Nemabiome sequencing reveals seasonal and age associated patterns of strongyle infection and high prevalence of Strongylus vulgaris in Alberta feral horses

Abstract

Unmanaged feral horses, naïve to dewormers, offer a unique opportunity to study natural communities of equine parasites. These communities may include parasites that are rare in managed populations, and these may be transmitted to domestic horses in areas where there is contact between feral and domestic equine populations. There have been only a few studies of gastrointestinal parasite populations in horses, and very few from North American equine populations. This study aimed to gain insights into parasite biology through identification of the strongyle parasite species infecting feral horses in Alberta, Canada, and to test for species-specific infection patterns across season and horse age. Fecal samples (N = 149) were collected from unique individuals in the Sundre Equine Management Zone (EMZ), Alberta, across two years: 2021 (N = 62) and 2022 (N = 87). In 2021, samples were collected in summer (N = 31; 8 foals, 5 subadults, 18 adults) and fall (N = 31; 5 foals, 1 subadult, 25 adults). In 2022, samples were collected in spring (N = 36; 4 subadults, 32 adults), summer (N = 41; 4 foals, 8 subadults, 29 adults), and fall (N = 20; 1 foal, 2 subadults, 17 adults). Fecal egg counts showed that these horses shed high numbers of strongyle eggs relative to domestic horse populations (mean = 1337.01 ± 961.81 epg), and nemabiome analyses identified a total of 34 strongyle species. Species richness and aggregate strongyle FECs were highest in subadults and during the summer, while lowest in foals and during the fall. There was a high prevalence of large strongyle species, especially Strongylus vulgaris (85.91 %), with strongyle species-specific prevalence and FECs strongly associated with age and season. Understanding the factors driving species-specific parasite infection provides important information on strongyle parasite ecology and may aid the development of targeted parasite control strategies.

Graphical abstract

Highlights

• •Nemabiome sequencing revealed 34 strongyle species in Alberta feral horses.
• •High prevalence of large strongyles and Strongylus vulgaris in Alberta feral horses.
• •Alberta feral horses may act as reservoirs for clinically relevant strongyles.
• •Strongyle species prevalence and abundance vary with season and horse age.

1. Introduction

Strongyle nematodes are the most common and historically important parasites of equines due to their ubiquity and negative impacts on horse health (Khan et al., 2015). Equine strongyles consist of 64 recognized species (Lichtenfels et al., 2008) and most horses are co-infected by a diverse community of multiple strongyle species (Bellaw and Nielsen, 2020; Poissant et al., 2021), some of which are associated with species-specific pathologies. The large migratory strongyle Strongylus vulgaris, is the most pathogenic, causing verminous arteritis leading to infarction, tissue damage, and potentially death (McCraw and Slocombe, 1976). Other Strongylus species migrate through tissues such as the liver and pancreas causing tissue lesions (Khan et al., 2015). Small strongyles, also referred to as cyathostomins, are non-migratory but encyst in the gastrointestinal mucosa during development. Infection by cyathostomins may lead to a clinical syndrome called larval cyathostominosis, associated with mass synchronous emergence of encysted larvae into the intestinal lumen (Love et al., 1999). This syndrome is characterised by diarrhea, lethargy, colic, inappetence, peripheral edema and weight loss in horses and can be fatal even with medical management (Murphy et al., 1997). Since pathogenicity varies between strongyle species, the community composition of strongyle parasites is likely an important determinant of the clinical manifestation of strongyle infection in equids.

Climatic conditions are known to influence the number of infective strongyle larvae in contaminated pastures (Nielsen et al., 2007; Rossanigo and Gruner, 1994), which can influence seasonal variations in parasite exposure rate. Strongyle egg shedding is greatest during warm and wet seasons in temperate regions (Sargison et al., 2022; Slivinska et al., 2020). Seasonal patterns are expected to reflect strongyle parasite biology, with warm and humid environments ideal for the development and survival of both their pre-infective (eggs to L2) and the infective (L3) stages (Leathwick et al., 2019). The seasonality of strongyle egg shedding has informed equine parasite management programs with yearly deworming in temperate regions typically commencing at the start of spring to coincide with the anticipated increase in egg shedding (Rendle et al., 2024).

Understanding the dynamics of equine gastrointestinal parasite populations over changing seasons is key for developing parasite control strategies to reduce the burden of parasitism and improve equine health (Abbas et al., 2023). Fecal egg counts (FECs) are a common method for quantifying strongyle egg shedding in horses, but have limitations (Nielsen, 2021). Although FECs have a weak correlation with adult worm burdens (Nielsen et al., 2010), they remain valuable for assessing egg shedding levels and informing targeted parasite control strategies (Nielsen, 2022). However, FECs are unable to resolve strongyle species-level identification (Lichtenfels et al., 2008), limiting their utility in the study of species-specific infection patterns which are known to vary between individual horses, social bands, and across seasons (Wood et al., 2013; Ahn et al., 2024). Since pathology may vary among strongyle species, the lack of seasonal species-specific epidemiological studies limits our ability to develop targeted management strategies to control specific pathogenic gastrointestinal nematodes.

While non-invasive identification of some strongyle species is possible through morphological identification of coprocultured third stage larvae (L3) (Russell, 1948), this technique has many limitations and requires significant training and time. To overcome the challenge of obtaining species-specific information non-invasively, nemabiome sequencing, which involves DNA metabarcoding of the ITS2 gene region, can be used (Avramenko et al., 2015). For equine strongyles, nemabiome sequencing has been applied using eggs isolated from feces, detecting the presence of 20 species (Mitchell et al., 2019). Subsequently, Poissant et al. (2021) developed a nemabiome assay to characterize equine strongyle parasites using L3 larvae cultured from fecal samples. Recent nemabiome studies of strongyle parasites in managed horses in Australia (Abbas et al., 2023), Scotland (Sargison et al., 2022), Thailand (Hamad et al., 2024) and in feral horses on Sable Island, Canada (Ahn et al., 2024), showed that equine strongyles exhibit species-specific seasonal patterns. This technique has also been utilized in egg reappearance period investigation (Nielsen et al., 2022), and anthelmintic resistance investigations (Bull et al., 2024), which revealed that resistance can be species-specific.

In this study, we apply nemabiome metabarcoding to the study of mixed strongyle parasite infections in an unmanaged population of feral horses in Alberta, Canada. Previous nemabiome sequencing of a limited number of samples from Alberta feral horses alluded to potential temporal variation in strongyle parasite communities, as well as the presence of the highly pathogenic Strongylus vulgaris (Poissant et al., 2021). While intense deworming programs were once thought to have eradicated S. vulgaris in domestic populations, recent evidence that this species is maintained in Alberta's domestic horse population (Domshy et al., 2024) highlights the clinical relevance of investigating species-specific patterns of strongyle infection. Given that unmanaged horse populations may act as reservoirs for such pathogenic species (Harvey et al., 2019), it is critical to understand the ecology of strongyle parasites in wild populations to assess the potential risk of spillover into domestic populations.

2. Materials and methods

2.1. Study population

The study was conducted at the Sundre Equine Management Zone (EMZ), which extends westward from Sundre, Alberta toward Banff National Park and supports numerous bands of feral horses. The Government of Alberta's feral horse minimum count in the Sundre EMZ was 969 horses in 2023 (Government of Alberta, 2023). Horses were categorized into three age groups based on visual observations: foals (0–1 year), sub-adult (>1–3 years) and adults (>3 years).

2.2. Fecal sample collection

Fecal samples from unique individuals were collected from June 2021 to November 2022, spanning three defined time periods: (i) Spring (April through June 20), (ii) Summer (June 21 to September 30), and (iii) Fall (October and November). Within this study period, sampling occurred during one spring season (Spring, 2022), two summer seasons (Summer, 2021 and Summer, 2022), and two fall seasons (Fall, 2021 and Fall, 2022). We note that no foals were sampled in Spring (2022). Sampling design is described in Supplementary Table 1.

Feces were collected within 1 h of observed voiding by a horse and care was taken to collect only feces that were not in direct contact with the ground. Feces dedicated to FECs were carefully collected using plastic bags, with excess air expelled, and immediately stored at 4 °C until the FECs were conducted within 7 days of collection. FECs were performed using a modified McMasters method (Taylor et al., 2015). Three grams of sample were placed in a plastic tube and the volume made up to 45 mL with floatation solution composed of 160 g of sodium chloride (NaCl) and 200 g of table sugar in 400 mL of distilled water. Eggs of strongyle species and ascarids were counted, then multiplied by 50 to obtain the number of eggs per gram (epg) of feces. While ascarids were not the focus of this study, ascarid infection patterns are also reported in the results.

The McMaster method has been widely used to assess strongyle infections in both domestic (Abbas et al., 2023; Kuzmina et al., 2016) and wild horse populations (Debeffe et al., 2016; Harvey et al., 2019; Slivinska et al., 2020) and remains the industry standard for fecal egg counting in equine studies despite its higher detection limit compared to other methods (Nielsen, 2021). Although its accuracy is lower, McMaster counts show trends similar to the more sensitive Mini-FLOTAC in naturally infected equine samples (Noel et al., 2017). Furthermore, its precision and accuracy improve with higher egg shedding, as demonstrated in fecal egg count reduction tests (Levecke et al., 2012), making it well-suited for assessing strongyle infection in wild horse populations where egg counts are typically high.

A subsample of feces was also collected for larval coprocultures. Approximately 30 g of fresh feces was collected in plastic cups, moistened with distilled water and mixed and covered with a perforated lid to provide aeration. Fecal samples were cultured for 21–26 days at room temperature (∼22 °C) and were misted with distilled water every 3–4 days. Following culturing, larvae were harvested using the glass-over-petri-dish-method (Roberts and O'Sullivan, 1950) and fixed in 70 % ethanol. The total worm number from each sample was estimated as the average number of larvae counted in two 10 μL aliquots of the sample. Larvae were stored at −80 °C until further processing.

2.3. DNA preparation, amplification, and sequencing

Preparation of L3 larvae for DNA extraction followed methods described in Poissant et al. (2021), with slight modifications. Approximately 1000 L3 larvae were aliquoted for DNA extraction and centrifuged at 14000 RPM for 2 min, followed by carefully removing the ethanol using a pipette. Samples were then placed on a SpeedVac DNA130 Vacuum Concentrator set to room temperature for 2 h to remove any remaining ethanol, and resuspended in 150 μL of lysis buffer (recipe in (Avramenko et al., 2015)). To lyse the worms, 6 μL of proteinase K (20 mg/mL) and two stainless steel beads (5 mm) were added, then samples were placed on a Vortex-Genie 2 for 10 min of bead beating. Samples were then placed on a shaker (300 rpm) in a 55 °C incubator for 2 h, after which the samples were placed on a heat plate set to 95 °C for 20 min to deactivate the proteinase K. DNA lysate was diluted to 1:10 using molecular grade water.

For each sample, PCR amplification was conducted using a KAPA HiFi Hotstart PCR plus dNTPs kit (Roche, USA) following manufacturer recommendations and following methods described in Poissant et al. (2021). The first PCR reaction had a total volume of 25 μL: 12.5 μL of molecular grade water, 5 μL of KAPA HiFi Buffer, 0.75 μL of each of the NC1 and NC2 primers (10 μM (Gasser et al., 1993);), 0.75 μL dNTPs (10 mM), 0.5 μL of KAPA HiFi Polymerase (0.5 U), and 5 μL of 1:10 DNA lysate as template. Thermocycling conditions for the first PCR were: 95 °C for 3 min, followed by 25 cycles of 98 °C for 20 s, 62 °C for 15 s, 72 °C for 15s, and a final extension at 72 °C for 2 min. Magnetic beads (Beckman Coulter, AMPure XP) were used to purify the first PCR product using a 1:1 ratio. A second PCR was conducted to add unique dual indexes for sequencing: 13.25 μL of molecular grade water, 5 μL of KAPA HiFi Buffer, 2.5 μL of a primer solution (5 μM of each unique dual index), 0.75 μL dNTPs (10 mM), 0,5 μL of KAPA HiFI Polymerase (0.5 U), and 3 μL of purified product from the first PCR was then conducted with the following thermocycling conditions: 98 °C for 45 s, followed by 7 cycles of 98 °C for 20 s, 63 °C for 20 s, and a final extension at 72 °C for 2 min. Magnetic bead purification was repeated for the second PCR product using again a 1:1 ratio. The DNA concentration of the purified product was quantified using a BioTek Take3 Plate Reader and approximately 100 ng of purified DNA from each sample was pooled into two libraries. DNA concentrations of the pooled libraries were quantified using a Qubit 1X dsDNA High Sensitivity Kit. Pooled libraries were sequenced across two separate runs, using 12 pM diluted libraries with 20 % PhiX sequencing control v3 spike in. One sequencing run (n = 62) was performed using an Illumina 500-cycle MiSeq Nano Kit v2 and only included samples from Alberta feral horses. Another sequencing run (n = 87) was performed using an Illumina 500-cycle MiSeq Kit v2, which included nemabiome samples from other horse populations. In total, 149 samples from Alberta feral horses were sequenced for this study.

2.4. Bioinformatics pipeline

The results of each sequencing run were separately processed using version 2 of a published bioinformatics pipeline (Poissant et al., 2021; available at https://data.mendeley.com/datasets/vhyysw8xt2/2), allowing up to 2.5 % base mismatches when merging the forward and reverse reads with the mergePairs function (Callahan et al., 2016). This percent mismatch has been used in previous studies (Ahn et al., 2024; Nielsen et al., 2022) and was chosen based on independent visual inspection of each sequencing run, with 2.5 % identified as an appropriate value that retained relatively abundant amplicons sequence variants (ASVs) while removing rare sequences which are likely due to technical errors. Filtered sequences were assigned taxonomy with assignTaxonomy set to ≥80 % confidence using a curated ITS2 reference database (v1.5.0) which consolidated synonymous species names and removed incomplete or poor quality sequences (Poissant et al., 2021; Workentine et al., 2020). A total of 291 sequences of equine strongyle species were included in the curated database, representing 43/64 recognized equine strongyle species (Lichtenfels et al., 2008) with no additional species compared to v1.4.0 (Ahn et al., 2024). Given that ∼1000 larvae were processed for nemabiome sequencing, species that represented less than 1/1000 proportion of the reads in a sample were removed as a final filtering step to limit false positives due to index hopping or contamination from aerosols.

To account for species-specific biases associated with nemabiome metabarcoding, read counts of Strongylus edentatus, Strongylus equinus, and Strongylus vulgaris were divided by previously estimated correction factors (S. edentatus: 1.523, S. equinus: 1.809, S. vulgaris: 4.648; Ahn et al., 2024). Following correction, species-specific fecal egg counts (FECs) were estimated by multiplying the aggregate FEC by species-specific relative abundances derived from nemabiome sequencing. Although L3 larvae and eggs are distinct sample types, studies have demonstrated a high correlation in species composition between L3 larvae and eggs collected from the same fecal sample when using nemabiome sequencing of the ITS2 region (Courtot et al., 2023; Redman et al., 2019), indicating that L3-derived relative abundances can reasonably approximate species-level egg shedding. This approach has been successfully applied in studies such as Bull et al. (2024) and Queiroz et al. (2020), where nemabiome sequencing enhanced species-level FEC assessments and enabled more precise evaluations of strongyle parasite communities. It is also suggested by the World Association for the Advancement of Veterinary Parasitology (WAAVP) guidelines for understanding parasite dynamics, including for equines (Kaplan et al., 2023). Four samples had 0 EPG following McMaster counts, though larvae were successfully harvested and sequenced. For these samples, the FEC was adjusted to 50 EPG, the minimum detectable threshold.

2.5. Statistical analysis

Data analysis was performed using R v.4.4.1. Generalized linear models (GLMs) were constructed using the package ‘glmmTMB’ (Brooks et al., 2017) for species richness, strongyle FEC, species-specific prevalence, and species-specific FEC. Strongyle FEC and species-specific FEC were cube-root transformed for analyses. Species richness models were constructed using a Poisson error family, strongyle FEC and species-specific FEC models were constructed using the Tweedie error family, with species-specific FEC accounting for zero-inflation (ziformula = ∼1), and prevalence models were constructed using a binomial error family. A total of 149 samples from unique individuals were included in this study. All models included the following fixed effects: age class (3-level factor: foal (N = 18), subadult (N = 18), adult (N = 113)), season (3-level factor: Spring (N = 36), Summer (N = 72), Fall (N = 41)), year (2-level factor: 2021 (N = 62), 2022 (N = 87)) and sex (2-level factor: male (N = 78, female (N = 71)). Significance of fixed effects was assessed using an ANOVA with type 2 Wald's test, with post-hoc Tukey's contrasts conducted to identify significance in pairwise differences among age classes and seasons. Species-specific models were only constructed for the species with >50 % population-wide prevalence.

To test for differences in the composition of strongyle parasite communities among horses, Jaccard and Bray-Curtis dissimilarity matrixes were constructed using the function diversity in the R package ‘vegan’ 2.6–4 (Oksanen J et al., 2022). Jaccard dissimilarity was used to compare strongyle community composition based on species presence-absence, while Bray-Curtis dissimilarity considers species-specific FEC (i.e., abundance). A PERMANOVA was conducted using the same model formulation as above with age class, season, year, and sex as fixed effects with 9999 permutations. The significance and proportion of variance in community dissimilarity explained by the fixed effects (R²) generated from PERMANOVAs were used for interpretations. Community dissimilarity was subsequently visualized through Principal coordinate analyses. Mean values are presented with ± Standard Deviation.

3. Results

3.1. Egg shedding patterns

Aggregate strongyle FEC was significantly associated with host age and season (Supplementary Table 2). Adult and subadult horses had similarly high FECs of 1428.45 ± 946.39 epg and 1797.22 ± 740.32 epg, respectively (Tukey's contrasts, p = 0.08). Both adults and subadults had significantly higher FECs compared to foals (302.78 ± 438.74 epg, Tukey's contrasts, p < 0.001).

Strongyle egg shedding was highest in Summer with an average of 1552.29 ± 1087.62 epg which was significantly greater than the average FEC in Fall (919.51 ± 676.01 epg, Tukey's contrasts, p < 0.001, Fig. 1), but similar to Spring (1381.94 ± 829.64 epg, Tukey's contrasts, p = 0.07, Fig. 1). Egg shedding was higher in 2022 (1436.95 ± 863.23 epg) compared to 2021 (1197.77 ± 1076.88 epg), though this difference was not significant (GLM, p = 0.19). While females appeared to have higher FEC (1407.69 ± 1079.38 epg) than males (1259.37 ± 814.02 epg), this difference was not significant (GLM, p = 0.80).

Fig. 1.

Age and seasonal variation in strongyle species richness and fecal egg counts from 149 feral horses from the Sundre Equine Management Zone in Alberta, Canada. Foals (N = 18) are 0–1 years old, Subadults (N = 18) are >1–3 years old, and Adults (N = 113) are >3 years old. Sampling occurred across three seasons: Spring (N = 49), Summer (N = 50), and Fall (N = 50).

Ascarid eggs were detected in 20.13 % of samples. Subadults had the highest prevalence of ascarid egg shedding (61.11 %, n = 11/18), followed by foals (33.33 %, n = 6/18) and adults (11.50 %, n = 13/113). Ascarid FEC was generally low ranging between 0 and 650 epg, with a mean of 39.43 ± 112.15 epg. Subadults had the highest ascarid FEC (138. 89 ± 185.94 epg), followed by foals (97.22 ± 201.08 epg) and adults (14.38 ± 49.75 epg).

3.2. Nemabioime sequencing and species richness

Mean sequencing depth across the 149 samples was 52,372.68 ± 47,860.81, ranging between 1983 and 214,513 raw reads per sample. Most ASVs were successfully assigned to a species with >80 % confidence, with a mean of 8.00 ± 0.52 % of reads being unclassified across samples (Fig. 2). Upon investigation, most of the unclassified reads were being assigned to Cylicostephanus calicatus or Cylicocylus ashworthi below the 80 % confidence threshold. Nemabiome sequencing detected a total of 34 strongyle species: 8 species of large strongyles and 26 species of cyathostomins (Table 1). Every horse was infected by multiple strongyle species, with total species numbers ranging from 4 to 29 species detected in a single horse.

Fig. 2.

The relative abundance of the 34 strongyle species detected in Alberta feral horses across season and horse age. Sample sizes for each category are displayed above each bar.

Table 1.

Population prevalence and mean strongyle fecal egg counts in feral horses from the Sundre Equine Management Zone, Alberta, Canada, based on 149 samples collected between June 2021 and November 2022.

Species	Prevalence (%)	Mean FEC ± SD (EPG)
Cylicostephanus longibursatus	98.66	149.02 ± 136.12
Cyathostomum catinatum	97.99	85.62 ± 88.25
Cylicocyclus ashworthi	95.97	131.00 ± 183.19
Cylicocyclus nassatus	95.30	312.89 ± 378.55
Cylicostephanus minutus	93.96	42.51 ± 50.49
Cylicocyclus leptostomum	90.60	92.5 ± 132.32
Cylicostephanus calicatus	89.26	14.59 ± 16.98
Cylicostephanus goldi	88.59	46.35 ± 63.61
Coronocyclus coronatus	86.58	47.3 ± 58.54
Strongylus vulgaris	85.91	31.03 ± 32.79
Strongylus edentatus	84.56	90.46 ± 112.65
Cylicocyclus insigne	70.47	19.83 ± 39.39
Strongylus equinus	69.80	101.12 ± 190.75
Cyathostomum pateratum	53.69	8.61 ± 25.01
Coronocyclus labratus	51.01	3.53 ± 6.58
Coronocyclus labiatus	49.66	6.54 ± 13.05
Cylicodontophorus bicoronatus	44.30	1.48 ± 3.32
Poteriostomum imparidentatum	43.62	4.65 ± 17.88
Triodontophorus serratus	37.58	1.66 ± 4.02
Poteriostomum ratzii	32.89	4.31 ± 10.34
Parapoteriostomum euproctus	30.20	2.19 ± 5.93
Cylicocyclus radiatus	28.86	3.61 ± 14.14
Petrovinema poculatum	28.86	1.67 ± 4.98
Cylicocyclus elongatus	27.52	1.85 ± 4.43
Triodontophorus brevicauda	26.17	8.15 ± 25.96
Gyalocephalus capitatus	25.50	0.66 ± 1.43
Parapoteriostomum mettami	22.82	1.1 ± 3.78
Craterostomum acuticaudatum	21.48	2.19 ± 7.56
Triodontophorus nipponicus	20.81	1.95 ± 8.38
Cylicostephanus bidentatus	18.79	2.06 ± 8.68
Tridentoinfundibulum gobi	13.42	1.15 ± 4.24
Cylicocyclus brevicapsulatus	11.41	3.34 ± 17.13
Cylicocyclus ultrajectinus	10.74	0.5 ± 2.16
Oesophagodontus robustus	06.04	0.18 ± 0.87

Species richness significantly differed between every age class (Supplementary Table 2), subadults were infected with the greatest number of species (20.838 ± 4.18 species, Tukey's contrasts p < 0.05), followed by adults (17.91± 4.50 species, Tukey's contrasts, p < 0.05), and foals had the lowest species richness (11.83 ± 3.96 species, Tukey's contrasts, p < 0.05). Female horses were infected by 18.53± 4.88 strongyle species, which was significantly more than males (16.44 ± 4.83 species, Tukey's contrasts, p = 0.02). Species richness was similar between Spring (18.86 ± 5.00 species) and Fall (19.21 ± 4.83 species, Tukey's contrasts, p = 0.95), though both had higher richness compared to Summer (15.90 ± 4.52 species, Tukey's contrasts, p < 0.05). While our GLMs indicated that richness was significantly higher in 2021 (GLM, p = 0.03), this was marginal (2021: 17.63 ± 5.00 species, 2022: 17.46 ± 4.95 species).

3.3. Specific species prevalence

Strongylus species were common in Alberta feral horses, with S. vulgaris as the most prevalent (85.91 %), followed closely by S. edentatus (84.56 %) and S. equinus (69.80 %). The prevalence of the three Strongylus species was significantly higher in adults compared to foals (Fig. 3, Tukey's contrasts, p < 0.01). While subadults had lower prevalences of S. vulgaris and S. equinus than adults, this difference was not statistically significant (p > 0.05). However, subadults had a significantly higher prevalence of S. edentatus than adults (Tukey's contrasts, p < 0.01, Fig. 3). Neither sex nor season significantly affected the prevalence of the three Strongylus species (Supplementary Table 2, GLM, p > 0.05).

Fig. 3.

Species-specific prevalence of the 15 most common strongyle parasite species by age class modeled in the study. Foals are 0–1 years old (N = 18), Subadults are >1–3 years old (N = 18), and Adults are >3 years old (N = 113). Significant differences are denoted by asterisks: p < 0.05 (∗), p < 0.01 (∗∗), and p < 0.001 (∗∗∗).

Of the 26 species of cyathostomins identified, 10 were very common with population-wide prevalences greater than 70 % (Table 1). Cylicostephanus longibursatus was the most prevalent cyathostomin (N = 147/149), followed by Cyathostomum catinatum (N = 146/149) and Cylicocyclus ashworthii (N = 143/149). In contrast, other species were rare, such as Cylicocyclus ultrajectinus which was only detected in N = 16/149 horses (Table 1). For the cyathostomin species modeled, horse age was significantly associated with the prevalence of Cylicostephanus insigne, and Cylicostephanus minutus, with adults having the highest, while subadults had the highest prevalence of Coronocyclus labratus (Fig. 3, Tukey's contrasts, p < 0.05). In contrast, sex, year, and season had less influence on cyathostomin species prevalences. Season was significant for C. coronatus and Cylicostephanus calicatus, both of which had low prevalence in Spring, in contrast to Cyathostomum pateratum which had the highest prevalence in Spring (Supplementary Fig. 1, Tukey's contrasts, p < 0.05). Coronocyclus labratus and C. insigne were more prevalent in 2021 (N = 40/62; 53/62, respectively) compared to 2022 (N = 35/87; 53/87, respectively) (Supplementary Table 2, GLM p < 0.05). Finally, Cylicostephanus goldi was significantly more prevalent in mares (Supplementary Table 2, GLM, p < 0.05), while there was no impact of sex on the prevalence of any other cyathostomin species.

3.4. Species-specific fecal egg count patterns

Strongylus vulgaris FECs were not significantly different between adults (36.13 ± 33.93 epg) and subadults (30.01 ± 24.19 epg), although both were significantly higher compared to S. vulgaris FECs in foals (0.04 ± 0.13 epg, Tukey's contrasts, p < 0.001, Fig. 4). Strongylus vulgaris FECs were significantly higher in Summer (38.84 ± 36.34 epg) compared to Fall (18.61 ± 29.58 epg) and Spring (29.54 ± 23.65 epg); although Spring had higher S. vulgaris FECs compared to Fall, this was not significantly different (Tukey's contrasts, p > 0.05). While the age-structured pattern observed for S. vulgaris was also observed for S. edentatus and S. equinus (Fig. 4), seasonal FEC patterns differed (Fig. 4). Strongylus edentatus egg shedding was significantly higher in Spring (138.82 ± 159.01 epg) and Summer (91.70± 92.29 epg) than Fall (45.83 ± 74.45 epg, Tukey's contrasts, p < 0.01). In contrast, S. equinus FECs did not change across seasons (Fig. 4). While S. equinus was more abundant in 2022 (135.30 ± 228.98 epg) compared to 2021 (53.16 ± 101.75 epg) (Supplementary Table 2, GLM p = 0.04), the high intrayear variability may limit the biological significance of this interannual difference. There were no significant differences in Strongylus spp. FECs between males and females.

Fig. 4.

Species-specific fecal egg counts of 15 strongyle species infecting Alberta feral horses by age class. Foals are 0–1years old (N = 18), Subadults are >1–3years old (N = 18) and Adults are >3years old (N = 113). Significant differences are denoted by asterisks: p < 0.05 (∗), p < 0.01 (∗∗), and p < 0.001 (∗∗∗).

In general, species-specific cyathostomin FECs were highest in subadults with the exception of Cylicostephanus longibursatus and Cylicostephanus goldi whose eggs were shed in greater quantities by adults than subadults (Fig. 5). For all cyathostomin species, foals had the lowest FECs compared to both adults and subadults (Fig. 5). The most common seasonal FEC patterns for cyathostomins was a rise in Spring to a peak in Summer followed by a decline in the Fall (Fig. 5). However, C. catinatum had the highest FECs in the Fall while C. goldi and C. pateratum had the highest FEC in the Spring (Fig. 5). Year was not associated with the FEC of any species modeled (Supplementary Table 2, Tukey's contrast, p > 0.05). Furthermore, cyathostomin FECs were generally unaffected by sex with only Cyathostomum catinatum FECs significantly higher in male horses (98.02 ± 89.92 epg) compared to females (76.53 ± 86.55 epg, GLM, p < 0.05) (see Fig. 6).

Fig. 5.

Species-specific fecal egg counts of 15 strongyle species infecting Alberta feral horses grouped by samples collected in Spring (N = 36), Summer (N = 72), and Fall (N = 41). Significant differences are denoted by asterisks: p < 0.05 (∗), p < 0.01 (∗∗), and p < 0.001 (∗∗∗).

Fig. 6.

Principal coordinate analysis of the (A, C) Jaccard and (B, D) Bray-Curtis dissimilarity matrices of parasitic strongyle community composition among N = 149 feral horses from the Sundre Equine Management Zone, Alberta, Canada. Points are colored by Age (Foals N = 18; Subadults N = 18; Adults N = 113) or Season (Spring N = 49; Summer N = 50; Fall N = 50). Each point represents a strongyle parasite community.

3.5. Patterns of strongyle community composition

Age explained the most variation in community composition based on Bray-Curtis Dissimilarity (PERMANOVA, R² = 0.21, p < 0.001), though less for Jaccard dissimilarity (PERMANOVA, R² = 0.14, p < 0.001). Season explained less variation than age, which did not appear to differ between Jaccard (PERMANOVA, R² = 0.080, p < 0.001) or Bray-Curtis (PERMANOVA, R² = 0.075, p < 0.001) dissimilarity Both year and sex were significant for both dissimilarity measures (PERMANOVA, p < 0.05), but their contributions to community composition were minimal, with year explaining R² = 0.023 (Jaccard) and R² = 0.010 (Bray-Curtis) of variance, and sex explaining R² = 0.018 (Jaccard) and R² = 0.013 (Bray-Curtis) of variance.

4. Discussion

Nemabiome sequencing revealed a diverse community of at least 34 strongyle species infecting the Sundre feral horse population in Alberta, Canada. This is 3 more species than found in a previous study on a limited number of feral horses from the same area (Poissant et al., 2021): Strongylus equinus, Cylicocylus radiatus, and Tridentoinfundibulum gobi. These horses have a greater species richness than has been reported in the feral horse population on Sable Island in Nova Scotia, Canada (25 species (Ahn et al., 2024)), as well as other managed horse populations (e.g., 20 species, Scotland (Sargison et al., 2022) and West Wales (Mitchell et al., 2019); 23 species, Australia (Abbas et al., 2023)). This study is one of the few species-specific strongyle parasite surveys of North American feral horse populations, indicating that large migratory strongyles are highly prevalent in unmanaged populations. The high prevalence of S. vulgaris (85.91 %) is of particular concern, given that this species is known for producing serious pathologic lesions and verminous arteritis. A high prevalence of S. vulgaris has also been reported in wild horses of Australia (Harvey et al., 2019); and Strongylus species are highly prevalent in Sable Island feral horses (Ahn et al., 2024; Jenkins et al., 2020). The high prevalence of S. vulgaris in Alberta feral horses, as well as other unmanaged horse populations, suggests that feral horses may serve as a reservoir for this species. Exposure to feral horse feces carries a risk of transmission of S. vulgaris between feral and domestic horses. While the risk of S. vulgaris in domestic populations was thought to have been eliminated through wide-scale deworming practices, recent studies suggest this is not the case. A recent post-mortem survey of domestic horses in Alberta found S. vulgaris-associated lesions in 17.3 % of necropsied horses over a 12 year period, indicating that S. vulgaris is still an important consideration in equine parasite management (Domshy et al., 2024). Further work is needed to explore the prevalence of S. vulgaris in domestic horse populations in the context of proximity and exposure to feral horses, and to investigate the potential of feral horses as a potential source of refugia as part of an integrated parasite management strategy in domestic horses.

Our results show high FECs in feral horses compared to previously reported values in domestic horses. Overall, we found that strongyle egg shedding increased from Spring to Summer and later declined in the Fall; this is similar to seasonal patterns of overall parasite egg shedding observed in other studies (Abbas et al., 2023; Sargison et al., 2022). However, our results indicate that seasonal patterns are species-specific, which may not be reflected in the overall strongyle FEC. This conclusion is supported by the significant amount of variation in strongyle parasite community dissimilarity explained by season for both Jaccard and Bray-Curtis dissimilarity. Even between the three Strongylus species, seasonal egg shedding patterns were not consistent. The contribution of S. edentatus, which is highest during Spring and Summer, may peak earlier in the year than that of S. vulgaris. In contrast, S. equinus does not show any seasonal trends. The feral horses of Alberta harbor a large community of strongyles consisting of several species. Since strongyle species is a determinant of pathology, parasite management strategies must be contextualized with knowledge of species-specific biology.

There was a distinct difference between the seasonality of species richness and FEC patterns. FECs increased from Spring to Summer unlike the species richness which decreased from a peak in Spring to its lowest value in Summer before a subsequent rise in the Fall. This may reflect changes in the strongyle parasite community structure throughout the seasons that may not be detected when only looking at aggregate strongyle FECs. Egg shedding by some species is lower in number compared to that of other species over some periods which may also account for the reduction in their detection in such periods. Larger strongyles are known to produce a greater number of eggs compared to smaller species (Kuzmina et al., 2012). Two large strongyles, S. edentatus and S. vulgaris had the highest FEC contributions in Summer, which may indicate that these two species specifically drive the seasonal patterns observed in aggregate strongyle FECs. In contrast, the high species richness in the Spring may reflect emergence of encysted cyathostomins following winter, with the decrease in Summer potentially indicating death of the previous year's adult cyathostomins, with an increase in richness in Fall reflecting new infections. Alternatively, seasonal variations in species richness and FECs may reflect the complex ecological relationship between the host, parasite and environment in terms of the flora in the environment being ingested by the feral horses. Malsa et al. (2024), reported a reduction in FECs and cyathostomin species richness in horses grazing chicory plants, as well as decreased larval development from the eggs shed by these horses. Various plants have been reported to contain bioactive compounds that are antiparasitic, such as tannins and terpenes (Olanrewaju et al., 2023; Rodríguez-Hernández et al., 2023), and the availability of different plant species may shift across seasons. Alberta feral horses have been shown to vary their habitat selection by season (Girard et al., 2013), but the impact of grazing/foraging behavior on their parasites is unknown. Although it may be the case that the dominance of high fecundity species in this study lowers the detection probability for rare and low fecundity ones since we randomly selected 1000 L3 larvae for nemabiome sequencing (i.e., lower likelihood of selecting a larva of a rare species). Given that equine strongyles vary in prevalence and fecundity, increasing the total number of larvae processed for nemabiome sequencing (e.g. 2500 L3 larvae as in Poissant et al., 2021) may provide more accurate estimates of species richness.

While our study provides valuable insights into the seasonal dynamics of strongyle egg shedding in Alberta's feral horses, certain limitations in study design and sample size should be acknowledged. Seasonal fluctuations in strongyle egg shedding are well-documented, with variations influenced by factors such as climate and pasture management practices (Abbas et al., 2023; Kuzmina et al., 2016), yet the uneven temporal coverage in our study may affect the generalizability of our findings. Although data collection spanned two years, only one spring sampling period was included, limiting our ability to capture inter-annual variability in springtime parasite dynamics, when egg shedding is typically highest in temperate regions like Alberta (Nielsen et al., 2007). While season was included as a fixed effect, the lack of spring replication may limit the interpretation of inter-annual differences. Furthermore, the sample sizes for specific age classes, notably foals and yearlings, were relatively small. Previous research has shown that age can significantly influence strongyle egg shedding patterns, with younger horses often exhibiting considerable variation (Ahn et al., 2024). The smaller sample sizes in these age groups may limit our ability to detect subtle differences or trends in species-specific prevalence estimates across seasons. Future studies with more balanced and comprehensive sampling across seasons and age groups would enhance the robustness of conclusions drawn about the ecology of strongyle infections in feral horse populations.

Intrinsic factors such as age and sex are known to be associated with strongyle parasite infection in horses (Ahn et al., 2024; Sallé et al., 2018). Horse age class appeared to be the most important consideration for strongyle parasite infection, particularly for abundance measures. For example, species-specific FEC of Cylicostephanus longibursatus was associated with age class but not for its prevalence, likely due to its high population wide prevalence limiting variation in prevalence across age class. Such patterns are likely driving the higher age-explained variation in Bray-Curtis dissimilarity compared to the Jaccard dissimilarity. Among Alberta feral horses, foals consistently had the lowest prevalence and infection intensity for strongyle species compared to adults and subadults, which may be a consequence of host and parasite life history. Foals may have the fewest encounters with parasites since they do not graze as much as subadults and adults, thereby decreasing exposure to contaminated herbage that harbors the infective stage of the parasites. Foals are also unlikely to shed strongyle eggs despite infection due to the long prepatent period of strongyle species. The variation in species-specific prevalence supports this expectation, with Strongylus species that have 6–10-month prepatent period not being common in foals (McCraw and Slocombe, 1978; Round, 1969). In contrast, foals are known to be primarily affected by ascarids (Nielsen, 2016). However, foals had low prevalence and intensity of ascarids in our study, which may indicate limited impact of ascarids for foals in Alberta feral horses. Interestingly, the high prevalence of ascarids in subadults, and the detection of ascarids in adults are unusual given they are rarely found in non-juveniles (Nielsen, 2016). The high prevalence of prepatent ascarid infection in adult Sable Island horses (Jenkins et al., 2020) suggests that the role of ascarids may be relevant beyond foals for unmanaged horse populations. Further work will be required to determine the extent of parasite transmission between feral and domestic horses in Alberta.

While factors such as sex and year may influence parasite infection dynamics (Ezenwa et al., 2012; Sallé et al., 2018; Wood et al., 2013), our study indicates their impact on strongyle community composition in Alberta feral horses. Sex differences in strongyle infections in horses have previously been associated with differences in energetic requirements between males and females (i.e., reproduction (Ahn et al., 2024; Debeffe et al., 2016)), which may be due to missing sex-specific data such as reproductive (e.g. pregnant, lactating) or social status (e.g. bands, bachelors) in our study. Similarly, the limited effect of year suggests that intra-annual (seasonal) dynamics may be more influential in shaping strongyle parasite infections (Slivinska et al., 2020). However, Wood et al. (2013) showed that year-to-year variation in aggregate FECs can be significant, likely driven by annual climate fluctuations. The limited year effects in our analysis may therefore reflect the constraints of our two-year sampling design, which might not have captured the full extent of interannual climate variability. Furthermore, our study included only a single spring sampling period, which may have further restricted our ability to detect interannual differences in parasite shedding during this critical season. To more accurately capture the influence of interannual climate variability, future work should incorporate longer-term sampling with direct environmental measurements such as rainfall, temperature, and humidity to better understand how climate-driven changes influence strongyle infection patterns over time.

We recognize that estimating species-specific egg shedding by multiplying larva-derived relative abundances by FEC is not without limitations, particularly due to potential biases between sample types. However, equine strongyle species generally exhibit consistent egg hatching patterns under optimal coproculture conditions (Mfitilodze and Hutchinson, 1987), suggesting that the species composition of L3 larvae reasonably reflects that of eggs. This approach is further supported by studies demonstrating high correlation in species composition between eggs and L3 larvae from the same samples when analyzed through nemabiome metabarcoding in both equine (Courtot et al., 2023) and ovine (Redman et al., 2019) strongyle infections. While the similarity between eggs and L3 larvae supports the use of larval relative abundances for species-specific estimates, we acknowledge that potential biases in nemabiome sequencing still exist. Differences in DNA yield across species have been linked to variations in the number of intestinal cells (Poissant et al., 2021). To address this, we applied correction factors to Strongylus edentatus, Strongylus equinus, and Strongylus vulgaris (Ahn et al., 2024), species known to have biased DNA representations during sequencing. In contrast, cyathostomin species possess consistent intestinal cell numbers (Russell, 1948) and demonstrate similar PCR efficiencies for the ITS2 region (Courtot et al., 2023), minimizing the risk of amplification bias. Species with differing cell counts, such as Triodontophorus (Russell, 1948), were rare and of low abundance in our dataset, limiting their influence on relative abundance estimates. Overall, the use of L3-derived relative abundances provides a practical and biologically relevant means of estimating species-specific fecal egg shedding, particularly in the context of assessing transmission risks from wild horse populations. This approach enables species-level resolution that is not achievable with traditional egg counting methods, providing the opportunity for greater insight into parasite community structure and transmission dynamics (Kaplan et al., 2023).

5. Conclusion

Here we report the first species-specific prevalences, and FECs for strongyle parasites in a large population of feral horses in the Sundre EMZ, Alberta. The parasite community in this horse population is exceptionally diverse compared to that of other horse populations, with a high prevalence of clinically relevant species such as S. vulgaris. Age and season influenced the patterns of species-specific prevalence and FEC. With limited human interventions, these feral horses provide an opportunity to study strongyle species in a wild population, albeit in close proximity to a large population of domestic horses that share the province. Further work is needed to determine the extent of parasite transmission between feral and domestic horses in Alberta, and how information on species-specific data can inform targeted strongyle management strategies.

CRediT authorship contribution statement

Grace Onyeche Ochigbo: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Sangwook Ahn: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Kobe Albert Belhumeur: Writing – review & editing, Methodology. Jocelyn Poissant: Writing – review & editing, Supervision, Resources, Conceptualization. Brielle Vastola Rosa: Writing – review & editing, Supervision, Resources, Methodology, Conceptualization.

Data availability

Sequencing data and accompanying metadata has been deposited in the NCBI SRA under the BioProject accession code PRJNA1269116.

Ethical standards

Sample collection and processing were performed under University of Calgary Animal Care Protocols AC21-0130 and AC22-0030.

Financial support

Funding for this project was provided by the Margaret Gunn Endowment for Animal Research (MGEAR). Grace Ochigbo was supported by an Alberta Graduate Excellence Scholarship (AGES). Sangwook Ahn was supported by a University of Calgary Eyes High Doctoral Recruitment and NSERC Canada Graduate Doctoral Scholarships.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.