteobacteria (1.38%). Differences approaching significance (p = 0.09) were found in the gastric content for Actinobacteria phylum between resistant and susceptible animals, showing an abundance of 10.2% in RG and 18.2% in SG.

The most abundant genus in gastric mucosa was Prevotella with accounted 15%, followed by Rikenellaceae RC9 (8.96%) and Aeriscardovia (8.70%) among other. In gastric content, Prevotella, with 14% of abundance, was followed by Aeriscardovia (13.42%) and Rikenellaceae RC9 (8.17%). Regarding the differences in genus abundance between groups, Serratia spp genus showed differences approaching significance (p = 0.07; RG 0.12% and SG 0.041%) in gastric mucosa. In gastric content, significant differences were found for Clostridum sensu stricto-1 (p = 0.05; RG 1.29% and SG 0.069%) and close to significance for Turicibacter (p = 0.07; RG 0.31% and SG 0.027%). The Spearman correlation coefficient between cFEC measured at the end of the first infection and the abundance of C. sensu stricto-1 (r= − 0.537; p= 0.08) in gastric content showed an approaching significance negative correlation; no correlations whereas found for other species between Serratia spp and cFEC (Table 1, Figs. 1, 2, 3).

Table 1 Principal taxa in gastric mucosa and gastric content samples between RG and SG in T. circumcincta sheep.infected.
Figure 1
figure 1

Relative percentage of abundance for the main taxa for phylum and genus level in resistant and susceptible groups.

Figure 2
figure 2

Boxplot showing significant or slight significant differences for RG and SG in gastric content (Clostridium sensu stricto 1; p = 0.08 and Turicibacter; p = 0.07), and gastric mucosa (Serrata spp ; p = 0.05). Significant differences are indicated by ** (p < 0.05) and slight significant differences by * (p between 0.05–0.01).

Figure 3
figure 3

Slight significant negative correlation between cFEC at the end of the first infection and Clostridium sensu stricto 1 abundance. Rho =− 0.537 and p = 0.08.

Alpha and Beta diversity

No significant differences between resistant and susceptible animals were observed in alpha diversity by any of the estimators used in the study. However, the gastric mucosa presented p values lower than gastric content (Tables 2, 3 and Fig. 4) (Chao index: p = 0.260; Shannon index: p = 0.243; and Simpson index: p = 0.271). Beta-analysis showed a trend to clustering for resistant animal samples from gastric mucosa; however no trend was detected for gastric content.

Table 2 Estimation of evenness and richness (Chao, Shannon and Simpson indexes) in gastric mucosa and gastric content samples during T. circumcincta infection. No significant differences between resistant and susceptible animals were observed.
Table 3 p values between RG and SG for each estimator and each sample.
Figure 4
figure 4

Estimation of richness (Shannon and Chao indexes) and evenness (Simpson index) in gastric mucosa (A) and gastric content (B) in RG (green) and SG (pink) experimentally infected with T. circumcincta. Not significance differences detected between groups.

The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Discussion

This study characterizes for first time the microbiome composition in gastric mucosa and content of Churra breed ewes with resistant and susceptible phenotype to the infection by T. circumcincta. The aim was to determinate if the microbiota could influence the resistant phenotype to the infection by T. circumcincta in Churra sheep. Animals were classified as resistant or susceptible according to their cFEC during a first experimental infection but also confirmed by IgA levels against L4 T. circumcincta by23. In that study, these resistant ewes showed higher IgA levels in serum at day 3 post-infection (p < 0.05) and close to significance at day 21 pi (p = 0.06); moreover, a strong negative correlation between cFEC and specific IgA was only significant in resistant ewes at day 3 pi (r =− 0.870; p < 0.05), but absent in susceptible ones. Several studies in different breed sheep have reported the association between high levels of specific IgA against the GIN and lower worm burden and FEC21,34.

Microbial diversity is evaluated using α-diversity indices such a richness or abundance of species from each community. Dissimilarities between communities are studied using β-diversities based on their ordination. In our study, neither gastric content microbiome nor gastric mucosa microbiome showed significant differences between groups in any α-diversity indices studied. On the contrary of our data35, reported significant differences in various microbial alpha-diversity indices between infected and non-infected group at 7 days post infection (dpi) by Haemonchus contortus. These differences could be due to all the animals of our study were infected36 studied abomasal microbiota composition in naive and immune calves infected with Ostertagia ostertagi to described and understand mechanism related with protective immunity. The authors suggested that unlike naive animals, the O. Ostertagi infection in immune cattle induced a minimal disruption in the abomasal microbiota and this may contribute to the development of long-term protective immunity.

In recent years, several studies have focused on the impact that GINs produced on host’s microbiome composition and how affects the parasite establishment and control of the infection comparing non-infected and infected animals37,38,39. It have been reported that GINs infections involve a change in the structure in the host digestive microbiome, inducing different physiological changes depending on the parasite specie or host29,37,38,40,41,42. The microbiome composition of abomasum from non-infected sheep is mainly represented by Bacteroidetes (≈65%) and Firmicutes (≈25%) phyla, but 7 days post-infection with H. contortus, Bacteroidetes increased (≈71%) and Firmicutes decreased (≈18%)35. In our study, Bacteroidetes (RG 46%; SG 49%) abundance in gastric mucosa was higher than Firmicutes (RG 33%; SG 25%) for both groups. Nevertheles, SG showed higher Bacteroidetes: Firmicutes ratio although differences were not statistically significant, probably because all animals were infected, as stated above43 sequenced gastric content from Merino sheep naturally infected with H. contortus field strains reporting higher relative abundance in Firmicutes (45%) than Bacteroidetes (26%). These results differ with most of studies and our own results, where Bacteroidetes (RG 46%; SG 40%) was more abundant than Firmicutes (RG 30%; SG 26%). Authors explained these variations in microbiome composition by factors that affect microbiome, as parasite burden, breed type, diets and different environmental condition. Besides, the relative abundance of Actinobacteria phylum in gastric content, which represented the 14.13% of the total bacteria, showed slight differences between resistant (10.2%) and susceptible (18.2%) sheep.

At genus level, Prevotella has been described as the most affected by GINs infections caused by H. contortus, Trichostrongylus colubriformis and T. circumcincta in sheep and goats. An increase in relative abundance in obligate anaerobes taxa, as Prevotellaceae family in lambs, sheep and goats infected with H. contortus, T. colubriformis and T. circumcincta was detected in infected animals compared with non-infected animals44,45,46,47. It could be explain by the Prevotellarole in protein degradation and energy host metabolism compensation44,48. Although in this study Prevotella was the most abundant genus, 15% in gastric mucosa and 14% in gastric content, no differences were shown between resistant and susceptible sheep in none sample presumably because all animals were infected.

Butyrate is a short-chain fatty acid (SCFA) that is formed during the microbial fermentation of dietary fiber of ruminants. This metabolite is present in low concentrations, and it seems to be involved not only in nutrition, but also as a potent inhibitor of intestinal inflammation47detected a decrease in metabolic pathway genes related to butyrate after an infection with H. contortus and T. circumcincta in lambs. Besides, this decrease in butyrate metabolism was following by the abundance reduction in some butyrate-producers bacteria species. On the other hand44, supposed that nematode infection modulates the gut butyrate biosynthesis by altering the abundance of butyrate-producing bacteria and they detected significant differences between non-infected and H. contortus infected goats in the relative abundance of the genus Butyrivibrio in rumen, which is a bacteria butyrate producer. In our study, differences were not found in Butyrivibrio genus in gastric content, but we found significant differences others butyrate-producers bacteria as C.sensu stricto-1 (RG 1.29% and SG 0.069%) and close to the significance limit with Turicibacter (RG 0.31% and SG 0.027%) between RG and SG in gastric content, been RG which account higher percentage49. Besides, this microorganism showed an approaching significant negative correlation with cFEC (r= − 0.537; p= 0.08). This data support the hypothesis that ewes with lower FEC have higher abundance of C. sensu strict-1.

The production of natural compounds with nematicidal activity synthesized by microorganisms is being a new focus in GINs infections chemical control investigation50 demonstrated that Serratia spp produces volatile compounds with 100% in vitro nematicidal activity against plant nematodes51 tasted in vitro isolated chitinases produced by Serratia sp. against H. contortus L3 obtaining 100% of larvicida activity, presumably because nematodes cuticle and eggs is constituted by chitin, been eggs who have higher levels52. Serratia spp was identified in gastric mucosa in our study, being the resistant group (0.12%) who had a higher abundance in comparison with the susceptible group (0.041%). This data may support the hypothesis that Serratia spp produces nematicidal compounds that collaborate in the control of the infection.

As a conclusion, our results suggest that resistant o susceptible phenotype to T. circumcincta infection could influence the microbiome composition, modifying the interaction with the host and in the last instance affecting the individual risk.

Methods

Ethical approval

All procedures involving animals in this study was performed in accordance to Spanish regulations regarding the protection of animals used for experimental and other scientific proposes (Royal Decree 53/2013), under the supervision of the Ethical and Animal Welfare Committee of University of León to after the approval of the competent body, Junta de Castilla y León.

All methods are reported in accordance with ARRIVE guidelines.

Animals and experimental design

The description of this section was previously published by53. Briefly, a total of 18 adult ewes (6–8 years old) belonging to a Churra breed flock were selected after measuring the number of eggs per gram (epg) in faeces in 109 grazing animals naturally infected with GIN. Those animals with the highest and lowest FEC values were selected for subsequent deworming and experimental infection with a single dose of 50,000 T. circumcincta L3. Thirty days after this first infection, ewes with the lowest and highest cumulative FEC (cFEC), 6 resistant (mean cFEC: 308 ± 338 epg) and 6 susceptible (mean cFEC: 5594 ± 2661 epg) to the infection, were selected. The individual data related to cFEC are shown in supplementary material (Table S1)Then, the same ewes were infected again but in this case with a single dose of 70,000 T. circumcincta L3; at day 7 pi, all animals were humanly euthanized for the collection of the samples. At that moment of the infection, the resistant ewes had a L4 burden 68% lower than susceptible ones.

Gastric content and abomasal tissue recovery

After sheep necropsy, the omaso and pylorus were tied using suture thread and immediately the abomasum was removed from all sheep. Abomasums were opened along the curvature and the inner surface was washed with tap water. : Both gastric contents and abomasal portions were immediately frozen in liquid nitrogen at the sampling site and then frozen at − 20 °C until use35.

Microbial DNA extraction

Microbial DNA was extracted from abomasal gastric mucosa and gastric content from each animal. Abomasum portions were scraped to obtain the gastric mucosa sample using a sterile slide without excessive pressure keeping the samples on ice to avoid DNA degradation. Microbial genomic DNA was extracted using Purelink Genomic DNA mini-kit (Invitrogen; REF K182000, Spain)38.

Gastric content was lyophilized and homogenised. Then genomic DNA was extracted using QIAampPowerFecal Pro DNA Kit (Qiagen; REF 51,805, Germany)54.

Both kits were used in accordance with manufacturer’s instructions. After microbial DNA extraction, DNA was quantified using Nanodrop® ND-1000 Spectrophotometer.

Total bacteria DNA

Total bacteria DNA was measured in all samples by quantitative real time PCR using forward primer (5’-GTG STG CAY GGY TGT CGT CA-3’) and reverse primer (5’- ACG TCR TCC MCA CCT TCC TC-3’) to calculate the number of copies in each samples, as previously described by55.

After extraction and quantification, samples were sent to amplify 16S rRNA hypervariable V3–V4 region. The sequencing was carried out by Teagasc Sequencing Centre (Moorepark, Fermoy, Ireland) service using 2 × 301 bp paired-end sequencing with Illumina MiSeq platform (Illumina, San Diego, CA, USA).

Bioinformatic processing and statistical analysis

After quality control, the sequencing raw data was aligned against the sheep reference genome (Oar_rambouillet_v1.0,https://www.ensembl.org/Ovis_aries_rambouillet/Info/Index) to remove host DNA sequences. The retained sequences from the Fastq file were filtered and trimmed to 280 (forward) and 210 bp (reverse) using the filter and Trimm function of the DADA2 package56. The paired reads were assembled into Amplicon Sequences Variants (ASV) and their taxonomic assignment was performed using the SILVA nr v.138 database57. Richness analyses were performed in R V4.1. ASVs and variables (phenotype and type of sample) were included in the estimation of alpha diversity index (Chao1 Rarefied Species, Shannon’s Diversity index and Simpson Dominance index) using Phyloseq package from R. Normality was check using Shapiro–Wilk test and the homogenized of the variance was tested with Levene’s test. The differences between groups were estimated using Kruskal–Wallis. Beta diversity was plotted using Non-linear Multi-Dimensional Scaling (NMDS) to explore the dissimilarities between pairs of samples using Bray–Curtis dissimilarity index, and Unweighted Unifrac index using Vegan package from R software. The Vegan envfit function was used to evaluate if the factors of study (phenotype and sample type) where associated to the NMDS ordinations; the significance of the fitted factors was estimated by using 999 permutations.

Relative abundances were calculated for each sample. Normality was check using Shapiro–Wilk test. Then, differences between groups were estimated using U-Mann–Whitney. Correlation between cFEC levels and relative abundance of bacterial species was measured by Spearman coefficient. The level of significance was determinate at p < 0.05 and p values between 0.05–0.1 were considered approaching significance.