Identification of potentially zoonotic parasites in captive orangutans and semi‐captive mandrills: Phylogeny and morphological comparison

Cysts and trophozoites of vestibuliferid ciliates and larvae of Strongyloides were found in fecal samples from captive orangutans Pongo pygmaeus and P. abelii from Czech and Slovak zoological gardens. As comparative material, ciliates from semi‐captive mandrills Mandrillus sphinx from Gabon were included in the study. Phylogenetic analysis of the detected vestibuliferid ciliates using ITS1‐5.8s‐rRNA‐ITS2 and partial 18S ribosomal deoxyribonucleic acid (rDNA) revealed that the ciliates from orangutans are conspecific with Balantioides coli lineage A, while the ciliates from mandrills clustered with Buxtonella‐like ciliates from other primates. Morphological examination of the cysts and trophozoites using light microscopy did not reveal differences robust enough to identify the genera of the ciliates. Phylogenetic analysis of detected L1 larvae of Strongyloides using partial cox1 revealed Strongyloides stercoralis clustering within the cox1 lineage A infecting dogs, humans, and other primates. The sequences of 18S rDNA support these results. As both B. coli and S. stercoralis are zoonotic parasites and the conditions in captive and semi‐captive settings may facilitate transmission to humans, prophylactic measures should reflect the findings.

Vertebrate Biology, Czech Academy of Sciences, Czech Republic, Grant/Award Number: RVO:68081766 Morphological examination of the cysts and trophozoites using light microscopy did not reveal differences robust enough to identify the genera of the ciliates.
Phylogenetic analysis of detected L1 larvae of Strongyloides using partial cox1 revealed Strongyloides stercoralis clustering within the cox1 lineage A infecting dogs, humans, and other primates. The sequences of 18S rDNA support these results. As both B. coli and S. stercoralis are zoonotic parasites and the conditions in captive and semi-captive settings may facilitate transmission to humans, prophylactic measures should reflect the findings.  (Islam et al., 2022;Labes et al., 2011;Sak et al., 2013). The close phylogenetic relationship between humans (Homo sapiens) and other primates further escalates the risk of potential pathogen transmission (including GI parasites). Such transmission occurs easily in conditions of close human-other primates contact, especially in captivity (Keita et al., 2014). In the wild, the risk of primate infection by pathogens originating from humans increases with increasing human pressure (Dunay et al., 2018;Kooriyama et al., 2013).
Proximity to other animal species in captive settings may facilitate transmission of parasites with low host specificity and these animals may serve as reservoirs of infection for primates (Modrý & Escalante, 2018). Additionally, limited space in captive settings leading to close contact with animals creates opportunities for parasite transmission (Sanchez & Pich, 2018).
Balantioides coli is classified as an intestinal commensal (Schuster & Ramirez-Avila, 2008), but can be pathogenic under certain conditions and cause clinical symptoms in humans and captive primates (Schovancová et al., 2013;Strait et al., 2012). In primates, infection may be asymptomatic or accompanied by dehydration, bloody or mucous diarrhea, weight loss, anorexia, lethargy, tenesmus, and rectal prolapse (Cockburn, 1948;Strait et al., 2012;Teare & Loomis, 1982). Some cases of balantidiosis can be even fatal (Lankester et al., 2008;Strait et al., 2012). The intensity of infection can be affected by a diet rich in starch (Schovancová et al., 2013) and also by stress, which generally leads to immunosuppression (de Oliveira et al., 2022). Balantioides coli is transmitted directly between individuals via the fecal-oral route, but one of the main sources of the infection can be domestic pigs (Sus scrofa domestica) and wild boars (Sus scrofa) (Schuster & Ramirez-Avila, 2008). Despite some knowledge about B. coli in primates, no information is available about whether infections of Buxtonella-like ciliates can manifest clinically in primates.
The genus Strongyloides (Nematoda: Chromadorea, Rhabditida) includes at least 50 species of parasitic rhabditid nematodes (Speare, 1989). Strongyloides stercoralis, S. fuelleborni, S. cebus, and several unidentified Strongyloides spp. are known to infect primates; however, the species greatly differ in their host spectrum (Bradbury et al., 2021). While S. stercoralis has a broad host specificity and has been detected in a range of hosts including humans, carnivores, and (mainly captive) primates, S. fuelleborni is restricted to primates (mostly free-ranging) with occasional spillover to humans (Bradbury et al., 2021). Studies using genotyping concluded that S. stercoralis likely originated in dogs (Canis lupus familiaris) and later adapted to human and other primate hosts (Barratt et al., 2019;Jaleta et al., 2017). The most commonly used markers for molecular analysis of Strongyloides include two hypervariable regions (HVR) of the 18S rDNA gene (HVR-I and HVR-IV) and a selected portion of the mitochondrial cytochrome c oxidase subunit 1 gene (cox1). These markers have been compared in different hosts and geographic regions, and haplotypes of the HVR-IV of 18S rDNA have been found to match cox1 haplotypes (Bradbury et al., 2021). A few studies have used 28S rDNA and parts of ITS genes for Strongyloides detection (e.g., Labes et al., 2011;Solórzano-García & Pérez-Ponce de León, 2017), but the number of such sequences is rather small and therefore these markers are usually not used for further analyses.
Strongyloides infection can be asymptomatic, while uncomplicated disease is manifested by GI, pulmonary, and dermatological symptoms.

In immunocompromised human patients, infection can result in a serious
systemic disease with fatal consequences Nutman, 2017). There are no reports on the clinical outcomes of Strongyloides infections in free-ranging apes or other primate species, but reports describe fatal cases in captive animals Strait et al., 2012). Kleinschmidt et al. (2018) reported fatal strongyloidiasis in a 5-month-old Sumatran orangutan (P. abelii). A fatal course of infection has been also reported in young chimpanzees (P. troglodytes) and gorillas (G. gorilla) (Penner, 1981). Orangutans below the age of 5 years appear to be the most prone to clinical disease (Labes et al., 2011;McClure et al., 1973;Uemura et al., 1979).
Our study aimed to survey a spectrum of parasites found in captive Bornean orangutans (P. pygmaeus) and Sumatran orangutans (P. abelii) in Czech and Slovak zoological gardens. Special attention was paid to potentially zoonotic GI parasites, specifically Strongyloides spp. and vestibuliferid ciliates. We further determined if the orangutans were infected with zoonotic vestibuliferid ciliates B. coli or probably nonzoonotic Buxtonella-like ciliates and identified Strongyloides species/haplotypes. As comparative material for the vestibuliferid ciliates, we used fecal samples from semi-free ranging mandrills (Mandrillus sphinx) from Gabon. First, we predict that the dominant detected GI parasites in studied captive orangutans from Czech and Slovak zoological gardens are nematodes of the genus Strongyloides and vestibuliferid ciliates, since these parasites were most frequently found in captive orangutans (Nurcahyo et al., 2017). Second, we assumed that potentially zoonotic B. coli and S. stercoralis are present in captive orangutans studied, since these species have already been identified in other captive primates using molecular analyses (Nurcahyo et al., 2017;Pomajbíková et al., 2013). Last, if wild primates of the family Cercopithecidae harbor potentially nonzoonotic Buxtonella-like ciliates Yan et al., 2018), we predict that these ciliates might be present in semi-free living mandrills. Detection and exact identification of parasites in captive settings is important, as unattended infections can cause serious health problems or even death in mostly young, old and weak individuals (Sanchez & Pich, 2018). Moreover, parasites with zoonotic potential may pose a risk to keepers and other staff working with captive animals. Detailed information about the parasites of captive animals is crucial to evaluate their impact on the health of free-ranging individuals, which is necessary for good conservation management.  were selected for further DNA barcoding and phylogenetic analyses. Individual Strongyloides larvae (N = 15) were collected from orangutan fecal samples (N = 4) after the Baermann larvoscopy method (Pafčo, 2018), placed in microcentrifuge tubes and frozen. Invisorb ® Spin Forensic Kit (STRATEC) was used to isolate DNA from individual Strongyloides larvae following the manufacturer's protocol. Three genetic markers were selected for amplification: cox1 and two HVR of 18S rDNA (HVR-I and HVR-IV) according to Barratt et al. (2019). Two primer pairs: TJ5207 forward and TJ5208 reverse (Jaleta et al., 2017) and
Amplification of all targeted Strongyloides and vestibuliferid ciliates markers was performed using Biometra T-personal thermocycler (Schoeller) under conditions specified in Table 1. The PCR products were separated by electrophoresis in a 1.5% agarose gel with Midori Green Advance (Nippon Genetics) and visualized on a UV transilluminator. The PCR products of the expected size were purified either directly or after cutting the band from the gel using Gel/PCR DNA Fragments Extraction Kit (Geneaid Biotech). All amplicons were Sanger sequenced in both directions commercially by Macrogen Europe (Amsterdam, Netherlands). If we detected mixed signals, the PCR product was cloned using pGEM ® -T Easy Vector Systems (Promega Corporation) and GenElute™ Plasmid Miniprep Kit (Sigma-Aldrich). The resulting cloning products were also sent for Sanger sequencing. All obtained sequences were checked and manually trimmed in Geneious 9.1.5 (www.geneious.com) and the identity was checked using BLAST (Altschul et al., 1990).

| Phylogenetic analyses
All sequences obtained were aligned with sequences from GenBank using Muscle alignment in Geneious 9.1.5 (Kearse et al., 2012).
For the 18S rDNA of vestibuliferid ciliates, the alignment (1629 bp) consisted of seven sequences obtained during the present project and T A B L E 1 PCR primers and reaction conditions.

Substitution model GTR + G was used as it was selected by
Modelgenerator (Keane et al., 2006).

| Potentially zoonotic parasites detected in captive orangutans
In total, 14 fecal samples of individual orangutans from five Czech and Slovak zoos were microscopically examined. Four samples were negative for any parasites/commensals (  Country/location Sex Host species/sample ID Strongyloides stercoralis Balantioides coli Entamoeba coli Chilomastix sp. Giardia sp.

| Phylogenetic analysis of vestibuliferid ciliates
Seven sequences (1145-1170 bp) of partial vestibuliferid ciliates 18S rDNA were obtained (two from Bornean orangutans, one from Sumatran orangutan, and four from mandrill samples). All sequences from orangutans were identical. In mandrills, one sequence differed by 2.8%. Sequences from orangutans differed from mandrill-derived sequences by 3.3-4.2%.
The resulting BI tree showed 19 separate clades corresponding to individual genera ( Figure 2). All the isolates thought to be B. coli from orangutans clustered within a highly supported (1.00 BI) sub-clade within the genus Balantioides clade comprising mainly isolates from other captive primates: western lowland gorillas (Gorilla gorilla gorilla), chimpanzees (Pan troglodytes), a bonobo (P. paniscus), and a hamadryas baboon, a few sequences from pigs (Sus scrofa domestica), from an ostrich (Struthio camelus), a common rhea (Rhea americana), and a Guinea pig (Cavia porcellus) (Figure 2). The other Balantioides sub-clade comprised sequences mainly from pigs, an Asian elephant (Elephas maximus indicus), a giant anteater (Myrmecophaga tridactyla), and a chimpanzee. The pairwise sequence distance (PSD) within the Balantioides clade did not reach over 2.1%.
In contrast, the sequences originating from mandrills clustered within a highly supported (0.9 BI) clade comprising only sequences labeled as Buxtonella-like and originating from various cercopithecids.
One sequence from a mandrill (OM423619) clustered separately from all other isolates in a highly supported (1.00 BI) sub-clade ( Figure 2). The sequence differed from the sequences within the major sub-clade by 2.7-3.2%, while the PSD within the major subclade was below 0.3%.
Eleven sequences (420-514 bp) of ITS were obtained during the current study (two sequences from Bornean orangutans, five clones of an isolate from a Sumatran orangutan, and four sequences from mandrill samples). All sequences from Bornean orangutans were 100% identical. In the Sumatran orangutan, the cloned sequences differed by 0.2-1.6%. In mandrills, one sequence differed by 6%.
Sequences from orangutans differed from mandrill-derived sequences by 8.4-11%. The ITS BI tree topology was similar to that of the 18S rDNA but showed significant intraspecific variability,   Table S1 for details).

| Haplotypes and phylogenetic analysis of S. stercoralis
Four sequences of partial cox1 (263-627 bp) and four sequences of each 18S rDNA HVR-I (391-457 bp) and HVR-IV (235-281 bp) were obtained from individual S. stercoralis larvae extracted from orangutan fecal samples (two from P. pygmaeus and two from P. abelii). A BLAST search showed that all four sequences of HVR-IV corresponded to haplotype A in Barratt et al. (2019), while in HVR-I region, haplotype VI was detected in P. pygmaeus and haplotype II in P. abelii, following Barratt et al. (2019). The two S. stercoralis cox1 sequences from P. abelii were identical, while the two sequences from P.
pygmaeus differed at a single-nucleotide polymorphis site. The PSD between the cox1 sequences from the two orangutan species was 2%. Both BI and ML cox1 phylogenetic trees were well-resolved with highly supported nodes and yielded the same general topology from Bornean orangutans were placed in a different sub-clade than sequences from Sumatran orangutans and these two sub-clades differed by 1.7-3.6%. The PSD within the sub-clades comprising the Bornean orangutan sequences and Sumatran orangutan sequences was below 2.9% and 3.1%, respectively. All sequences of Strongyloides obtained in this study were uploaded to GenBank under the accession numbers OM423625-OM423632 (18S rDNA); OM392049-OM392052 (cox1) (see Supporting Information: Table S1 for details). Parasites of orangutans are overall much less studied than those of African great apes (Modrý & Escalante, 2018). Captive, semicaptive, and wild orangutans are commonly diagnosed with protists and helminths with relatively low host specificity often in combination with orangutan-specific parasites (Modrý & Escalante, 2018). We detected Giardia intestinalis, which is a typical parasite with low host specificity. Infections in primates can occur asymptomatically or with clinical symptoms such as diarrhea or vomiting (Strait et al., 2012) that were not observed during our study. We also detected E. coli and Chilomastix sp., which are also protists commonly occurring in great apes (Jirků-Pomajbíková & Vlčková, 2018;Kváč & McEvoy, 2018).
However, the most commonly detected GI parasites in captive and free-ranging orangutans are nematodes of the genus Strongyloides and vestibuliferid ciliates B. coli (Foitová et al., 2009;Nurcahyo et al., 2017), both of which were among the most common parasites in our study, consistent with our first hypothesis.
The common occurrence of ciliates identified as B. coli is reported in most of the microscopy-based studies focused on parasitofauna of orangutans (e.g., Kilbourn et al., 2003;Kuze et al., 2010;Labes et al., 2011;Mul et al., 2007). Studies of mandrills, the source of our comparative material, also mention detection of B. coli (Poirotte et al., 2016;Setchell et al., 2007). However, recent studies (Chistyakova et al., 2014;Pomajbíková et al., 2013;Yan et al., 2018)  Our phylogenies of the 18S rDNA and ITS regions showed almost identical results and clearly assigned the vestibuliferid ciliates of orangutans in the sub-clade A of B. coli, consistent with our second hypothesis. This was previously detected in various primates, including captive African great apes, several captive macaque species (Barbosa et al., 2017;Pomajbíková et al., 2013), and one isolate from a free-ranging mountain gorilla (Hassell et al., 2013), extending the spectrum of this sub-clade hosts to captive Bornean and Sumatran orangutans. In contrast, the ciliates from our comparative samples from mandrills clustered with Buxtonella-like ciliates from other cercopithecids, consistent with our last hypotheses.
Although many studies report the occurrence of B. coli in captive/semi-captive and free-ranging orangutans (Nurcahyo et al., 2017), supporting molecular data is lacking. Here, we, for the first time, performed molecular-phylogenetic analyses and confirmed the occurrence of B. coli in captive Bornean and Sumatran orangutans. In the past, it was impossible to distinguish whether orangutans (including captive, semi-captive, and wild) are infected by Buxtonella-like or B. coli. Balantoides coli is often detected in captive primates, including great apes (Pomajbíková et al., 2010).
In contrast, Buxtonella-like ciliates seem to be restricted to primates of the family Cercopithecidae both in captivity and in the wild Yan et al., 2018), which is supported by our results for semi-captive mandrills. So far, it appears that B. coli is less host-specific and may occur in various primates as well as in humans, while Buxtonella-like infects mostly primates other than great apes. Implementation of molecular phylogeny is necessary to clarify the distribution and host specificity of the two vestibuliferid ciliates in wild, semi-captive and captive primate populations.
The occurrence of potentially pathogenic and zoonotic B. coli in captive primates in zoos and occasionally in sanctuaries is noteworthy and deserves further attention. The parasite appears to be rarely detected in free-ranging great apes (Hassell et al., 2013), whereas in captivity, there may be factors that increase the susceptibility of primates to infection. In case of B. coli, there is a positive correlation recorded between the number of ciliates in fecal samples of captive chimpanzees and starch content in the diet (Schovancová et al., 2013). Optimal nutrition of great apes in (semi)captive settings is of utmost importance to prevent metabolic diseases, dental problems, or abnormal behavior patterns (Edwards & Ullrey, 1999;Kuhar et al., 2013;Plowman, 2013). Therefore, the role of diet in orangutans infected with B. coli should be investigated in future studies.
Sequencing confirmed the existence of the two types of ciliates in our sample set. However, the question arises as to the identification of the two taxa based on their morphological characteristics observed under the microscope. In general, the cysts formed by vestibuliferid ciliates vary in size and shape even within a species, which complicates their identification (Kuze et al., 2010;Mul et al., 2007;Pafčo et al., 2018;Yan et al., 2018). Although the cysts appeared slightly different in microscopy (see large vacuole in the cysts of Buxtonella-like vestibuliferids, Figure 1c), we did not find robust features that distinguish between B. coli and Buxtonella-like ciliates in primate fecal samples using microscopic analyses.
From the clinical perspective, strongyloidiasis is the most important parasitosis of captive and semi-captive orangutans (Kleinschmidt et al., 2018;Labes et al., 2011). It is assumed that wild primates are infected with S. fuelleborni, while S. stercoralis is typical for captive animals, and is probably transmitted to primates from other susceptible hosts (Bradbury et al., 2021). A study providing exact Strongyloides species determination in free-ranging, semi-captive, and captive Bornean orangutans using ITS and 18S rDNA analysis revealed that free-ranging animals were infected with S. fuelleborni, while S. stercoralis was identified only in captive individuals (Labes et al., 2011).
Moreover, Frias et al. (2018) found only S. fuelleborni in free-ranging Bornean orangutans in the Lower Kinabatangan Wildlife Sanctuary, Sabah, Malaysia. In contrast, S. fuelleborni has never been found in orangutans kept in zoological gardens. However, the transmission of S.
fuelleborni between primates and animal caretakers in captive conditions has been demonstrated between Southern pig-tailed macaques (Macaca nemestrina) and their owners in Southern Thailand (Janwan et al., 2020) and transmission is also very possible in case of S. stercoralis between captive primates and the staff and vice versa (Bradbury et al., 2021).
In all orangutans from both Slovak and Czech zoos, S. stercoralis was detected by coproscopic detection of L1 larvae and amplification of the partial sequence of cox1, HVR-I, and HVR-IV regions of 18S rDNA confirmed its identity, which supported our hypothesis. Based on cox1, two lineages A and B of S. stercoralis are distinguished, with lineage A occurring in dogs, humans, and other primates, while lineage B has been recorded in dogs only (Bradbury et al., 2021). All cox1 sequences obtained in the present study clustered within the lineage A (highlighted in dark blue in Figure 3). Similarly, only haplotype A of HVR-IV, so far detected in humans, chimpanzees, and dogs (Bradbury et al., 2021) was detected in our samples. In contrast, we found haplotypes II and VI of the HVR-I. Although the significance of the HVR-I is not yet known, our results indicate possible interspecific variability. In general, our results support the hypothesis that S. stercoralis occurs not only in dogs and humans, but also in Phylogenetic tree derived by Bayesian inference and maximum likelihood for Strongyloides spp. cox1 (722 bp) sequences obtained in this study (in bold) and available in the GenBank database. The tree was constructed from a muscle alignment and calculated using the GTR + G model for nucleotide substitutions. Branch lengths indicate expected numbers of substitutions per nucleotide site. Numbers at the nodes show posterior probabilities under the Bayesian inference/bootstrap support values for maximum likelihood. Nodal supports <50 are marked with an asterisk. A total of 18S rDNA haplotypes are marked in the tree. captive primates, whereas S. fuelleborni is restricted to free-ranging primates.
We detected zoonotic parasites with possible pathogenic potential, S. stercoralis and B. coli, in Bornean and Sumatran orangutans kept in several Czech and Slovak zoological gardens.
Application of molecular, sequencing, and phylogenetic tools for exact identification of vestibuliferid ciliates is crucial as morphological distinction is difficult. We provided sequence data for S. stercoralis from captive orangutans and contributed to knowledge of the genetic diversity of the parasite as data from primates are very rare in public databases (Bradbury et al., 2021). Although none of the animals exhibited any clinical signs, these infections can lead, under some circumstances, to serious illness and even to death (Strait et al., 2012).
Parasite monitoring is crucial for institutional health programs and reducing the risk of transmission to humans working with captive primates. Effective management, which includes monitoring of parasitic diseases, is a paramount component of the preventative medicine program of any facility that wants to reduce the risk of clinical illness in animals as well as the risk of zoonotic transmission (Sanchez & Pich, 2018). Regular parasite screening provides a base for targeted treatment, while regular and blind application of antiparasitic drugs can lead to resistance to the products and reduce their effect in the future, as known in livestock (e.g., Geurden et al., 2015).

ACKNOWLEDGMENTS
The authors thank all zoological gardens for providing the material, and especially to curators, zookeepers, and veterinarians who participated on the sample collection. We also thank the animal care staff at the CIRMF primate center for their help. This project was

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.