Outbreak of acute larval cyathostominosis – A “perfect storm” of inflammation and dysbiosis

Abstract Background Cyathostomins are prevalent and pathogenic intestinal helminths of horses, causing acute and chronic disease, including acute larval cyathostominosis, which has a mortality rate of 50%. Factors determining individual susceptibility to acute larval cyathostominosis are unknown. Investigation of these factors could lead to novel treatment and prevention strategies. Objectives To investigate clinicopathological and faecal microbiota changes associated with disease in individual horses in an acute larval cyathostominosis outbreak. Study design Case series. Methods The study population was a herd of 23 mixed breed horses in Ireland. The outbreak occurred in November 2018. Fourteen horses were clinically affected. Clinical status was monitored and recorded. Blood and faecal sampling allowed clinicopathological, faecal 16s rRNA gene sequencing and faecal egg count analyses. Results Two horses were euthanised, whilst 12 recovered. Common clinical signs included loose faecal consistency, weight loss and pyrexia. Consistent clinicopathological findings were borderline anaemia, leucocytosis, thrombocytosis, hyperfibrinogenaemia, hyperglobulinaemia and a reverse A: G ratio. Decreased alpha‐diversity of the faecal microbiota and greater relative abundance of the genus Streptococcus, class Bacilli, order Lactobacillales and family Streptococcaceae, and family Prevotelleceae was found in clinically affected horses compared to their clinically normal cohorts. An increase in obligate fibrolytic bacteria was seen in the clinically normal group compared to the clinical group. Histopathological findings of the colon and caecum revealed a severe necrotising typhlocolitis associated with cyathostomin larvae and bacterial overgrowth in the mucosa of the large intestine. Main limitations The study population in this outbreak is small. There are several confounding factors limiting this to a descriptive case series. Faecal microbiota has been shown to reflect the large intestinal microbiota but do not represent changes directly. Conclusions These findings suggest that acute larval cyathostominosis is associated with dysbiosis of the gut microbiota as well as the inflammatory stimulus of numerous emerging larvae leading to structural and functional pathology of the large intestine.


| INTRODUC TI ON
Cyathostomins are pervasive parasitic helminths, currently considered to be one of the most pathogenic endoparasites of equids. [1][2][3][4] There are over 40 species of cyathostomins that infect horses and they are ubiquitous in grazing horses, with prevalence of 89%-100% reported in the literature. [4][5][6][7] Infection occurs through the faecal oral route, and larval stages of the parasites travel to the caecum and colon where they encyst in the intestinal mucosa and may become inhibited for up to three years before re-emerging to mature to adult worms. 8 Heavy burdens accumulating in the large intestinal lumen and mucosa result in the development of clinical disease which can manifest in several different forms 9 including ill-thrift and weight loss, with or without diarrhoea, 10,11 colic associated with caecal compromise [11][12][13][14] or a severe local and systemic inflammatory response syndrome termed acute larval cyathostominosis. [14][15][16] Acute larval cyathostominosis is associated with the emergence "en masse" of previously encysted larvae from the mucosa of the colon and caecum resulting in severe typhlocolitis. Acute larval cyathostominosis has a high fatality rate with up to 50% mortality reported. 13,14,16 In the literature, acute larval cyathostominosis is noted as having an obscure pathogenesis with reference to many risk factors age, season and period since last anthelmintic treated 17 but also altered host immunity, stress, dietary changes and high intestinal helminth burden. 10,16,18 It most commonly develops in late winter/ early spring. 13,16,17,19 However, not all horses with recognised risk factors develop the disease and not all horses who develop acute larval cyathostominosis have these risk factors. 10,20 Therefore, it is recognised that other, as yet unknown, factors contribute to individual susceptibility to acute larval cyathostominosis.
The immunoregulatory effects of helminths in a wide variety of host species (human and animal). commonly involve a combination of Th2 and Treg responses (reviewed in 21 ). No such studies have yet been carried out in horses. Furthermore, the treatment of helminth infections has been shown to influence the inflammatory dynamics of the host's immune response. [22][23][24][25] Although not many studies have evaluated the effects of cyathostomins on the immune response in horses, there is some evidence of a similar immunoregulatory response. 26 In fact, it has already been hypothesised that removal of adults could restore a previously suppressed immune response 10 and it has been shown that anthelmintic treatment in horses with a substantial helminth burden is associated with an inflammatory response. [27][28][29] Moreover, there is growing evidence that helminths interact with the intestinal microbiota in horses. [29][30][31] Dysbiosis of the microbiota has been shown to be associated with colitis 32 and to precede colic. 33 The role of intestinal microbiota in the development of acute larval cyathostominosis has not been well-described.
Here, we describe an outbreak of acute larval cyathostominosis during November and December 2018 in a herd of horses on an equine rescue facility. The clinical presentation, clinicopathological parameters and gut microbiota analysis of the clinically affected horses, their clinically normal pasture mates and the post-mortem findings on the non-surviving cases are presented. There has been no case series to date that has outlined the changes in the faecal microbiota associated with this disease. This report suggests potential new approaches to understanding the complex tripartite relationship between the equine faecal microbiota, the helminth population (helminthome) and host immune responses in acute larval cyathostominosis.

| Study population
This outbreak occurred in the midlands of Ireland, in mid-November 2018. The study population consisted of 23 mixed breed horses, mostly made up of cob ponies, large cobs and ponies, with ages ranging from 1.5 years to 12 years of age, 16 males, and 7 females. The horses had been rescued from various locations around the midlands by a charity organisation and it was planned to rehome them in time. Animals had been individually introduced into the herd over a 2-year period prior to the outbreak (range: 2 weeks to 2 years). The group were kept on a 15acre plot of land that was considered an out-farm. They were fed hay from the ground, as little grass was available, and had free access to water. The horses in this herd were selected for the pasture because they were considered "good doers" i.e. maintain and gain weight easily.
Horses were introduced and removed from the herd regularly without quarantine protocols. No pasture faecal decontamination processes were employed, and no co-grazing practices with other species were used. Anthelmintic treatment prior to the outbreak was intermittent and inconsistent; a variety of anthelmintic drugs had been used from 12 months to 3 weeks prior to onset of clinical signs.
Due to the high level of pasture contamination, the time of year and the age demographic of the horses, these horses were deemed an at risk population for acute larval cyathostominosis. 17 Therefore, horses were suspected of having clinical cyathostominosis if they presented with the following clinical signs that were shown to be associated with acute larval cyathostominosis, acute weight loss, diarrhoea/soft faeces, pyrexia, dullness and colic. 13,14,16,18,19,34 The horses showing clinical signs were moved to individual stables on the main farm to prevent further exposure to infective larvae and to facilitate monitoring and treatment. The rest of the herd remained on pasture due to restricted land availability despite the risks associated with such highly contaminated pasture.

| Treatment and clinical monitoring
Clinically affected horses were treated on the main farm with anthelmintics, anti-inflammatory drugs and supportive therapy as appropriate for each case. They were monitored continuously in their individual stables on the main farm. Monitoring included estimated feed and water intake, faecal consistency, demeanour and daily rectal temperature measurement. The animals had a full clinical examination performed by a veterinary practitioner on weekly visits or more frequently when deemed necessary due to deterioration of clinical status.

| Haematology and biochemistry
Blood samples were taken by jugular venepuncture into plain, EDTA, lithium heparin or sodium citrate Vacutainer® tubes as appropriate.
Samples were collected from horses when clinical findings consistent with acute larval cyathostominosis first presented and thereafter were collected weekly from affected horses to monitor response to treatment. Blood samples were also taken from clinically normal horses under five years old that were co-grazing on the out-farm, as these were considered an at risk population. 17 Haematological analysis was performed on blood samples collected in EDTA. Total protein, albumin and globulin concentration were measured in blood samples collected in lithium heparin. Fibrinogen concentration was measured in blood samples collected in sodium citrate and ELISA for encysted cyathostomins in samples collected in plain Vacutainer® tubes.

| Small redworm blood test
Blood samples were left to coagulate at room temperature for 24 hours prior to centrifugation at 1000g for 10 minutes. The serum was removed and stored at −20°C. A commercialised cyathostominspecific enzyme linked immunosorbent assay (ELISA) (Austin Davis Biologics Ltd) was performed as per manufacturer technical guidelines (www.austi ndavis.co.uk). This test detects IgG (T) antibodies specific to a combination of larval cyathostomin antigens from three common cyathostomin species. 35 Results are reported as 'Serum Scores' which are relative concentrations of specific IgG (T) derived from ELISA absorbance and the use of ELISA calibration curves.
Serum scores are reported together with statistically derived probabilities (using logistic regression models) that a horse is infected with a cyathostomin burden greater than a given threshold.

| Faecal sampling
Faecal samples were taken from horses who presented with clinical signs of acute larval cyathostominosis and from pasture mates, which were of similar age and clinically normal.

| Faecal sample preparation and 16S rRNA gene sequence analysis
Faecal samples for microbiota analysis were processed as described previously. 29 Briefly, faecal samples for microbiota analysis were ho- The sequences obtained were filtered on the basis of quality (removal of low-quality nucleotides at the 3' end and removal of window 10 nt with low average quality) and length (removal of sequences with less than 200 bp with prinsEquation 37 , and the paired-end reads with a minimum overlap of 50 bp were joined using Fastq-join. 38 Finally, all single files were processed to a final filtering sequence mean quality score >25 bp. The sequences were cleaned of replicates, and unique sequences and chimeras were checked against the GOLD database (https://gold.jgi.doe.gov) using the closed reference Usearch v7.0 algorithm. 39 The resulting sequences were matched at operational taxonomic unit level (OTU; with 97% identity level) using UPARSE-OUT algorithm with Usearch v7.0 algorithm. The taxonomic assignment of these OTUs was matched to results in the Ribosomal Database Project. 40 Alpha and Beta diversities were determined using QIIME. 41 Beta-dispersion was quantified with betadisper (vegan::betadisper) and additional analyses were performed with the R package phylosEquation. 42

| Post-mortem examination
Post-mortem examination was performed on horses who were euthanised. Intestinal tissue samples from the ventral and dorsal colon, caecum and small intestine, were fixed in 10% neutral buffered formalin. Following fixation, tissues were embedded in paraffin wax, sectioned at 4μm and stained with Gill®-2 Haematoxylin and Eosin (H&E) for histopathological examination.

| Data analysis
Statistical analyses were performed and graphical outputs of 16s rRNA gene high throughput sequencing data were generated with various packages in R. 42 For beta-diversity analysis, dissimilarity matrix between samples was calculated with the Bray-Curtis method transforming the data into a logarithmic scale, studying the effects on microbiota composition. Variability between samples by Permutational Multivariate Analysis of Variance was calcu- Student T-tests were used to determine differences in haematological parameters or plasma protein concentrations between the clinically normal horses and the clinically affected horses at first presentation Paired T-tests were used to investigate changes in these parameters in repeated samples in the treated horses in the weeks after treatment. Statistical significance was established at P < .05.

| Clinical presentation
Of the 23 horses grazing on the out-farm where the outbreak occurred, 14 showed overt clinical signs consistent with acute larval cyathostominosis (9 males and 5 females; 1.5-6 years old). Clinical signs appeared in 11 of the horses (8 males and 3 females; 1.5-6 years old) over a period of one week. Three further horses presented with acute signs four weeks later (1 male and 2 females; 3-4 years old) ( Figure 1).
There were a variety of clinical signs, with the most common being F I G U R E 1 Schematic outline of the study population and timeline of outbreak. "Clinical" -horses who presented with clinical signs. "Not sampled" -horses who were not considered in the at-risk cohort. "Clinically normal" -horses that were considered at risk and had clinicopathological findings of chronic inflammation. "On pasture" -before outbreak, "November" and "December" timeline during the outbreak [Colour figure can be viewed at wileyonlinelibrary.com]

| Anthelmintic treatment history
The prior anthelmintic treatment of the herd was inconsistent (Table S1). However, the seven most severely affected horses had been treated with an ivermectin or a combination of ivermectin/ praziquantel product three weeks prior to first observation of clinical signs (Table S1). Three horses that had weight loss only had been dosed with ivermectin 12 weeks prior to presentation.

| Treatment
The first four horses to present were initially treated by a veterinary practitioner who was not part of the study team, with a single dose (10 mg/kg) bwt of fenbendazole and a single dose (25 gs) of a trimethoprim/sulphonamide product (Table S1). This treatment was immediately discontinued to allow for assessment of the horses but occurred one day prior to faecal sampling in   (Table S1). Three of these horses presented with weight loss 26 days later, blood samples were taken at first clinical presentation and were included in the clinically affected group.

| Haematology and biochemistry
Consistent findings in blood samples from all clinically affected and clinically normal horses included borderline anaemia, leucocytosis, thrombocytosis, lymphocytosis, hyperfibrinogenaemia, hyperglobulinaemia and a reverse albumin: globulin (A: G) ratio (Table S2). Clinically affected horses had higher neutrophil counts (P = .01) and lower albumin (P = .002) and total serum protein (P = .02) concentrations than clinically normal horses. Total serum protein concentrations, however, remained within the reference range. The more severely affected cases presented with neutrophilia with a left shift and monocytosis.
Following treatment, the left shift took up to two weeks to resolve.
There was no significant difference in total white blood cell or neutro-

| Small redworm blood test
Six of the clinically affected horses were included in this analysis. All horses had very high Serum Scores, ranging from 53.7 to 70.9 therefore indicating a 76.9% to 91.7% (respectively) probability that the horses had > 10 000 total worm burden 35 (www.austi ndavis.co.uk).

| Faecal egg counts
Faecal samples were obtained at presentation from all clinically affected horses (except horse 11) and four clinically normal pasture mates. Only one clinically affected horse had a faecal egg count of over 200epg at presentation (Table S1). In the horses treated with moxidectin, FEC were also taken two weeks after treatment to ensure efficacy of treatment and were negative.

| Faecal microbiota analysis
Prior to the implementation of the authors' treatment protocol, fae- In order to ensure that beta diversity estimates were not affected by the administration of antibiotics, we excluded the three animals that had received treatment prior to sampling from this analysis. We also included sex as a possible variance factor before investigation the clinical factor of interest.
Beta-diversity measurements showed a significantly in-

| Outcome
In

| D ISCUSS I ON
We report here on an outbreak of acute larval cyathostominosis in In this outbreak, as in other cases in the literature, there was a lack of pasture hygiene practices such as faecal removal, rest periods or cross grazing. 15,19 This pasture had more than the recommended one horse per 1-2 acres 45 with up to 25 horses on the 15 acre pasture at times. Furthermore, although these horses were in good condition, they were originally rescued on welfare grounds, often due to lack of concern for the animal's basic requirements including preventative healthcare. Therefore, it is likely that these horses would be at risk of accumulating high cyathostomin burdens prior to arrival on farm.
Additionally, no consistent anthelmintic regime had been in place. Although acute larval cyathostominosis can occur in regularly dewormed horses, 15,19,20  There is growing concern, due to the decrease in egg reappearance time after seemingly effective moxidectin treatment (>95% FECRT), that resistance is building in some species. [46][47][48][49] In the past growing resistance to anthelmintics in cyathostomins was associated perceived increase in the incidence of ALC. 14 present. This has been reported previously, with an improvement in horses after initiation of treatment followed by a worsening or reoccurrence of signs and death in some cases. 10,15,53 We suggest that in this herd, horses were all affected with chronic cyathostominosis, but that individual horses suffered from different intensities of disease, with some remaining clinically normal. Those horses that presented with severe clinical signs were either overwhelmed by infectious burdens or had experienced a trigger to induce synchronised larval re-emergence and an acute severe typhlocolitis.
The cause of the severe post-mortem changes in the two horses which were euthanised is unclear. The mucosal changes are in general similar to those described in other reports. 16,19,53 However, the bacterial over-growth has not been previously described and is an indication that a disruption in the bacterial population could be a contributing factor to the deterioration of some cases. The severe disruption of the mucosal barrier in combination with the profound inflammatory response and bacterial overgrowth provides a reason for the endotoxic signs associated with the most severely affected horses.
Corticosteroids in conjunction with anthelmintic treatment, most often moxidectin, are recommended for the treatment of larval cyathostominosis, together with supportive therapy. 18,45,52,55,56 In addition to anthelmintic treatment with moxidectin, corticosteroids were administered to clinically affected horses and tapered slowly according to clinical response, and we did not observe any adverse effects attributable directly to the steroid use. Notably two horses, which showed only mild transient diarrhoea, did not receive corticosteroids; one of these horses re-presented within four weeks with signs of acute severe weight loss. A combination of moxidectin and corticosteroids appears to be safe and effective in the treatment of horses suspected to be suffering from cyathostominosis and at risk of acute larval cyathostominosis.
In this study, we considered it appropriate to administer broad spectrum antibiotics to individual high-risk cases i.e. persistent in the horse. In addition, beta-diversity analysis showed an obvious increased dispersion of the clinically affected horses, reflecting the dramatic changes to their microbiota when compared to their clinically normal co-grazing herd-mates. Although three of these horses received antibiotic and anthelmintic treatment the day before sampling, these changes were also evident in the other clinically affected horses.
In horses with cyathostomin burdens, anthelmintic treatment is primarily effective against adult worms in the gut lumen, and removal of the adults is likely to break a negative-feedback loop that is suspected to prevent hypobiotic larvae resuming development and emerging into the lumen. 8,64 While some emergent larvae will be killed by anthelmintic treatment, we hypothesise that the dominant effect in this situation is a concatenation of inflammation arising from removal of adult cyathostomins and a wave of emerging larvae, whilst also disturbing the delicate balance between the helminth and host, and in particular the potentially beneficial anti-inflammatory influence that prevails in steady state. The inflammatory profile of the horses in this acute larval cyathostominosis outbreak, with features associated with gut microbiota dysbiosis as well as intestinal and systemic inflammatory responses, is consistent with this hypothesis.
The differences between communities of bacteria associated with clinically affected and clinically normal horses assessed using the LefSe algorithm, were interesting but not conclusive. Prevotella is considered a commensal and is associated with the colonic microbiota community. 65  and it more commonly presents with lethargy, anorexia and fever, less frequently with GIT related signs. 76 The faecal microbiota has been shown to reflect the large intestinal microbiota, but of course does not represent changes directly. 67 However, faecal sampling is readily accessible and non-invasive, and thus a valuable tool in investigations of disorders of the equine gut, and particularly the large intestine. Three of the clinically affected horses had been treated with one dose of antibiotics and anthelmintics one day prior to initial faecal sampling. We acknowledge that this treatment may have perturbed the faecal microbiota in these animals.

| CON CLUS ION
Together with our previous study 29 and findings described here of the severe typhlocolitis, marked weight loss and endotoxaemia seen in some of the cases in this outbreak, we suggest that the systemic inflammation and acute clinical syndrome that is seen in acute larval cyathostominosis could arise from the "perfect storm" of an inflammatory stimulus of numerous emerging larvae and dysbiosis of the gut microbiota leading to structural and functional pathology of the large intestine. Further investigation into the relationship between cyathostomins, the gut bacterial microbiota and inflammatory changes in normal, sub-clinical and clinically affected horses will be useful in developing biomarkers that can be used for risk assessment and diagnosis.

E TH I C A L A N I M A L R E S E A RCH
Research ethics committee oversight not required by this journal: retrospective analysis of clinical data.

OWN E R I N FO R M E D CO N S E NT
The data used in this study was collected form clinical samples taken during the course of the disease outbreak.

ACK N OWLED G EM ENTS
We thank My Lovely Horse Rescue for their assistance and cooperation during this outbreak and their ongoing support in our research into parasite related disease. We are grateful to Austin Davis Biologics Ltd for testing the serum samples in the commercialised small redworm blood test.

CO N FLI C T O F I NTE R E S T S
No competing interests have been declared.