Abstract
The goal of this study was to understand the composition and existence of the resident uterine microbiome in healthy mares and to establish the presence of a core microbiome for the healthy equine uterus. We analyzed the microbiomes of 35 healthy mares that are long-time residents of three farms in Oklahoma, Louisiana, and Australia as well as that of 19 mares purchased from scattered owners in the Southern Mid-Western states of the United States. Over 6 million paired-end reads of the V4 region of the 16S rRNA gene were obtained resulting in 19,542 unique Amplicon Sequence Variants (ASVs). ASVs were assigned to 17 known phyla and 213 known genera. Most abundant genera across all animals were Pseudomonas (27%) followed by Lonsdalea (8%), Lactobacillus (7.5%), Escherichia/Shigella (4.5%), and Prevotella (3%). Oklahoma and Louisiana samples were dominated by Pseudomonas (75%). Lonsdalea (28%) was the most abundant genus in the Australian samples but was not found in any other region. Microbial diversity, richness, and evenness of the equine uterine microbiome is largely dependent on the geographical location of the animal. However, we observed a core uterine microbiome consisting of Lactobacillus, Escherichia/Shigella, Streptococcus, Blautia, Staphylococcus, Klebsiella, Acinetobacter, and Peptoanaerobacter.
Similar content being viewed by others
Introduction
For decades the standard dogma was that the uterus was maintained as a sterile environment despite the vagina being colonized with commensal bacteria1. Even though culture-based studies laid out the foundation of our understanding of the reproductive tract microbiota as far back as three decades ago, there was an indication that culture-based methodologies in aiming to identifying microbial populations may underestimate diversity and overestimate the role of culturable bacteria2. Focusing on culturable bacteria, often the minority members of microbial communities, enhances the risk of missing and therapeutically disrupting those microbes that are more abundant3,4,5. Furthermore, the general consensus now is that culture-based technologies detects less than 10% of the resident microbial community in a sample6.
With the advent of 16S rRNA gene-based bacterial detection and identification techniques, the sterile womb paradigm has been challenged through fluorescence in situ hybridization (FISH) with 16S rRNA targeted probes in pregnant and non-pregnant women7. Different microbial communities with variable richness and diversity have been found among other species in distinct areas of the reproductive tract8. A recent study on the resident microbiota of the canine vagina (cranial) and endometrium revealed that, they both are home to a diverse microbiome; more specifically, a distinct dissimilarity exists in the structure and diversity of the endometrial microbiome in comparison to the microbiome of the cranial vagina9. During another study based on 16S rRNA gene amplicon sequencing of bovine, ovine, and porcine vaginal microbiomes were found to be distinctly different from how they were described using culture-based techniques10. Therefore, it is clear that the traditional culture-based techniques lack in the ability to capture the inherent diversity of the native microbial ecosystems that prevail within the mammalian reproductive tract.
The clinical literature continues to describe the uterus of a normal (healthy) mare as a site that does not host a resident microbial community11. However, previous studies based on in vitro culturing indicate that, a variety of microorganisms may be present in the uterus of a clinically healthy mare12. However, failure to remove pathogenic bacteria, sperm and other inflammatory substances may cause post breeding endometritis in mares13. Furthermore, several bacterial genera such as Streptococcus, Escherichia and Staphylococcus, working as opportunistic pathogens in the equine uterus, are known to cause pregnancy loss and infertility in mares due to infectious bacterial endometritis14,15.
The importance of having a comprehensive understanding of the resident microbiota of the reproductive tract of a healthy mare, or any domestic animal species, lies in the need to understand the role played by these commensal microorganisms and how dysbiosis may impact the overall health of the animal. Furthermore, decoding the composition of the uterine microbiome plays a vital role in comprehending the effect of broad spectrum antibiotics on shifting the overall composition of the resident microbiota possibly playing important roles in the reproductive process and fertility16.
We began our metagenomics investigations in the mare to document the underestimation of microbial diversity with common culture methods relative to the presence of microbial DNA using the approach over 10 years ago17. Subsequently, Schnobrich et al.18 described Next Generation Sequencing (NGS), culture, and cytology results from 29 mares with suspected endometritis. The main problem with these analyses were the dependence on culture and sequencing depth.
In this study, we attempted to comprehensively analyze the resident microbiome of 54 clinically healthy mares from four different geographical locations in two continents and attempted to establish a stringent core microbiome of the healthy equine uterus. To the best of our knowledge, this is the first extensive study based on 16S rRNA gene amplicon sequencing, conducted to gain a comprehensive understanding about the uterine microbiome of the healthy mare.
Results
Sequencing results and data analysis statistics
Sequence analysis of 16S rDNA amplicons resulted in 6,486,441 paired-end reads from 56 samples at 115,829 ± 47,952 reads per sample ranging from 32,222 to 147,965. After removal of low-quality and ambiguous reads, 5,868,329 sequence reads were retained for downstream analysis. From these sequence reads, a further 3,784,490 was retained after sequence variant inferring, sequence read merging and chimera removal, which gave rise to 19,776 unique amplicon sequence variants. About 96% of the inferred sequence variants were retained during chimera removal, thus indicating that initial data cleansing steps were successful. Subsampling based on least number sequences for a sample was performed on all samples to normalize data and remove sequence bias before any of the downstream processing.
A taxonomy rarefaction curve indicated that the sequencing depth used in this study was sufficient to saturate species richness in all samples (Fig. 1).
Alpha diversity
Alpha diversity estimates for species richness and evenness were calculated for geographically grouped samples. Chao1 and Observed ASV metrices were used to calculate species richness, while Shannon and Simpsons Indices were used to calculate evenness19. Between group comparisons for Chao1 and Observed ASV metrices were performed using ANOVA, while Kruskal–Wallis test was used for Shannon and Simpson indices20. All four alpha diversity measures indicated that the uterine microbiomes of mares from four different geographical regions were significantly different in species richness and evenness (p < 0.05) (Fig. 2, Table 1). Australian samples were the highest in species richness, while Oklahoma samples were leading in species evenness.
To further ascertain the source of this intergroup difference in alpha diversity, pair-wise comparisons were conducted using Tukeys HSD and Dunn Test with B–H correction21 on species richness and evenness metrices respectively (Table 2). The pair-wise comparisons between the Dispersed group and all other groups were significantly different (p-adj < 0.05) in both species’ richness and evenness. Pairwise comparisons between Oklahoma and Australia deemed not significant for both species richness and evenness (p-adj > 0.05), while both Louisiana–Australia and Oklahoma–Louisiana comparisons revealed to be statistically significant in terms of species richness (p-adj < 0.05).
Beta diversity
Beta diversity values for all samples were calculated at genus level using the Bray–Curtis dissimilarity matrix and plotted using NMDS (Non-metric Multidimensional Scaling) ordination method available in MicrobiomeAnalyst22. Samples formed clear clusters based on their geographic origin (Fig. 3). Those originating from Oklahoma and Louisiana formed close, overlapping but distinct clusters, whereas samples from the Dispersed group and Australian animals clustered clearly apart. Permutational Multivariate Analysis of Variance (PERMANOVA) on the beta diversity values23 revealed that the distinct clustering visible on NMDS plots are statistically significant (F-value: 23.894; R-squared: 0.5891; p-value < 0.001), thus confirming that there is a significant difference in the composition or the microbial community structure recovered from endometrial lavage samples of animals from the Dispersed group, Oklahoma, Louisiana and Australia.
Hierarchical clustering
A dendrogram depicting the hierarchical clustering of the samples was drawn at genus level using Bray–Curtis index as the distance measure and Ward method as the clustering algorithm24 (Fig. 4).
Similar to the beta-diversity plots, three clear clusters were observed in the dendrogram as well. Australian and Dispersed samples (other than sample A616) formed clear separate clusters far apart, while the Oklahoma and Louisiana samples grouped together to form most of the third cluster. However, the separation between Louisiana and Oklahoma groups was less apparent when compared to the beta diversity plots.
Microbial community structure
Sequence analysis of 16S rDNA amplicons resulted in 6,486,441 paired-end reads. Subsequent removal of low-quality, ambiguous, and non-bacterial sequences resulted in 3,716,157 sequence reads. These resulted in 19,542 unique amplicon sequence variants (ASVs) that were used to study the composition of the equine endometrial microbiome. These 19,542 ASVs were assigned to 17 known phyla (Supplementary Table S1) and 213 known genera (Supplementary Table S2) by DADA2. Proteobacteria (~ 48%) was the most abundant phylum, followed by Firmicutes (30%), Bacteroidetes (12%), Actinobacteria (5%), Tenericutes (2%), and Kiritimatiellaeota (1%). These 6 phyla accounted for 98% of total abundance. Proteobacteria, Firmicutes, Bacteroidetes and Actinobacteria and Kiritimatiellaeota were observed across all samples. Tenericutes were only observed in samples from Oklahoma and Louisiana. Firmicutes were the most abundant phylum (52%) in Louisiana samples, while both Proteobacteria (36%) and Firmicutes (36%) emerged as the top two equally abundant phyla in Oklahoma. Following the overall trend, samples from both Australia (40%) and Dispersed (80%) were dominated by Proteobacteria.
At genus level, Pseudomonas emerged as the most abundant (27%) genus cumulative across all samples. The rest of the top 5 genera were Lonsdalea (8%), Lactobacillus (7.5%), Escherichia/Shigella (4.5%) and Prevotella_9 (3%), amounting to 51% of the total abundance. The most abundant genus in Australian samples was Lonsdalea (28%). Interestingly, Lonsdalea was not detected in any of the other samples. Samples from Oklahoma and Louisiana were dominated by Lactobacillus (14%) and Escherichia/Shigella (13%). These samples also had higher species diversity than Australian or Diverse groups. Out of the 19,542 ASVs analyzed, 9334 ASVs were not assigned a taxonomic classification at genus level. However, the vast majority of these had a taxonomic classification at Phylum level or below, only 371 ASVs remained unclassified at phylum level.
The Core-microbiome
The core microbiome was defined as genera present in all samples at an abundance of at least 0.1%25. Eligible genera from each group were overlapped using a Venn diagram26. A collection of eight genera consisting of Lactobacillus, Escherichia/Shigella, Streptococcus, Blautia, Staphylococcus, Klebsiella, Acinetobacter, and Peptoanaerobacter were present in all samples and was designated as the core microbiome of healthy equine uterus.
Phylogeny of the normal mare microbiome
Out of the 104 ASVs used for this analysis, one, was not assigned with a classification at genus level but belonged to order Clostridiales. In the phylogenetic tree illustrated in Fig. 5, this was observed to correctly cluster alongside other Clostridiales. Lonsdalea, previously unknown to inhabit the equine reproductive tract, and the most abundant genus in the Australian samples, was observed to accurately cluster alongside Escherichia/Shigella which share the same taxonomic family. This further confirms the accuracy of assignment to amplicon sequence variants and their subsequent taxonomic classification.
Differentially abundant taxa
LEfSe or linear discriminant analysis (LDA) effect size, an algorithm for class comparison of high dimensional data27 and DESeq2, a method utilized for differential analysis of count data28, both implemented in Microbiome Analyst was used to identify differentially abundant taxa, or biomarkers, based on the geographic origin of the endometrial lavage samples. DESeq2 was able to identify 92 significant features based on false discovery rate (FDR < 0.05), and LEfSe identified 194 (adjusted FDR < 0.05, LDA > 2.0) significant features.
A consensus list of 92 significant features was arrived upon (Supplementary Table S3) based on the overlapping features from the two independent lists derived above19. Pseudomonas, Lonsdalea, Lactobacillus, Escherichia/Shigella, and Psychrobacter emerged as the top 5 biomarkers with a LDA score > 5.0. Pseudomonas, Lonsdalea, and Lactobacillus were significant in the Dispersed, Australian and Louisiana samples respectively, while both Escherichia/Shigella, and Psychrobacter appeared to be significant in Oklahoma samples (Fig. 6).
Discussion
The mammalian uterus has long been considered a sterile environment to sustain life1. However, this paradigm was challenged via some seminal studies in humans7,29. Since that time, the presence of a rich and dynamic microbiome in the mammalian endometrium was established in several species9,23,30,31,32,33. We and others have previously established that the uteri of clinically healthy mares host a resident microbiome12,17,34. The importance in understanding this potential shift in accepted dogma is underscored by the fact that many pregnancy related complications in humans35, and other mammals, especially those in important livestock species36,37, are well known to be of bacterial origin. Equine endometritis occurs in 25–60% of breeding mares38,39 resulting in infertility, postpartum metritis as well as septicemia and fatality in newborn foals40. Historically, genera such as Staphylococcus, Escherichia, Pseudomonas, and Klebsiella were considered to be invasive pathogens indicating dysbiosis in mares41. However, this and our previous studies indicate that these four genera are common members of healthy mare uterine microbiome underscoring the importance of understanding the structure and diversity of healthy uterine microbiome of the mare and to define a “core microbiome”.
The alpha and beta diversity of bacterial communities recovered from equine uteri in this study indicate that the species diversity, richness and evenness of the healthy mare uterine microbiome defer significantly based on the geographical origin of the animal (Fig. 2, Table 1). The pairwise comparisons of alpha diversity indices indicate that microbiome of animals listed as Dispersed is significantly different in species richness and evenness from those of Oklahoma, Louisiana and Australian animals.
Similarly, beta diversity analyses using NMDS plots (Fig. 3) show that Oklahoma and Louisiana samples cluster together and the Dispersed group, as expected, forms a separate cluster with increased species diversity. The Australian samples, that were farthest away geographically, formed a tight cluster away from the other three groups. The clustering makes sense as Oklahoma, Louisiana, and Australian animals were long-term residents of respective university farms and were expected to cluster together. Dispersed animals were purchased from an Oklahoma horse purveyor who collects animals from a dispersed area in Southern Midwestern states of United States. As expected, they were the most diverse, yet they were closer to US samples than Australian samples. Australian animals clustered together away from other animals. Hierarchical clustering (Fig. 4) also depicts three clear clusters with Oklahoma and Louisiana samples clustering together, dispersed samples mostly clustering on their own and the Australian samples forming their own distinct cluster.
We generated 6,486,441 paired-end sequences in this study that yielded 19,542 unique ASVs. They were assigned to 17 known phyla and 213 known genera. Consistent with other mammalian uterine microbiomes9,23, the equine uterus was dominated by Proteobacteria (48%), Firmicutes (30%), and Bacteroidetes (12%). As expected, there were geographical differences in the phyla present in samples. Tenericutes were only observed in horses from the university farms in Oklahoma and Louisiana. They were not observed in dispersed animals, many of whom would have been from Oklahoma. This is surprising due to the fact that they were one of the top 6 phyla in the study, but their inconsistency in reproductive tract metagenomes have been observed before42.
At the genus level Pseudomonas was the most abundant genus (27%). This is consistent with our previous work where we identified Pseudomonas as the most abundant genus in the canine endometrium as well9. The prevalence of Pseudomonas in the present study was largely driven by it dominating the uteri of animals from the Dispersed group (Fig. 7). One could speculate the presence to differences in management of those animals prior to our acquisition. Pseudomonas aeruginosa is thought to be an opportunistic pathogen that is responsible for many instances of equine endometritis43,44. Indirect evidence suggests that organisms belonging to the genus Pseudomonas in our study consists of several species. We are unable to reliably assign species with the SILVA database we used for taxonomic classification45. However, we are in the process of isolating and fully characterizing the Pseudomonas spp. identified in equine uteri.
Species composition of Australian samples revealed an interesting trend. The dominant genus in many samples were Lonsdalea. In 5/14 samples Lonsdalea represented over 60% of the organisms identified. In another two samples they exceeded 20%. This was an intriguing finding as Lonsdalea has not previously been reported in mammals. It is a relatively new bacterial genus that previously belonged to the genus Brenneria. Lonsdalea (as well as Brenneria) is a known phytopathogen that belong to Enterobacteriaceae family and is a pathogen of poplar and oak trees46. Lonsdalea quercina has been isolated from the surfaces of flight membranes of lesser horseshoe bats in the Czech Republic47. Lonsdalea was not detected at all in any of the other locations. However, organisms belonging to the genus Erwinia was identified in 5 animals from the Oklahoma group and one animal from the Louisiana group with a groupwise relative abundance of 0.94%. This is interesting as Lonsdalea and Brenneria were formerly classified as Erwinia prior to recent taxonomic reclassification47. Erwinia and Brenneria were also isolated from the gut microbiome of redbanded stink bugs (Piezodorus guildinii)48. We considered the possibility of Lonsdalea being an environmental contaminant in the Australian samples. However, negative controls including those for lavage medium and equipment failed to detect any Lonsdalea. The large contribution of this one genus to Australian samples, and its presence in majority of the animals tested is intriguing and rules out a transient environmental infection. As our Australian samples came from one semi-closed herd in one location, further investigation is required before the significance of Lonsdalea in equine uteri could be determined. It is also possible that Lonsdalea found in Australia and Erwinia found in Oklahoma and Louisiana belong to a hitherto unclassified organism(s) belonging to the same superfamily.
One of our major goals in this study was to identify and establish a core microbiome for the healthy mare uterus. We used a comparatively strict criterion of a groupwise prevalence of 100% and a 0.1% groupwise abundance25. We identified a core microbiome constituting of the following 8 genera in the mare endometrium. They were Lactobacillus, Escherichia/Shigella, Streptococcus, Blautia, Staphylococcus, Klebsiella, Acinetobacter, and Peptoanaerobacter (Fig. 8).
Lactobacillus (7.5%) had the third highest relative abundance across all samples (after Pseudomonas and Lonsdalea). Lactobacillus has long been observed as a dominant commensal organism in the human vagina and is known to play a role in gestational length, parturition and maintaining a healthy vaginal microbiome49. Furthermore, Lactobacilli are reported as a dominant member of the human uterine microbiome as well with implications in implantation success50,51. Previous studies in the mare reported that the presence of Lactobacillus in the equine uterus was low compared to humans and was reported at < 2% in abundance52. In this study, we found Lactobacillus to be the most prevalent genus in the core microbiome with an abundance of 7.5%. Escherichia/Shigella, Streptococcus and Klebsiella have previously been reported from uterine lavages of mares suspected of endometritis in a previous study18. Blautia and Acinetobacter have not been reported from the equine uterus to date, however, they have been reported as abundant in the human uterus50 and in the human gut microbiome53. Peptoanaerobacter has not been reported from reproductive tracts before but has been identified associated with human periodontal diseases54.
We were cognizant of the debate surrounding the placental microbiome and the role of reagent contamination55,56. Unlike the placenta, the uterine microbiome has a reasonable biomass. However, we collected negative controls for equipment and reagents at each stage of analysis to account for any reagent contamination and removed all detected sequences from further analysis57.
From this study, we conclude that the healthy equine uterus harbors a rich and diverse microbiome. We have illustrated the structure and composition of this microbiome with detailed analysis of endometrial microbiomes obtained from animals from four distinct geographical regions in two continents. We acknowledge that differences in climate and feed could have contributed to some of the variations we have identified. We bridged a research gap that existed on the commensal uterine microbiome of a healthy mare. The total composition of the microbiome varies with the geographical region while still maintaining a distinct core microbiome across all animals tested. We do realize that only a limited number of geographical locations were assayed in this study. We have established a platform for future research on elucidating the uterine microbiome at species level using full-length 16S rDNA sequencing and establishing the functional attributes of the microbiome utilizing whole metagenome sequencing58,59,60. We hope that other researchers would build on this study and further refine the healthy equine uterine microbiome. Furthermore, this paves the way for developing microbes as pre- or probiotics to enhance and maintain the natural balance of the uterine microbiome, thus improving the overall reproductive health of the mare.
Methods
Animals
Fifty-four geographically dispersed and clinically mature healthy mares between 4 to 18 years of age were used for the study. Out of the 54 animals, 9, 12, and 14 were located on university farms located in Stillwater, Oklahoma, Baton Rouge, Louisiana, and Wagga Wagga, NSW, Australia, respectively. These mares were long-term residents of their respective farms. They are labeled Oklahoma, Louisiana, and Australia in this study. The remaining 19 mares were purchased from an Oklahoma horse purveyor who is a source to regional veterinary practices for embryo recipient mares and collects animals from a wide multi-state area in the Southern Midwestern states of the US. They are labeled “Dispersed” in the study. These mares were quarantined for 4 weeks upon arrival at the Oklahoma State University farm and were allowed to acclimatize on pasture for 2 + months before samples were collected. Dispersed mares and Oklahoma mares shared the same facility and identical management conditions for multiple months although they never co-mingled in the same pasture. All mares used in the study were anatomically and clinically normal with no history of reproductive problems. The mares were given a complete pre-study breeding soundness exam so that the study included only historically reproductively normal mares with no clinical evidence of endometritis. Mares used in this study were of different breeds. All animal studies were approved by the Institutional Animal Care and Use Committees of Oklahoma State University (VM-17-25 and IACUC 20-50), Louisiana State University (#17-046), and Charles Sturt University (A18086). This study is reported here in accordance with ARRIVE guidelines61.
Small volume lavage (SVL)
Following a thorough cleaning of the perineum and external genitalia, uterine luminal endometrial samples were collected using saline infused through an aseptic protected double-guarded system. To avoid contact with the vestibule and cranial vagina the first protective sleeve containing a double-guarded system was carried to the external os of the cervix where it was brought into contact with the cervix. A second tube (36 Fr. uterine lavage catheter) containing the sample catheter was inserted well into the cervix and the sample catheter (stallion urethral catheter) was then introduced deep into body of the uterus. 150 ml of saline was infused through the sample catheter, allowed 30 s of contact within the lumen prior to efflux collection via gravity flow and or aspiration. The initial 50 ml were not saved to ensure the sample collected had been within the uterine lumen. The samples were immediately placed on ice and transported to the laboratory for processing. This technique was proven to be sufficient for uterine microbial sampling in previous studies62,63. Control samples consisting of the saline flushed through the protected double-guarded system, including the same lubricants used for sample collection from both the US and Australia were collected and analyzed alongside the uterine samples. SVL samples from Louisiana and Australia were shipped to Oklahoma State University for further analysis.
DNA isolation
All DNA extractions used in this study were performed in the same laboratory by a single investigator using identical reagents and laboratory conditions. Fifty mL of SVL obtained for each sample was spun down at 4000 rpm for 30 min at 4 °C. The pellet was then re-suspended in 1.5 ml of the same solution. Two hundred fifty µl of the resuspension was used for total DNA extraction using QiAamp DNA mini kit® (Qiagen, Germantown, MD) based on manufacturer’s instructions. DNA quality and quantity was measured using a NanoDrop® ND-1000 spectrophotometer. ZymoBIOMICS® Microbial Community Standard (Zymobiomics, Irvine, CA Cat. D6300) containing a known composition of microbiota was used as the positive control and molecular grade water was used instead of the sample in the negative control. The microbial community standard was used as per manufacturer’s instructions. All samples including the negative and positive controls were subjected to sequence analysis.
PCR amplification and DNA sequencing
The PCR amplification of extracted DNA and 16S amplicon sequencing was performed by Novogene Corporation Inc, Sacramento, CA. The V4 region of the 16S ribosomal RNA gene (~ 250 bp) was amplified using the specific primer pair 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′-GGAC TACHVGGGTWTCTAAT-3′), with the forward primer carrying the barcode sequence. The Phusion® High-Fidelity PCR Master Mix (New England Biolabs) was used to carry out all PCR reactions.
The PCR products were purified by running them on a 2% agarose gel and extracting them using a Qiagen® Gel Extraction Kit. The sequencing library preparation was done using a NEBNext® UltraTM® DNA Library Prep Kit for Illumina. It was quality tested using Qubit 3.0 fluorometer and quantified using Q-PCR. Subsequently, an Illumina HiSeq 2500 platform was utilized for the sequencing of the library, generating ~ 250 bp paired-end raw sequence reads.
Negative controls
Negative controls were collected from equipment used in SVL collection, laboratory reagents etc. and was subjected to DNA extraction and subsequent sequence analysis. Only one sample for laboratory reagents produced detectable DNA after amplification and was sequenced. Resulting reads were deleted from the subsequent analyses using microDecon57, a subtraction tool, prior to statistical analyses.
Data analysis
The computational analysis of 16S V4 amplicon sequence data was conducted using DADA2 version 1.8, according to the SOP for Illumina data64. All DADA2 codes were run using R-Studio version 1.1.4 on R version 3.6.1. Briefly, the quality scores of forward and reverse sequence reads were plotted to identify any drastic drops in sequence quality, specifically towards the ends of the sequences, and no such significant drops were observed. Hence, no specific truncation of the ends of sequences were done during the pre-processing step. However, any sequences that had ambiguous bases and an expected error greater than 2 were removed using the ‘filterAndTrim’ command in DADA2. Subsequently, the error rates for the amplicon sequence data were modeled using ‘learnErrors’function in DADA2 and the sequence reads were then dereplicated using the function derepFastq. The dereplicated sequences were then used for the inference of sequence variants for each sample using the core sample inference algorithms available in DADA2, implemented in the function ‘dada’. Using the mergePairs function, the corresponding forward and reverse inferred sequence variants were then merged to arrive at the full de-noised sequence contigs. The ‘makeSequenceTable’ function was run on these contigs to create the amplicon sequence variant (ASV) table. The ASV table was then subjected to chimera removal and taxonomic classification via removeBimeraDenovo and assignTaxonomy functions. The DADA formatted version of the Silva version 132 reference database45 was used as the training dataset for taxonomic classification. All sequences of non-bacterial origin (i.e., chloroplasts, mitochondria, Archaea and Eukaryotes) were removed. The ASV table and the taxonomy file generated by DADA2 was then used for statistical analysis and data visualization, using MicrobiomeAnalyst22. Multiple sequence alignment and phylogenetic analysis of representative ASVs were conducted using MEGA765.
Data filtering and normalization
Low abundance features, defined as those with less than 10 read counts in at least 20% of the samples were removed prior to data normalization. Rarefying of data based on minimum library size, and Total Sum Scaling was performed as data normalization steps to negate the effect of uneven sequencing depth22. The resulting filtered and normalized data were used for subsequent statistical analyses and data visualizations.
Phylogenetic analysis
A total of 104 ASVs along with their taxonomic classifications were used to construct a phylogenetic tree to depict the evolutionary relationships between some of the most abundant bacterial genera found in the current study. ASVs with a relative abundance > 0.05% were selected and when several qualifying ASVs were assigned with the same genus, the representative ASV with the highest relative abundance was chosen. Multiple sequence alignment and phylogenetic tree building (via MUSCLE) was done using algorithms implemented in MEGA765.
Ethical approval
All methods were performed in accordance with the relevant guidelines and regulations of respective institutions. Animal Use and Care were covered under protocols VM-17-25 and IACUC 20-50 (Oklahoma State University), #17-046 (Louisiana State University), and A18086 (Charles Sturt University).
Data availability
The raw sequencing data used in this study is available at the Sequence Read Archive (SRA) of NCBI under the accession number PRJNA784331.
References
Ansbacher, R., Boyson, W. A. & Morris, J. A. Sterility of the uterine cavity. Am. J. Obstet. Gynecol. 99, 394–396. https://doi.org/10.1016/S0002-9378(16)34549-5 (1967).
Whitman, W. B., Coleman, D. C. & Wiebe, W. J. Prokaryotes: The unseen majority. Proc. Natl. Acad. Sci. USA 95, 6578–6583. https://doi.org/10.1073/pnas.95.12.6578 (1998).
Shade, A. et al. Culturing captures members of the soil rare biosphere. Environ. Microbiol. 14, 2247–2252. https://doi.org/10.1111/j.1462-2920.2012.02817.x (2012).
Handelsman, J. Metagenomics: Application of genomics to uncultured microorganisms (vol 68, pg 669, 2004). Microbiol. Mol. Biol. Rev. 69, 195–195. https://doi.org/10.1128/mmbr.69.1.195.2005 (2005).
Silva, E. N. et al. Amoxicillin-induced gut dysbiosis influences estrous cycle in mice and cytokine expression in the ovary and the caecum. Am. J. Reprod. Immunol. https://doi.org/10.1111/aji.13247 (2020).
Theron, J. & Cloete, T. E. Molecular techniques for determining microbial diversity and community structure in natural environments. Crit. Rev. Microbiol. 26, 37–57 (2000).
Swidsinski, A. et al. Presence of a polymicrobial endometrial biofilm in patients with bacterial vaginosis. PLoS One https://doi.org/10.1371/journal.pone.0053997 (2013).
Heil, B. A., Paccamonti, D. L. & Sones, J. L. Role for the mammalian female reproductive tract microbiome in pregnancy outcomes. Physiol. Genom. 51, 390–399. https://doi.org/10.1152/physiolgenomics.00045.2019 (2019).
Lyman, C. C. et al. Canine endometrial and vaginal microbiomes reveal distinct and complex ecosystems. PLoS One https://doi.org/10.1371/journal.pone.0210157 (2019).
Swartz, J. D. et al. Characterization of the vaginal microbiota of ewes and cows reveals a unique microbiota with low levels of lactobacilli and near-neutral pH. Front. Vet. Sci. 1, 19. https://doi.org/10.3389/fvets.2014.00019 (2014).
Ferris, R. A. et al. Model of chronic equine endometritis involving a Pseudomonas aeruginosa biofilm. Infect. Immun. https://doi.org/10.1128/IAI.00332-17 (2017).
Davis, H. A., Stanton, M. B., Thungrat, K. & Boothe, D. M. Uterine bacterial isolates from mares and their resistance to antimicrobials: 8296 cases (2003–2008). J. Am. Vet. Med. Assoc. 242, 977–983. https://doi.org/10.2460/javma.242.7.977 (2013).
Troedsson, M. H. Mating-induced endometritis: Physiology or pathology?. Vet. J. 199, 9–10. https://doi.org/10.1016/j.tvjl.2013.10.012 (2014).
Rebordão, M. R. et al. Physiopathologic mechanisms involved in mare endometrosis. Reprod. Domest. Anim. 49, 82–87. https://doi.org/10.1111/rda.12397 (2014).
Ferris, R. A. Endometritis: Diagnostic tools for infectious endometritis. Vet. Clin. N. Am. Equine Pract. 32, 481–498. https://doi.org/10.1016/j.cveq.2016.08.001 (2016).
Moreno, I. & Simon, C. Deciphering the effect of reproductive tract microbiota on human reproduction. Reprod. Med. Biol. 18, 40–50. https://doi.org/10.1002/rmb2.12249 (2019).
Rock, K. S. et al. Detectable differences in the endometrial microbiome between normal and susceptible mares using metagenomic profiling and conventional bacterial culture. Clin. Theriogenol. 3, 376 (2011).
Schnobrich, M. R., Bradecamp, E. A. & Scoggin, C. F. Next generation DNA sequencing, culture and cytology results in 29 mares with suspected endometritis. Clin. Theriogenol. 10, 338 (2018).
Ocejo, M., Oporto, B. & Hurtado, A. 16S rRNA amplicon sequencing characterization of caecal microbiome composition of broilers and free-range slow-growing chickens throughout their productive lifespan. Sci. Rep. https://doi.org/10.1038/s41598-019-39323-x (2019).
Clemmons, B. A. et al. Vaginal and uterine bacterial communities in postpartum lactating cows. Front. Microbiol. https://doi.org/10.3389/fmicb.2017.01047 (2017).
Mangiafico, S. S. An R Companion for the Handbook of Biological Statistics. https://rcompanion.org/rcompanion/d_06.html (2015).
Dhariwal, A. et al. MicrobiomeAnalyst: A web-based tool for comprehensive statistical, visual and meta-analysis of microbiome data. Nucleic Acids Res. 45, W180–W188. https://doi.org/10.1093/nar/gkx295 (2017).
Moore, S. G., Ericsson, A. C., Poock, S. E., Melendez, P. & Lucy, M. C. Hot topic: 16S rRNA gene sequencing reveals the microbiome of the virgin and pregnant bovine uterus. J. Dairy Sci. 100, 4953–4960. https://doi.org/10.3168/jds.2017-12592 (2017).
Yang, D. & Xu, W. Clustering in human microbiome sequencing data: A distance-based unsupervised learning model. Microorganisms 8, 1–18. https://doi.org/10.3390/microorganisms8101612 (2020).
Risely, A. Applying the core microbiome to understand host–microbe systems. J. Anim. Ecol. 89, 1549–1558. https://doi.org/10.1111/1365-2656.13229 (2020).
Heberle, H., Meirelles, G., da Silva, F. R., Telles, G. P. & Minghim, R. InteractiVenn: A web-based tool for the analysis of sets through Venn diagrams. BMC Bioinform. https://doi.org/10.1186/s12859-015-0611-3 (2015).
Segata, N. et al. Metagenomic biomarker discovery and explanation. Genome Biol. 12, R60 (2011).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. https://doi.org/10.1186/s13059-014-0550-8 (2014).
Wassenaar, T. M. & Panigrahi, P. Is a foetus developing in a sterile environment?. Lett. Appl. Microbiol. 59, 572–579. https://doi.org/10.1111/lam.12334 (2014).
Karstrup, C. C. et al. Presence and localization of bacteria in the bovine endometrium postpartum using fluorescence in situ hybridization. Theriogenology 92, 167–175. https://doi.org/10.1016/j.theriogenology.2017.01.026 (2017).
Franasiak, J. M. & Scott, R. T. Reproductive tract microbiome in assisted reproductive technologies. Fertil. Steril. 104, 1364–1371. https://doi.org/10.1016/j.fertnstert.2015.10.012 (2015).
Verstraelen, H. et al. Characterisation of the human uterine microbiome in non-pregnant women through deep sequencing of the V1–2 region of the 16S rRNA gene. PeerJ https://doi.org/10.7717/peerj.1602 (2016).
Yang, X. et al. The normal vaginal and uterine bacterial microbiome in giant pandas (Ailuropoda melanoleuca). Microbiol. Res. 199, 1–9. https://doi.org/10.1016/j.micres.2017.01.003 (2017).
Holyoak, G. R., Lyman, C. C., Wieneke, X. & DeSilva, U. The equine endometrial microbiome. Clin. Theriogenol. 10, 273–277 (2018).
Pellati, D. et al. Genital tract infections and infertility. Eur. J. Obstet. Gynecol. Reprod. Biol. 140, 3–11. https://doi.org/10.1016/j.ejogrb.2008.03.009 (2008).
Hong, C. B. et al. Etiology and pathology of equine placentitis. J. Vet. Diagn. Investig. 5, 56–63. https://doi.org/10.1177/104063879300500113 (1993).
Johnson, C. T., Lupson, G. R. & Lawrence, K. E. The bovine placentome in bacterial and mycotic abortions. Vet. Rec. 134, 263–266. https://doi.org/10.1136/vr.134.11.263 (1994).
Hurtgen, J. P. Pathogenesis and treatment of endometritis in the mare: A review. Theriogenology 66, 560–566. https://doi.org/10.1016/j.theriogenology.2006.04.006 (2006).
Leblanc, M. & Causey, R. Clinical and subclinical endometritis in the mare: Both threats to fertility. Reprod. Domest. Anim. 44, 10–22. https://doi.org/10.1111/j.1439-0531.2009.01485.x (2009).
Martabano House, A. Septiciemia in Foals. https://thehorse.com/125261/septicemia-in-foals/ (2008).
Benko, T. et al. Incidence of bacterial pathogens in equine uterine swabs, their antibiotic resistance patterns, and selected reproductive indices in English thoroughbred mares during the foal heat cycle. Vet. Med. 60, 613–620. https://doi.org/10.17221/8529-VETMED (2015).
Ong, C. T. et al. Interrogating the bovine reproductive tract metagenomes using culture-independent approaches: A systematic review. Anim. Microbiome 3, 41. https://doi.org/10.1186/s42523-021-00106-3 (2021).
Allen, J. L., Begg, A. P. & Browning, G. F. Outbreak of equine endometritis caused by a genotypically identical strain of Pseudomonas aeruginosa. J. Vet. Diagn. Investig. 23, 1236–1239. https://doi.org/10.1177/1040638711425589 (2011).
Anzai, T., Kamada, M. & Kanemaru, T. Serotypes and drug susceptibility of Pseudomonas aeruginosa isolated from mares with metritis. Bull. Equine Res. Inst. 1991, 12–20 (1991).
Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res. 41, D590–D596 (2013).
Berruete, I. M. et al. First report of bark canker disease of poplar caused by Lonsdalea quercina subp. Populi in Spain. Plant Dis. 100, 2159. https://doi.org/10.1094/PDIS-03-16-0405-PDN (2016).
Kovacova, V. et al. Recovery of a phytopathogenic bacterium Lonsdalea quercina from a lesser horseshoe bat in Moravian karst, Czech Republic. For. Pathol. https://doi.org/10.1111/efp.12379 (2018).
Husseneder, C., Park, J. S., Howells, A., Tikhe, C. V. & Davis, J. A. Bacteria associated with Piezodorus guildinii (Hemiptera: Pentatomidae), with special reference to those transmitted by feeding. Environ. Entomol. 46, 159–166. https://doi.org/10.1093/ee/nvw112 (2017).
DiGiulio, D. B. et al. Temporal and spatial variation of the human microbiota during pregnancy. Proc. Natl. Acad. Sci. USA 112, 11060–11065. https://doi.org/10.1073/pnas.1502875112 (2015).
Benner, M., Ferwerda, G., Joosten, I. & van der Molen, R. G. How uterine microbiota might be responsible for a receptive, fertile endometrium. Hum. Reprod. Update 24, 393–415. https://doi.org/10.1093/humupd/dmy012 (2018).
Moreno, I. et al. Evidence that the endometrial microbiota has an effect on implantation success or failure. Am. J. Obstet. Gynecol. 215, 684–703. https://doi.org/10.1016/j.ajog.2016.09.075 (2016).
Barba, M. et al. Vaginal microbiota is stable throughout the estrous cycle in Arabian maress. Animals 10, 1–14. https://doi.org/10.3390/ani10112020 (2020).
Li, C., Tucker-Kellogg, L. & Nagarajan, N. System biology modeling with compositional microbiome data reveals personalized gut microbial dynamics and keystone species. bioRxiv, 288803 (2018).
Flores, E. J. et al. Peptoanaerobacter stomatis primes human neutrophils and induces granule exocytosis. Infect. Immun. https://doi.org/10.1128/IAI.01043-16 (2017).
Zheng, J. et al. The placental microbiome varies in association with low birth weight in full-term neonates. Nutrients 7, 6924–6937. https://doi.org/10.3390/nu7085315 (2015).
Briana, D. D., Papaevangelou, V. & Malamitsi-Puchner, A. The jury is still out on the existence of a placental microbiome. Acta Paediatr. Int. J. Paediatr. 110, 2958–2963. https://doi.org/10.1111/apa.16048 (2021).
McKnight, D. T. et al. microDecon: A highly accurate read-subtraction tool for the post-sequencing removal of contamination in metabarcoding studies. Environ. DNA 1, 14–25. https://doi.org/10.1002/edn3.11 (2019).
Stinson, L. F., Boyce, M. C., Payne, M. S. & Keelan, J. A. The not-so-sterile womb: Evidence that the human fetus is exposed to bacteria prior to birth. Front. Microbiol. https://doi.org/10.3389/fmicb.2019.01124 (2019).
Chen, C. et al. The microbiota continuum along the female reproductive tract and its relation to uterine-related diseases. Nat. Commun. https://doi.org/10.1038/s41467-017-00901-0 (2017).
Bicalho, M. L. S., Machado, V. S., Higgins, C. H., Lima, F. S. & Bicalho, R. C. Genetic and functional analysis of the bovine uterine microbiota. Part I: Metritis versus healthy cows. J. Dairy Sci. 100, 3850–3862. https://doi.org/10.3168/jds.2016-12058 (2017).
Percie du Sert, N. et al. The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. PLoS Biol. 18, e3000410. https://doi.org/10.1371/journal.pbio.3000410 (2020).
Heil, B. A. et al. Metagenetic characterization of the resident equine uterine microbiome using multiple techniques. J. Equine Vet. Sci. https://doi.org/10.1016/j.jevs.2018.05.156 (2018).
Diel de Amorim, M. et al. Comparison of clinical signs, endometrial culture, endometrial cytology, uterine low-volume lavage, and uterine biopsy and combinations in the diagnosis of equine endometritis. J. Equine Vet. Sci. 44, 54–61. https://doi.org/10.1016/j.jevs.2015.10.012 (2016).
Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583. https://doi.org/10.1038/nmeth.3869 (2016).
Kumar, S., Stecher, G. & Tamura, K. MEGA7: Molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874. https://doi.org/10.1093/molbev/msw054 (2016).
Acknowledgements
This work was supported by Bullock Endowed Professorship (GRH) and Oklahoma Agricultural Experimentation HATCH Grant OKL03192 (UD).
Author information
Authors and Affiliations
Contributions
G.R.H., C.C.L. and U.D. contributed to the conception and design of the study. G.R.H. and U.D. provided the funding. G.R.H., J.L.S., and A.G. procured the animals and performed sample collection. H.U.P. and X.W. preformed laboratory analyses and bioinformatic analyses. H.U.P. wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Holyoak, G.R., Premathilake, H.U., Lyman, C.C. et al. The healthy equine uterus harbors a distinct core microbiome plus a rich and diverse microbiome that varies with geographical location. Sci Rep 12, 14790 (2022). https://doi.org/10.1038/s41598-022-18971-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-18971-6
- Springer Nature Limited
This article is cited by
-
Isolation of lactic acid bacteria from the reproductive tract of mares as potentially beneficial strains to prevent equine endometritis
Veterinary Research Communications (2024)