Aspects of Genetic Diversity, Host Specificity and Public Health Significance of Single‐Celled Intestinal Parasites Commonly Observed in Humans and Mostly Referred to as ‘Non‐Pathogenic’

Authors: Christen Rune Stensvold

Abstract

Clinical microbiology involves the detection and differentiation of primarily bacteria, viruses, parasites and fungi in patients with infections. Billions of people may be colonised by one or more species of common luminal intestinal parasitic protists (CLIPPs) that are often detected in clinical microbiology laboratories; still, our knowledge on these organisms' impact on global health is very limited. The genera Blastocystis, Dientamoeba, Entamoeba, Endolimax and Iodamoeba comprise CLIPPs species, the life cycles of which, as opposed to single‐celled pathogenic intestinal parasites (e.g., microsporidia and sporozoa), do probably not include gut‐invasive stages that could result in pathological processes and thereby disease (except for Entamoeba histolytica). All five genera are parasites in the sense that they use a host to complete their life cycle; still, by many specialists, these are considered to be of limited clinical relevance and could possibly be referred to as ‘eukaryotic endobionts’ or even ‘endosymbionts’, in case they would have health‐protective effects. The articles included in this thesis exemplify the work and the data that support the view that it might be more relevant to study these genera in a public health and gut ecology context than in a clinical microbiology context. Essential to investigating the impact of intestinal parasites on health and disease are accurate diagnostic tools, including DNA‐based technology such as PCR and sequencing, plus accurate reference databases. Small subunit (SSU) ribosomal RNA (rRNA) genes consistently present in both pro‐ and eukaryotic organisms are today avidly used as taxonomic markers. DNA‐based methods have been developed for genetic characterisation of microorganisms and provided data on species/subtypes/genotypes, etc. Metagenomics and metabarcoding (the use of low‐specific PCR coupled with next‐generation sequencing) can provide information on co‐infection/co‐colonisation with other organisms and enable screening for genetic diversity, even in complex matrices. By developing and implementing sensitive and specific DNA‐based diagnostic tools and typing assays primarily based on the SSU rRNA gene, we have increased insight into the diversity, distribution and significance of CLIPPs. With these tools, we have shown that the genera Blastocystis and Dientamoeba are far more common than previously thought. Only 10–15 years ago, hypotheses on their distribution typically relied on data generated by traditional parasitological diagnostic methods, such as light microscopy. Hence, we have shown that most older children in Nigeria host Blastocystis, and that most children in day‐care institutions in Denmark, if not all, get colonised by Dientamoeba at some point. Single‐celled non‐pathogenic intestinal parasites can be hosted by patients with diarrhoea and functional or inflammatory bowel diseases. However, emerging data appear to suggest that CLIPPs are generally more common in gut‐healthy individuals than in patients with gastrointestinal symptoms. The research we have carried out on associations between CLIPPs and gut bacteria suggests that colonisation with these parasites is seen primarily in individuals with a healthy ‘gut flora’ (eubiosis). This observation should prompt future research projects focusing on the use of CLIPPs as biomarkers, and it should be investigated to which extent manipulation with CLIPPs could lead to changes in the gut flora and thereby be used as probiotics. In the event that it makes sense to speak of ‘infection’ by CLIPPs, we still lack tools to differentiate between colonisation and infection. We have known for decades that morphologically similar parasites can differ in terms of clinical impact and be genetically distinct, a feature that we refer to as ‘cryptic genetic diversity’. One example is E. histolytica, which cannot be differentiated from Entamoeba dispar by cyst morphological features. However, whereas E. histolytica can be invasive and give rise to amoebic dysentery and amoebiasis, E. dispar is by most specialists considered non‐invasive and generally non‐pathogenic. This insight led us to investigate genetic diversity among other species of Entamoeba as well as other CLIPPs genera. If we could demonstrate similar—or higher—degrees of diversity within Blastocystis, Dientamoeba, Endolimax and Iodamoeba, these differences might be key to explaining differences in parasite phenotype and thereby differences in the ability of the parasites to cause symptoms. Despite the disclosure of striking genetic diversity among some CLIPPs, we have found little support for such theories; however, more studies are needed. As for Dientamoeba, we have observed a more or less clonal expansion of one of the two genotypes known to exist, and this genotype appears to have global predominance. In contrast, extensive genetic diversity is observed between and within subtypes of Blastocystis: to date, more than 30 species, the so‐called subtypes, have been acknowledged. We, and many others, have sought to identify whether one or more of these subtypes could be linked to the development of intestinal symptoms, but there is little evidence to support this hypothesis. We know that Subtypes 1–4 reflect about 95% of Blastocystis colonisation in humans, and we have shown that individuals with zoonotic subtypes (e.g., ST6, ST7 and ST8) might typically experience symptoms. We have disclosed astonishing genetic variation among other CLIPPs, which has led to the recognition of Iodamoeba bütschlii, Endolimax nana, Entamoeba coli and Entamoeba hartmanni as species complexes, where each species should be regarded as a complex of species (referred to as ‘subtypes’ or ‘ribosomal lineages’) with overlapping morphology. And where E. histolytica and E. dispar differ by only 1%–2% diversity across the SSU rRNA gene, we have observed up to at least 10% and 30% genetic difference among ribosomal lineages within E. coli and I. bütschlii, respectively, challenging species concepts currently applied. Our research has resulted in the recognition of three ribosomal lineages within both E. coli and E. hartmanni, as well as two ribosomal lineages of E. nana and I. bütschlii. Moreover, we have discovered a new lineage of Entamoeba moshkovskii. Molecular characterisation of intestinal parasites collected from different host species (humans, non‐human primates, other mammals, birds, etc.) can help identify opportunity for transmission between human and non‐human hosts. We have shown that pigs can host a few species/lineages that can readily colonise humans, such as Entamoeba hartmanni and I. bütschlii. Many other species of larger mammals are common hosts of Blastocystis and Entamoeba. However, for the two latter genera, the species/genetic variants observed in non‐human hosts are typically different from those observed in humans, which could indicate that many species of CLIPPs have adapted to their respective hosts over a long period, resulting in relatively high host specificity. For Blastocystis, we have shown that even though a given subtype may be found in more than one host species, it is possible to demonstrate cryptic host specificity at allele level. For instance, even though both human and non‐human primates can be colonised by ST3, host species‐specific strains of ST3 circulate within these two host populations. With regards to E. coli, it is possible that ST1 has adapted to human hosts, while E. coli ST2 has adapted to a broader host range, including non‐human primates and rodents. It has become clear that CLIPPs are common colonisers of the human background population, and even though we cannot disprove the existence of infection by any of these, it should be reasonable to consider clinical and medical intervention redundant in most cases. Perhaps it might even be so that one should try not to eradicate these organisms from the gut when first established. However, more studies are warranted to elucidate the significance of the pronounced genetic diversity observed in some CLIPPs with regards to transmission patterns and clinical significance. Future research in CLIPPs should also include studies that can elucidate those factors that favour colonisation with CLIPPs and what role CLIPPs have in host–gut ecology, metabolism and overall health condition. Finally, as human and non‐human hosts share these parasitic genera, and as some protozoa possibly contribute to overall gut function in ruminants, it would be interesting to study these in domesticated and wild animals to learn more about the role of these parasites in health and disease, including investigations into whether some CLIPPs might be endosymbionts.

Introduction

Common Luminal Intestinal Parasitic Protists (CLIPPs)—Overall Classification

Any non‐fungal eukaryotic organism living in/on and completing at least part of its life cycle in humans could be referred to as a ‘parasite’. Parasites comprise those affecting the skin (ectoparasites) and those living inside the body (endoparasites). Endoparasites can for practical reasons be divided into tissue, blood and digestive‐tract parasites. Parasites may be found in the human digestive tract from the oral cavity to the distal colon and anal canal. A simplistic but potentially useful categorisation of organisms parasitising the human intestine is provided in Figure 1. Overall, intestinal parasites of humans are either multicellular (metazoa; the ‘worms’) or single‐celled (parasitic protists, most of which are protozoa). Previously considered protozoa, the genera Enterocytozoon and Encephalitozoon are now taxonomically classified as fungi; in the clinical microbiology laboratory (CML), however, these are still mostly dealt with in the parasitology section of the CML. The life cycle of intestinal microsporidia largely resembles that of sporozoa, such as Cryptosporidium, with similar transmission modes and involving an intracellular life cycle stage and the shedding of spores in faeces that can be detected by for instance staining of faecal concentrates or DNA‐based methods. Also Blastocystis is an ‘outsider’, as it belongs to the group of Stramenopiles (heterokonts). An overall breakdown of the single‐celled intestinal parasitic genera commonly or less commonly (depending on the population examined) observed in human stool samples is provided in Table 1. Importantly, many of these genera can be encountered in non‐human hosts.

While some single‐celled intestinal parasites are mostly observed in patients with gastrointestinal symptoms, the reverse is true for others. For instance, with respect to individuals tested in Denmark, genera such as Cryptosporidium, Giardia and Enterocytozoon are only very rarely found in asymptomatic individuals, whereas these are increasingly being acknowledged as causes of (outbreaks of) diarrhoea and gastroenteritis in Scandinavia [1, 2, 3, 4, 5, 6, 7, 8]. Meanwhile, several other single‐celled intestinal parasite genera appear to be much more common in gut‐healthy individuals than in patients with intestinal symptoms and individuals with underlying diseases; these parasites are the ones that are the focus of this thesis.

In the scientific literature, the term ‘parasite infection’ may often be encountered. While it makes sense to define cryptosporidiosis as a disease caused by a Cryptosporidium infection because of the invasive stage of the Cryptosporidium life cycle, the pathology caused including the triggering of an immune response, it may be misleading to refer to, for example, a ‘Blastocystis infection’, in the event that no invasive stage, pathology or even immune reaction is involved during intestinal colonisation. Hence, it may be practical to differentiate between ‘infection’ and ‘colonisation’, where the latter does not typically activate the immune system or involve any pathology. Hence, it appears relevant to speak of ‘cryptosporidiosis’, whereas the term ‘blastocystosis’ would typically make little sense. This line of thinking also underlies the reason why I coined the term ‘common luminal intestinal parasitic protists’ (CLIPPs), where the word ‘luminal’ is meant to indicate the absence of gut‐invasive properties, and where ‘protists’ were chosen rather than ‘protozoa’, which term would exclude Blastocystis [9].

The life cycles of most CLIPPs involve both a trophozoite(‐like) stage and a cyst stage; however, for some (e.g., Entamoeba gingivalis and Dientamoeba fragilis), cyst stages remain to be identified and confirmed, respectively. Transmission occurs via the faecal‐oral route.

Although perceived as non‐pathogenic by most specialists, it has been customary in the CML to report these parasites when found in faecal samples, as the presence of these organisms indicate exposure to faecal contamination and prompt further investigations for pathogenic organisms.

Detection and Differentiation of CLIPPs

Except for Blastocystis, the organisms of study in this thesis were described in remarkable detail by Clifford Dobell (1886–1949), a British protozoologist. Most of Dobell's observations were published during the 1910s and 1920s and had vast implications for future parasitological evaluation of faecal samples. Since then, classical examination for single‐celled parasites in human stool samples has involved concentration of parasitic elements and removal of fat and debris from the samples to enable detection of (oo)cysts of protozoa and ova of helminths, and microscopy of fresh faeces and/or permanent staining of fixed faecal material has been used to detect trophozoites of protozoa and stages of Blastocystis. Identification of protozoa such as Entamoeba involves morphological description of any cysts observed, including information on size, number of nuclei, position of karyosome, presence/absence of peripheral chromatin in the nuclei and chromatid bodies among other features and/or description of trophozoite stages, which may also exhibit morphological hallmarks. Indeed, data on morphology combined with information on the sampled host have been used to name parasites using the classical binomial nomenclature. However, the use of DNA‐based methods has made it clear that detection and differentiation of single‐celled intestinal parasites based on light microscopy alone has at least two major The first limitation has to do with the limited sensitivity of microscopy‐based methods compared with DNA‐based methods [10, 11] which is an important point when the aim is to separate colonised from non‐colonised individuals. The second limitation pertains to the fact that morphologically similar organisms may exhibit extensive genetic diversity, a feature referred to as ‘cryptic genetic diversity’ (see below).

Over the past two decades, direct diagnosis of intestinal parasites has changed from relying predominantly on microscopy, including staining of faecal concentrates, to relying primarily on DNA‐based methods, at least in the modern CML [11, 12]. The incentive for the use of DNA‐based methods has been highlighted by Stensvold et al. [10] and Verweij and Stensvold [11] and includes increased sensitivity, automatability, standardisability and the opportunity easily to store and use positive DNAs for molecular typing where needed. DNA‐based methods can be tailored to specific needs and include singleplex, duplex and multiplex real‐time PCR assays, and, more lately, metabarcoding assays, just to mention some options that can be applied to genomic DNA extracted either directly from a sample or after any pre‐DNA extraction step that would aim to enhance the sensitivity and overall quality of the assay. PCR with Sanger sequencing has been used extensively for differentiation of parasitic species or subspecies/subtypes/genotypes, depending on the parasite in question.

To this end, it should be mentioned that the nuclear small subunit (SSU) ribosomal RNA (rRNA) gene has over time proved a quite robust taxonomic identifier for intestinal parasites. The nucleotide database at the National Center for Biotechnology Information (NCBI) (referred to in this thesis as either ‘NCBI nucleotide database’, ‘NCBI database’ or ‘GenBank’) and other databases are continuously being updated with sequence data from activities involving Sanger sequencing, but also next‐generation sequencing (NGS). Metabarcoding assays have been developed and already used extensively. One such example is the metabarcoding assay in place at Statens Serum Institut (SSI), which enables detection and differentiation of nuclear small subunit rRNA genes (also known as 16S and 18S genes) of bacteria, parasites and fungi, which means that this one‐fits‐most approach can be used to detect a variety of organisms simultaneously in anyone sample, although with varying sensitivity [13, 14, 15]. This ‘meta‐ribosomalomics’ approach is particularly useful to identify mixed colonisation by genetically related species or subtypes that cannot be differentiated otherwise (e.g., by microscopy or even other NGS‐based assays [16]) and to assist in mapping the ‘eukaryome’ of selected host species [17].

Given the likelihood of CLIPPs being confined to the gut lumen not triggering immune responses observed for parasites that cause pathology or those that at least are invasive (e.g., Toxoplasma), one would not expect serology to be a relevant diagnostic modality for CLIPPs detection. To my knowledge, Blastocystis is the only CLIPP where serological diagnosis has been attempted [18]. Hence, indirect detection methods appear to be of little relevance to CLIPPs.

Cryptic Genetic Diversity

Molecular characterisation of parasites is useful in epidemiological surveillance, including outbreak investigations, and for identification of lineages that differ in prevalence among healthy and diseased individuals, those that differ geospatially, and for revealing transmission patterns and host reservoirs. Within the field of intestinal parasite research, the relevance of molecular characterisation has been highlighted by the observation of cryptic genetic diversity within Entamoeba. Organisms that share identical morphological features but that differ genetically (and maybe also phenotypically) to an extent where it could be argued that they reflect different species and not only different genotypes may be referred to as ‘cryptic species’ [19]. In the early nineties, it was finally confirmed that some morphologically similar Entamoeba parasites were genetically different, leading to the separation of E. dispar from E. histolytica [20, 21]. Since then, E. dispar has been considered largely non‐pathogenic, while E. histolytica is known for its dramatic pathological potential [22, 23, 24]. This recognition has had important consequences for the diagnosis and management of patients with Entamoeba‐positive stool samples.

Entamoeba cysts with eight nuclei and with overlapping size ranges may be found in faeces from both primates and rodents; however, the cysts from rodents may genetically differ from those of primates by more than 17% across the entire SSU rRNA gene, and even within *Entamoeba coli—*which is the octonucleated cyst producer observed in humans—up to at least 12% divergence has been observed (see Section 2.3.1). Even higher differences (~30%) have been observed for morphologically similar cysts of Iodamoeba (see Section 2.5); to put this into perspective, some species—for instance E. dispar and E. histolytica differ by only 1%–2% across the entire SSU rRNA gene (Novati et al. mentioned 1.7% nucleotide (nt) substitutions [25]), and several species of Cryptosporidium differ by only 2%–4%.

Another example has to do with the species name ‘Blastocystis hominis’, a term now considered invalid [26] and which was applied to Blastocystis identified in stool samples. However, Blastocystis with indistinguishable morphological features can be found in humans and a vast variety of non‐human hosts, and studies of the genetic makeup of these morphologically similar organisms indicates the existence of multiple species (currently referred to as ‘subtypes’; see Section 2.1). Hence, molecular methods enable a higher and more accurate discriminatory ability compared with microscopy, and obviously, molecular methods are particularly useful for studying organisms for which cyst morphology data are not available (e.g., D. fragilis and E. gingivalis).

On the other hand, basing taxonomic inferences on molecular data only also has its limitations, which is why it has become common to use alternatives to the traditional Latin binomial nomenclature to delineate species at least for a preliminary period until data from sampling of relevant hosts are sufficient to make robust inferences on host specificity and until there are morphological data that can be paired with the DNA data. Terms commonly in use are ‘subtypes’ (e.g., Blastocystis and some species of Entamoeba) and ‘ribosomal lineages’ (e.g., Entamoeba, Iodamoeba and Endolimax).

There are no rules set in stone as to how genetically different organisms have to be in order to be considered two different species/lineages. For Blastocystis, we now recommend that a DNA sequence can be a candidate for a new subtype number if it covers > 80% of the entire SSU rRNA gene and differs by > 4% from previously sequenced complete Blastocystis SSU genes [27]. For practical reasons, colleagues provided definitions of species, subtypes, ribosomal lineages and conditional lineages for studies of genetic diversity of Entamoeba [28] (Table 2).

Importantly, terminology should be practical and pragmatic, and it is beyond the scope of the present research to venture into more theoretical discussions on species concepts and the more theoretical aspects on which taxonomical inferences are based.

Molecular Characterisation of Parasites

Molecular characterisation of parasite genera from human and non‐human hosts can help us identify whether transmission between host species, for instance between non‐human and human hosts, might take place. In other words, molecular characterisation can help us identify whether zoonotic transmission is likely for any parasite species in question by comparing parasite data from human and non‐human hosts. An intestinal parasite for which molecular characterisation has proven particularly useful in order to delineate patterns and the extent of zoonotic transmission is Cryptosporidium [4]. Also research into Blastocystis has had a particular focus on genetic characterisation both in order to identify host specificity and to enable discovery of potentially pathogenic variants (see Section 2.1). Various genes are used as markers; for CLIPPs, the SSU rRNA gene generally holds a lot of information, while for others, it is quite conserved across the genus, which is why other genes, typically house‐keeping genes (e.g., actin, elongation factor 1‐alpha, heat‐shock protein 70), or genus‐specific genes (e.g., glycoprotein 60 or beta‐giardin for Cryptosporidium and Giardia, respectively) are chosen for better discrimination.

For CLIPP genera such as Entamoeba and Blastocystis molecular characterisation is relatively straightforward and has largely been based on specific PCR and subsequent Sanger sequencing. However, obtaining Endolimax and Iodamoeba SSU rDNA sequences can be challenging because of the typical absence of cultured material and the fact that SSU rDNA sequences of these parasites are relatively long (~2.5 kbp) [29, 30]. PCR using low‐specificity eukaryotic primers preferentially amplifies any shorter and more abundant SSU rDNA from co‐infecting/co‐colonising organisms present in the intestine. This is often Blastocystis sp., which is frequently observed in Endolimax‐ and Iodamoeba‐positive samples, as SSU rDNA sequences of Blastocystis are around 700 bp shorter than those of Endolimax and Iodamoeba. Even when specific amplification of Iodamoeba‐ or Endolimax‐specific DNA is successful, Sanger sequencing of the PCR product will often result in a sequence of low quality (i.e., with double peaks and sequence patterns lacking synchronisation) because of high intra‐genome variation among the ribosomal gene copies, including differences in homopolymere length [31]. This makes Sanger sequencing of PCR products problematic and unable to clarify genetic diversity when used alone. Therefore, when relying on Sanger sequencing, a cloning step prior to sequencing has proven necessary for Iodamoeba and Endolimax SSU rDNA sequencing [29, 32]. This is probably the reason why accumulation of SSU rDNA sequences of these genera in the NCBI database has been relatively slow. Indeed, at the time of writing (January 2023), the number of Iodamoeba‐ and *Endolimax‐*specific DNA sequences in GenBank are 79 and 34, respectively; these are quite modest numbers when compared to the number of sequences available for Entamoeba (~300,000) and Blastocystis (~75,000), and, for Iodamoeba, more than half of the sequences stem from one single study only and are practically identical.

As opposed to other areas within the field of microbiology, whole‐genome sequencing (WGS) is not yet a viable option for molecular characterisation of parasites. Very few parasites are readily established in culture and they are generally difficult to isolate in other ways. Moreover, whereas viruses and bacteria have relatively short genomes (e.g., ~30 kb for severe acute respiratory syndrome coronavirus 2), parasites have genomes of typically about ~20 megabases (mb), which makes WGS expensive and time‐consuming. And as CLIPPs and many other parasites are difficult to isolate from other organisms, the presence of competing template is typically an issue in amplification and sequencing procedures. Moreover, and not surprisingly given the situation described above, the reference data pipeline for WGS is still rudimentary, so even if WGS data would be available, genome data analyses would be challenging and time consuming. Nevertheless, given the diversity within CLIPPs already recognised, it would be fair to expect vast whole‐genomic diversity, which could again be reflected in very diverse influences on the host and, maybe especially, host–gut microbiota across CLIPPs; therefore, the mapping of genomes should receive priority.

Nevertheless, lately, the advantages of screening for genetic variation among CLIPPs using metabarcoding has become evident. Here, shorter fragments (a few hundred bp) of PCR‐amplified SSU rRNA genes are sequenced using NGS technology, and consensus sequences can be generated manually from the total sequence output using DNA sequence clustering tools and become subject to genetic analysis, including phylogenetic analysis. The method may even be sufficient for establishing hypotheses on new species/ribosomal lineages and can be used to demonstrate intra‐genome variation in the genera Iodamoeba and Endolimax. However, near‐complete SSU rDNA sequences should be obtained for potentially new lineage whenever possible for more robust phylogenetic inferences, and here, other technologies, such as MinION sequencing, are currently being tested.

During the period in which the publications included in the present thesis were published, research into genetic and host specificity of parasites went from relying mainly on traditional parasitological methods (including morphological observations) and conventional PCR and sequencing to relying increasingly on NGS‐based approaches such as metabarcoding (or ‘meta‐ribosomalomics’). This is reflected in the selection of publications included in this thesis, of which six involved the use of data obtained by metabarcoding. In the absence of morphological data, the applicability of the concept of cryptic species is challenged and delineation of species would have to rely solely on phylogenetic species concepts. Nevertheless, metabarcoding has been of particular use in the mapping of parasites in wastewater samples, where it contributes to capturing of the diversity and relative prevalence of parasites in the society from which the wastewater is sourced [33].

Gut Microbiota Profiling

Despite trailblazing observations on microeukaryotic diversity in the human distal gut already in 2008 [34], until quite recently, high‐throughput sequencing technologies used in studies of gut microbiomes have largely focussed merely on bacterial taxa. It may be so that the bacteria in the gut outnumber other members of the gut microbiome, namely archaea, fungi and non‐fungal eukaryotes including CLIPPs; however, the latter may with their much larger genomes play roles that would be ignored, if studies relied on analysis of bacterial taxa only [35]. An advantage of the use of meta‐ribosomalomics is the ability to compare bacterial communities in the digestive tract in individuals with and without parasites to test for significant differences in bacterial composition and to hypothesise to which extent such differences might be driven by parasite colonisation/infection.

Studying the distribution of parasitic genera common in humans in non‐human hosts and the relationship between such parasites and bacterial communities in non‐human hosts can help us hypothesise on the clinical and public health significance of these parasites in humans. Studies of wildlife mammals sampled in their natural habitats appear particularly interesting, as the gut microbiomes of these animals may largely be unaffected by factors reflecting human intervention and therefore would appear more original, potentially reminiscent of the original human gut microbiome given the evolutionary relationships between humans and larger mammals. Gut microbiota profiling of experimental animals before and after exposure to parasite inoculation can also help (1) identifying factors that might drive the establishment of parasites in the intestine and (2) explore the impact of parasite establishment on gut ecology and host immune responses [36, 37].

Aim of Studies

The primary aim of the work reflected in this thesis have been to expand the knowledge on the genetic diversity and host specificity of CLIPPs found in the human digestive tract and mostly referred to as non‐pathogenic, namely Blastocystis, Dientamoeba, Endolimax, Entamoeba and Iodamoeba. A secondary, and perhaps a more presumptuous aim was to try to inform public health policy makers and stake holders regarding the significance of single‐celled parasitic genera commonly found in humans by obtaining a robust and reliable impression of the colonisation rate of some of these genera in both human and non‐human hosts and by studying digestive tract microbiota profiles in association with the presence/absence of intestinal parasite colonisation.

Genetic Diversity and Host Specificity of CLIPPs Observed in Humans

In this section, there will be a brief account of the genetic diversity and host specificity of the most common CLIPPs found in humans, namely Blastocystis, Dientamoeba, Endolimax, Entamoeba and Iodamoeba.

Before that, I would like to highlight briefly a few methodological features that pertained to much of the work related to obtaining and analysing genomic data in many of the Genomic DNA was extracted from faecal/environmental samples or cyst preparations using typically the QIAamp DNA Stool Mini Kit (QIAGEN, Hilden, Germany) or the EasyMag or eMAG (NUCLISENS; bioMérieux Inc.) protocols.Conventional PCRs used primers as appropriate, and Sanger sequencing was used to sequence PCR products. Two of the studies involved TA cloning [29, 32].Sequence alignments for the generation of consensus sequences and alignments for phylogenetic analyses were generated using a variety of free software; first and foremost different versions of MEGA [38, 39, 40], but also the online software Multalin (http://multalin.toulouse.inra.fr/multalin/) and Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). In cases where consensus sequences were generated from sequence outputs from amplicon‐based NGS, it could not be ruled out that a very minor fraction of the SNPs identified would reflect sequencing error rather than biological polymorphism. However, none of our analyses would be particularly sensitive to these errors, as these errors would either be diluted out by the vast variation seen between some subtypes (e.g., E. gingivalis; Section 2.3.3), or they would be excluded from our multiple sequence alignments, as we would generally exclude bp positions where a SNP would be present in one sequence at a position where all other sequences were similar and if the sequence would otherwise be similar to another sequence in the alignment.For mitochondrion‐like organelle (MLO) genome assembly [41], Staden Package [42] was used.Metabarcoding was used in some studies to identify positive samples and to screen for genetic diversity. Fasta files extracted from the BION server available at SSI were used to generate consensus sequences as appropriate [13, 33, 43, 44, 45].What is referred to as ‘barcoding’ [46, 47, 48] was used to identify subtypes of Blastocystis [47, 48].Phylogenetic analyses used distance‐based analysis (neighbor‐joining method), maximum likelihood and Bayesian analysis as relevant [29, 32, 44, 49, 50, 51, 52, 53].Two articles included diversity analysis of gut or oral cavity bacterial microbiota in relationship with parasite colonisation, where the analyses included were alpha and beta diversity analyses and linear discriminant analysis effect size (LEfSe) analysis [43, 45].

Small subunit rDNA sequences were sourced from the NCBI nucleotide database (https://www.ncbi.nlm.nih.gov/nucleotide/). Briefly, the species or genus name (as applicable and relevant) was entered in the search filed (in citation marks) along with Boolean operators as applicable (e.g., “Entamoeba coli” AND “ribsosomal”), and the resulting list of hits was then subject to filtering using the filters and other features available at the site where relevant to ensure the inclusion of 18S sequences only and exported to Excel using the feature ‘Send to’, by choosing ‘Complete record’ + ‘File’ and the ‘INSDSeqXML’ option under ‘Format’. The downloaded file would then be opened in Excel and sort functions used to identify relevant features, such as host information, sample material, date of data deposition, country of origin, where available. Excel pivot functions were used particularly for Blastocystis.

Near‐exhaustive summaries of SSU rDNA gene sequences are provided for all the organisms, but in the thesis proper, only summaries for non‐Blastocystis organisms were included, as the number of Blastocystis SSU rDNA sequences in the NCBI database is around 10,000 and therefore would take up too much space in this thesis, in my opinion. However, a full repository of Blastocytis SSU rDNA sequences is available at GitHub (https://github.com/Entamoeba/DMSc‐Thesis/blob/main/Blastocystis%20SSUs%20from%20GenBank%20‐%20Stensvold%2021012023.zip). The data downloaded from the NCBI database for the tables were edited only to a minimum (e.g., spelling errors and information on host, where this had not been provided but was given elsewhere).

At least three issues generally hamper attempts to generate exhaustive species‐specific SSU rDNA sequence repositories for CLIPPs—these are described and exemplified If only sequences listed specifically as for instance ‘Entamoeba coli’ or ‘Entamoeba polecki’ are selected, then information from some of the many sequences deposited as ‘Entamoeba sp.’ or ‘uncultured Entamoeba clone’ that are in fact ‘Entamoeba coli’ or ‘Entamoeba polecki’ is missed.Examples are as KY658179 is listed as ‘Entamoeba sp.’ but is ‘Entamoeba coli ST3’JX131936 and JX131943 are listed as ‘Uncultured Entamoeba clone’, but are in fact ‘Entamoeba hartmanni’ Sequences are seen listed as, for example, ‘Entamoeba polecki’ but cannot be verified as such (and maybe not even Entamoeba). Examples: AB845670, AB845671.An Entamoeba coli sequence was observed listed as ‘Entamoeba muris’ in GenBank (FN396613), and would therefore be missed, if only ‘Entamoeba coli’ sequences were queried. There are also examples of situations where an Entamoeba has been misclassified as a completely different genus (e.g., KU886548, which is listed as Terfezia, but which is Entamoeba gingivalis).

Another example of GenBank sequence entries where things have gone really wrong can be observed for the sequences MW133761–MW133772, which are all individually listed as ‘Entamoeba sp.’ For this batch of sequences, only two are indeed Entamoeba, namely MW133765 and MW133768, which are both Entamoeba coli.

When assessing the validity of sequences potentially reflecting new intra‐specific lineages, it may be helpful if already two or more lineages have been acknowledged within a species. If the variation observed in a new sequence is limited to parts of the gene where variation exists between the lineages already acknowledged, this could indicate that the sequence is valid (and not artefact). Meanwhile, if random SNPs are observed in places where the gene is otherwise conserved (typically between species of a genus), the sequence might be considered an artefact until confirmed. Sequence chimaeras (see Section 2.3.6), of which there are quite a few in GenBank, can distort phylogenetic analyses and should be recognised and removed from analyses.

Finally, it should maybe be mentioned that according to PubMed, the number of publications related to Blastocystis is 10 times the number of that of for instance Iodamoeba‐associated publications, and therefore, the section on Blastocystis is more elaborate than sections on most of the other CLIPPs included.

Blastocystis

Aspects of Biology, Life Cycle, Diagnosis and Epidemiology

The most common intestinal parasite in the human population might be a parasite that genetically does not resemble any of the other organisms parasitising on the human gut, namely Blastocystis. An oomycete sharing a most common ancestor with Proteromonas [54, 55], which is a flagellated single‐celled parasite of the hind gut of lizards, it might colonise at least one billion people worldwide, and easily two, if data from recent DNA‐based surveys of Blastocystis in European study populations can be extrapolated to the rest of the world. Indeed, in some populations, Blastocystis appears to be more or less an obligate finding in stool samples [56, 57]. Blastocystis is one of at least two Stramenopiles organisms known to parasitise on humans, the other one being Pythium, a cause of keratitis and endophthalmitis among other clinical manifestations [58].

Nevertheless, the life cycle and transmission of Blastocystis is quite reminiscent of that of many other CLIPPs, comprising a trophozoite‐like stage (typically referred to as the ‘vacuolar form’) and a cyst stage [59] that makes Blastocystis amenable to faecal‐oral transmission. There is some evidence that cysts might not always be detectable in Blastocystis carriers and that the detection of cysts in human faecal samples is independent of subtype and estimated gastrointestinal transit time [60]. Other stages reported include the granular and the amoeboid stages; however, it is not clear to which extent these stages reflect artefact stages. At least the granular stage may often be encountered in cultures and may reflect cells about to undergo apoptosis, as granular cells are often seen in cultures that have not been maintained, whereas rarely seen in fresh cultures (personal observations).

Indeed, Blastocystis is one of the few CLIPPs that can relatively easily be cultured [60, 61, 62], which expands the research opportunity for this parasite; however, Blastocystis is difficult to grow in the absence of bacteria in the culture medium. Cultures can be cryopreserved and thawed for future use [61]. With the specificity of culture being optimal [63], short‐term in vitro culture as a diagnostic method appears superior to the traditional formol ethyl acetate concentration technique (FECT), but inferior to DNA‐based detection [62]. Culture also appears to have high efficacy for Blastocystis in non‐human hosts such as NHP and pigs [64, 65], although a more detailed view of the applicability to non‐human samples of short term in vitro culture as a diagnostic method is warranted. A positive culture result would evince colonisation and the presence of a live isolate in the host from which the sample is sourced, whereas a positive PCR result only shows that DNA of Blastocystis (whether dead or live) has been detected in the sample.

After developing several DNA‐based methods for detection and subtyping [16, 62, 66], we developed a real‐time PCR for more sensitive detection of Blastocystis [67], and this method was recently recommended for use in diagnostic laboratories [68]. Real‐time PCR is possibly one of the most sensitive methods to distinguish between carriers and non‐carriers, and with faecal DNA already available, subtyping using state‐of‐the art methods [41, 47, 48] can be performed for research purposes. Recently, metabarcoding was proved to be a quite efficient way of detecting Blastocystis [13, 14], with the added benefit of enabling subtype determination based on the data output, including differentiation of multiple subtypes in any given sample [33].

Blastocystis is a rare example of an anaerobic eukaryote with organelles that have retained some mitochondrial characteristics, including a mitochondrial genome, and so, unlike most of the other CLIPPs included in this thesis, Blastocystis has not only one but two a nuclear genome of 12.9–18.8 mb (depending on ST) encoding 5713–6544 proteins, and a ‘mitochondrial’ genome of 27.7–29.3 kb [59]. The mitochondrial genome is that of the so‐called MLO, and the first MLO genomes were published by Perez‐Brocal and [69] and Wawrzyniak et al. [70] in 2008; these genomes represented Subtypes 1 and 4 (for more information on subtypes, see Section 2.1.2). A few years later, our group sequenced a few additional MLO genomes [41, 71] for two particular (1) to learn more about the genes contained in the MLO and (2) to develop multi‐locus sequence typing systems that might yield more genetic resolution than mere SSU rDNA sequence analysis.

Denoeud et al. published the first nuclear genome of Blastocystis in 2011 [72], which may very well be the smallest genome ever sequenced of a Stramenopiles organism, being just under 19 mb in size and containing about 6000 genes.

Maybe one of the most interesting take‐home message from studies of both MLOs and the nuclear genome is that Blastocystis might be able to tolerate environmental exposure that is not strictly anaerobic [73]. Still, Blastocystis appears to thrive in gut ecological niches that are dominated by anaerobes [45]. In humans, it is likely that Blastocystis may primarily be lodged in the terminal ileum, the caecum, and the proximal part of the colon.

The fact that Blastocystis is one of very few parasites found in humans that can easily be cultured [61, 62], makes it more amenable to genome‐based studies than many other parasites. This is already reflected by the fact that there are several draft nuclear genomes in GenBank, some of which were deposited by our group, and there are also genomes representing the MLO for quite a few of the subtypes. Another potential opportunity of culture could be the induction of the cyst stage so that cysts could be produced for experimental studies; however, to my knowledge, a robust protocol for cyst induction is still to be developed.

Data on Blastocystis in the environment are scare. Maybe unsurprisingly, this parasite was an obligate finding in our study of Swedish wastewater samples [33]. In Thailand, a variety of subtypes were identified from Blastocystis cultured from water and soil samples [74, 75]. The parasite can survive exposure to chlorine and hydrogen peroxide [76].

In some human populations, Blastocystis is almost an obligate finding. All 93 children sampled in Senegal by El Safadi et al. tested positive [56], and in a study that we carried out on samples from 199 Nigerian children aged 2–14 years, we identified a positive rate of 84%, with prevalence increasing by age [57]. In children < 4 years of age, Blastocystis was not nearly as common as in older children, and, in general, there was an almost linear relationship between colonisation rate and age. As exposure to Blastocystis could be similar across all age groups, the association between age and Blastocystis colonisation independently corroborated by relatively high‐powered studies from Brazil [77] and Libya [78] should be investigated in greater detail. It could be hypothesised that Blastocystis is dependent on a more mature (diverse) microbiota in order to be able to establish.

In our study of samples from Nigerian children, we found a clear inverse association between Blastocystis colonisation and malaria infection (P < 0.0001) [57]; however, children with malaria being younger than children with no malaria, this finding was attributed to the age effect of Blastocystis colonisation.

In Denmark, Blastocystis and Dientamoeba are by far the most common CLIPPs observed, but where D. fragilis is mostly seen in younger children, Blastocystis tends to colonise a wider range of age groups, being more common in adolescents and adults [79]. In adults, it is not uncommon to see both.

Blastocystis appear to be able to colonise the human intestinal tract for many years. Indeed, in our study led by Dr Scanlan, we observed that Blastocystis was present in a subset of healthy adult individuals sampled over a period of time between 6 and 10 years, indicating that it is capable of long‐term host colonisation [80]. We based this assumption on the observations that the test individuals had tested positive for the same SSU rDNA sequence allele (see below) on samples taken several years apart.

The clinical significance of Blastocystis has been subject to a very heated debate over the past few decades. It has been speculated that Blastocystis could induce disease through elicitation of toxic‐allergic reactions, degradation of human secretory immunoglobulin A by proteases, changes in epithelial permeability, induction of apoptosis of host intestinal cells and disruption of the epithelial barrier function and/or modulation of immune response and cytokine release from colonic epithelial cells [81]. Some years back, we focussed on some of the limitations associated with pursuing the clinical significance of Blastocystis [81]. Case control, cohort and randomised controlled clinical trials that could shed light on the role of Blastocystis in health and disease remain limited and are usually hampered by the use of methods of limited efficacy, including insensitive methods for evaluation of medical intervention. Moreover, Blastocystis is often observed together with other parasites, such as D. fragilis and species of Entamoeba, so investigating the isolated clinical effect of Blastocystis carriage is far from straightforward.

Nevertheless, we reviewed the efficiency of drugs used to eliminate Blastocystis, finding that the efficacy of metronidazole (MZ), which has been recommended [82] and used with an aim to ‘cure’ what has been called ‘blastocystosis’, was extremely limited [83, 84]. In vitro, MZ has high efficiency; however, we produced data indicating that herbal extracts of Mallotus oppositifolius, a medicinal plant traditionally used in sub‐Saharan Africa to treat or alleviate stomach ache, diarrhoea/dysentery and diabetes, are almost just as efficient [85].

I was involved in a longitudinal, prospective case study, where 11 symptomatic patients positive for Blastocystis underwent outpatient clinical assessment to exclude other diagnoses before being treated for 14 days with either MZ 400 mg × 3/daily or trimethoprim (TMP)/sulfamethoxazole (SXT) 160/800 mg × 2 daily; none of the patients tested negative for Blastocystis following therapy [86].

I have moreover been involved in three cases for which data on the effect of treatment were One was the case of a woman with combined Blastocystis subtype 9 (ST9) and D. fragilis colonisation, who failed to experience clinical or microbiological effect of various treatments (Table 3) [84]. In another case, we were following a woman with ST8 as the only identified potential cause of intestinal symptoms, who achieved clinical and microbiological cure after completing 10 days of treatment with thrice‐daily doses of 80 mg TMP/400 mg SXT; previous treatments with MZ had proven futile [87]. In the third case, a 20‐year‐old man with ST2 colonisation, who developed generalised urticarial, also failed to respond to MZ treatment and experienced clinical and microbiological cure only after adding paromomycin to the MZ treatment [88]. To my knowledge, no drug has to date proved 100% efficient in terms of eliminating Blastocystis.

Still, it could easily be argued that Blastocystis should not generally be included in routine diagnostic work‐ups in clinical microbiology laboratories but rather be subject to scrutiny in research studies aiming to map the public health and gastroenterological impact of presence and absence of the parasite [89]. This conclusion is based on two major Firstly, in Denmark, Blastocystis appears to be more common in the background population (positivity rate, 22%), while less common in patients with functional bowel disease (positivity rate, 15%) [90] and least common in patients with inflammatory bowel disease and acute diarrhoea (positivity rates, 0%–5%) [91, 92] (see Section 4), which could suggest a protective role of Blastocystis. Secondly, since long, Blastocystis has by our own research groups (and since then also by many others) been hypothesised to be an indicator of a healthy gut microbiome and a normal body mass index (see Section 4) [45, 93, 94, 95, 96]. Thirdly, eradicating Blastocystis from the gut can be extremely difficult, and the treatments might do more harm than good, also in the long run. Finally, and very importantly, the current focus on antimicrobial stewardship should be prioritised. Overuse of for instance MZ can lead to antimicrobial resistance in both bacteria and parasites.

Genetic Diversity and Host Specificity of Blastocystis

In 2006, we revisited the terminology of Blastocystis, which by then had been subject to a perplexing number of classification systems. In the process, we decided to abandon the species name ‘Blastocystis hominis’ for reasons stated above (see Section 1.3), and consensus was also reached to adapt the subtype system still in place, and which is based primarily on the relative amount of diversity across Blastocystis SSU rRNA genes [26]. Since then, subtyping studies of Blastocystis from humans, other mammals and birds have led to a vast increase in publications on Blastocystis. This means that Blastocystis sequences from hosts such as amphibians, reptiles and insects are not included in the subtype terminology.

Before this time, molecular characterisation of Blastocystis had often been performed using the so‐called sequence‐tagged site (STS) primers originally developed by Yoshikawa in 1998 [97]. These primers had been designed from random amplified polymorphic DNA sequences, with the nature of the DNA targets as well as their copy numbers remaining unknown. When I compared this method with barcoding published in 2006 by Scicluna et al. [46], I found that the latter method was more sensitive [48]. False‐negative results by the STS assay were not linked exclusively to certain subtypes or alleles, and evidence of substantial genetic variation in STS loci was obtained. Over the next many years, the use of the STS method diminished, while the barcode method has been used by quite a few research teams to date. This has been advantageous for Blastocystis research, as standardisation of methods allows for inter‐study comparisons, and DNA sequences obtained by the barcode method hold more information than just the subtype [41].

When the terminology was introduced, a total of nine subtypes were acknowledged [26]. Two years later, we published evidence of a new subtype, ST10, which we had identified in several Danish cows, but also in a sheep, a roe deer and in a lemur from a zoo [51]. Again a few years later, the number of subtypes increased to 17, as a few new subtypes were Our Australian colleagues identified ST11 from elephants, ST12 from giraffes, and ST13 from a quokka [98]; moreover, Fayer et al. introduced ST14 based on sequences obtained from cattle in the US [99]. In our large survey from 2013 going out from London School of Hygiene and Tropical Medicine and led by Dr Alfellani, we corroborated the validity of ST13, which had now been found in a mouse deer, and ST14, found in cattle and a mouflon [100]. Moreover, Subtypes 15, 16 and 17 were introduced as new subtypes found in camel and gibbon (ST15), kangaroo (ST16; data deposited in GenBank by Hisao Yoshikawa) and a gundi (ST17).

Over the next 10 years, many more subtypes made it into the Blastocystis terminology, some of which were found invalid, and where we realised that the sequences that had been proposed as new subtypes represented sequence chimaeras [27]; these subtypes (ST18, ST19, ST20 and ST22) are now considered redundant. At the time of writing (January 2023), at least 34 subtypes have been acknowledged, and as the ‘tree’ has expanded, it is becoming clear that the subtype collection might benefit from a re‐evaluation to keep terminology practical (Figure 2). For instance, it might be useful to collapse ST24, ST25 with ST14 for two important (1) They cluster with very high bootstrap value and appear to exhibit less genetic diversity than what is observed within a subtype such as ST7; in fact, the two subtypes are not receiving individual support in Figure 2, which is based on 1189 positions (and not near‐complete genes, as these are not available for all subtypes), but are rather engulfed by ST14; (2) there is an overlap in host range. A somewhat similar situation is seen for ST21, ST26, ST30 and ST32, all of which are from ruminants, and which could be regarded as one subtype. However, any changes to the terminology might not be imminent, as it is perfectly operable for the time being, and it has not yet proved unuseful as such.

There are examples of sequences in GenBank that might represent new subtypes but for which additional sequencing and revisions of phylogenetic analyses are required. The sequences published in the study on Blastocystis in rabbits in China from 2022 by Su et al. is just one such example [101].

Efforts to identify whether ‘new’ sequences could represent a new subtype have been facilitated by sequencing technologies that were not in place until relatively recently. Especially MinION sequencing has proved particularly relevant, as it allows for cost‐effective sequencing of near‐complete SSU rRNA genes [102, 103], which was previously a laborious and difficult activity, but which was recommended to increase the quality of phylogenetic analyses [27, 73]. My own experience with MinION sequencing technology is limited, but we used it in a study of Blastocystis and Entamoeba in muskoxen (article in preparation).

There are now close to 12,000 nuclear SSU rDNA sequences (18S) for Blastocystis in the NCBI database (table S1 available on GitHub: https://github.com/Entamoeba/DMSc‐Thesis/blob/main/Blastocystis%20SSUs%20from%20GenBank%20‐%20Stensvold%2021012023.zip), and almost 10,000 for which host information was provided, so it is beyond the scope of this thesis to provide an exhaustive review of the host specificity of Blastocystis. The number of Blastocystis‐associated publications has sky‐rocketed over the past few years, with almost 750 being included in PubMed over the past 5 years. I will, however, try and draw up some features of Blastocystis host specificity in the following.

Our understanding of Blastocystis host specificity has been building up for now more than 15 years. Maybe more than 95% of colonised humans in Denmark and the rest of Europe have one or more of the Subtypes 1, 2, 3 and 4 [80, 92, 104, 105, 106, 107, 108], and globally, ST3 appears to be the far most predominant subtype of the four in humans [109]. Interestingly, outside Europe, mainly Subtypes 1, 2 and 3 predominate, while ST4 is rare [14, 104, 110, 111, 112, 113]. Other subtypes may rarely be seen in human hosts, and it is therefore likely that humans are not natural hosts of such other subtypes.

Looking at non‐human primates sampled in the old world, including both apes and monkeys, these tend to harbour the same subtypes as seen in humans, except for ST4. Meanwhile, we have published data suggesting that ST5 might be a common finding in apes, although potentially not in monkeys [114]; however, we found a few different species of cercopithecines in Thailand colonised by ST5 [65]. The predominance of ST1–ST3 in non‐human primates [114], which happen to be the STs most commonly associated with humans, might suggest that these STs have a shared co‐evolutionary history with humans and their closest living relatives. ST4, by contrast, is uncommon in non‐human primates, whereas it is common in humans in Europe; contrary to ST1–ST3, ST4 from humans appears genetically conserved (see below), indicating a recent entry in the human population [115].

There is a fair chance that ST3 is the most common subtype in both human and non‐human primates, and one might easily jump to the conclusion that NHPs constitute a reservoir for human ST3 colonisation; however, based on MLO genomic analysis, we produced evidence of cryptic genetic diversity within ST3, suggesting host‐adapted genotypes (i.e., variants within a subtype) of ST3, something that we confirmed by allele analysis of ST3 found in cercopithecines sampled in Thailand [65]. The cryptic diversity disclosed by studying MLO genome data was to a high degree reflected in ST3 SSU rDNA sequences, which is why the ‘allele’ concept and the publically available MLST database (https://pubmlst.org/organisms/blastocystis‐spp) was introduced [41]. The allele system takes advantage of any intra‐ST variation detected in Blastocystis. Sequences that differ down to one SNP can be submitted to the MLST database and is assigned a unique allele number. These allele numbers can then be used in studies calling for more subtle discrimination of strains than can be provided merely by providing information on subtype.

If we look specifically at arboreal monkeys, we found ST8 in monkeys of Asian or South American origin [114], but apart from these, ST8 has only relatively rarely been reported in primates.

From the research undertaken so far, the extent to which livestock and other synanthropic animals contribute to transmission of Blastocystis to humans remains somewhat unclear. A preponderance of ST10 and ST14 is seen in ruminant hosts sampled across the world, neither of which are often observed in humans. Meanwhile, the subtypes typically seen in suid hosts are Subtypes 1, 3, 5 and 15 [13, 116], and even ST2 has been found in pigs [116]. Of these, both ST1, ST2 and ST3 are commonly seen in humans, but we currently do not know if ST1, ST2 and ST3 strains from pigs are different to ST1, ST2 and ST3 strains colonising humans, as the allele system may not hold sufficient discrimination to tell these apart [13]. Analysis of complete SSU rDNAs and/or MLO genome markers might enable researchers to provide answers to this question.

Avian subtypes, such as ST6 and ST7, may occasionally be seen in human faeces [117], and ST8 is also occasionally seen. We also found ST8 in quite a few samples in our study of Swedish wastewater [33]. As ST8 is uncommon in humans, and as many other parasites found in the wastewater material were parasites known to colonise humans, we speculated that ST8 might stem from an animal that can live in sewers, for instance rats, and that rats might even be natural hosts of ST8. Galán‐Puchades et al. identified a Blastocystis colonisation rate of 83.5% in rats sampled in parks and sewers of Barcelona [118]; however, no subtyping of these samples was performed in the study. ST4, which is the sister taxon of ST8 (actually, ST8 was split out from ST4 when we established the consensus terminology), has been shown to colonise rats and other rodents easily [119, 120, 121, 122] (see below), so the hypothesis might not be farfetched. More extensive sampling of rats is necessary to complete this picture. Of note, ST1 of human origin can be established experimentally in rats, with colonisation lasting for more than a year [37].

Until very recently, ST9 appeared to be an extremely rare finding and potentially restricted to humans hosts [84]; however, in 2021, Liu et al. published extensive evidence of ST9 in peafowl in China, and poultry was also identified as hosts of ST9 in Malaysia (GenBank entry KX234596). Indeed, it would not be surprising if birds were confirmed natural hosts of ST9, as this subtype is a sister taxon to ST6 (ST9 was split out from ST6 in when we established the consensus terminology), and as also ST27, which is also observed in birds, clusters together with ST6, ST7 and ST9 (Figure 2).

Many subtypes have more or less exclusively been detected in ruminants or at least in large, mainly herbivorous or omnivorous mammals, including marsupials. These include the subtypes found in the upper third of the tree (Figure 2), which is the part of the tree that I would refer to as the ST5‐ST14 clade, and which includes the Subtypes 5, 12, 13, 14, 21, 24, 25, 26, 30, 31 and 32. An interesting feature of this clade is that white‐tailed deer have been found to host quite a few of the subtypes in this clade, including ST14, ST21, ST30 and ST31. Although much more limited in the number of subtypes, there is another artiodactyle‐specific clade made up by ST10 and ST23, which again includes white‐tailed deer as host among cattle and dromedaries.

The base of the phylogenetic tree has been expanded a bit recently by a number of new subtypes that tend to sit on relatively long branches, which indicates a substantial degree of genetic diversity. Some of these have been found in hosts that might not previously have received much attention in Blastocystis research, such as bats and heteromyids.

Overall, the most clear host specificity is seen mainly for the ST5‐ST14 clade, comprising ‘artiodactyle subtypes’, which could be considered an analogue of the ‘bovis complex’ of Entamoeba (see Section 2.3) and for the clade made up by the avian subtypes (mainly ST6 and ST7). Apart from this, the lack of clear co‐evolutionary trends could indicate that Blastocystis has been entering most major host groups more than once.

When having a rough look at the breakdown of hosts for the > 9800 SSU rDNA sequences in GenBank for which host information is available, humans account for a good half of these (56%). Cattle contribute at least 10% of all sequences, and suids 7%. Non‐human primates account for at least 7.5% of the sequences, avian hosts 6%, and goats and sheep together 3% of the sequences. Larger mammals like elephants, giraffes, lamas, pandas, horses and deer contribute about 5%–6%, and rodents 1%. Other sequences are from lagomorphs, marsupials, reptiles and even fish (Clupea harengus and Pollachius virens) and cockroaches. Strictly carnivorous mammals and omnivorous scavenger animals, on the other hand are relatively scarcely represented, with about 34 sequences from dogs (0.3%), 22 from cats (0.2%), 19 from foxes (0.2%), 11 (0.1%) sequences from bears and 8 (< 0.1%) from raccoon dogs. To this end, carnivores do not appear to host any particular subtype but rather a large variety, including Subtypes 1–4, ST7, ST10 and ST14, to mention some, which together with the small number of sequences and the possibly very low prevalence of Blastocystis could indicate that carnivores may not be natural hosts of Blastocystis (also see Section 5). A study adding support to this hypothesis is the one by Heitlinger et al., who characterised the eukaryotic and bacterial faecal microbiota of 42 spotted hyenas, and who did not report any finding of Blastocystis [123].

We (and many other teams) have highlighted the apparent absence or rarity of ST4 in human populations outside Europe [14, 57, 111, 124, 125, 126, 127, 128, 129]. Although rare in humans outside Europe, this subtype has meanwhile been an occasional finding in non‐human hosts. In a metanalysis of data published from the Americas, Jiménez et al. identified the observation of ST4 in 1.7% of humans sampled and in 7.2% of samples from non‐human hosts [130]. Extending the metanalysis approach to global scale, Barati et al. [120] identified 18 studies (16 of which were studies of rodents sampled outside of Europe) that had used molecular methods to detect and differentiate Blastocystis subtypes in rodent samples, and ST4 was the most common among these.

Genetic variation exists within ST4, and already in 2011, we could differentiate at least two major clades; one that appeared common in both human and rodent hosts and one that we thought might be limited exclusively to rodent hosts [92]. However, the year after, we had produced evidence of a ST4 sequence from a human clustering in the clade thought to be strictly rodent (JN682513) (Figure 3). There are still two major clades of ST4, of which one (Clade 2) appears to be ‘predominantly rodent’, while the other one (Clade 1) contains sequences from both rodents, humans and other mammals, a finding more recently corroborated by Katsumata et al. [119]. An interesting feature of these two clades is that where Clade 1 is practically clonal (i.e., with very limited genetic variation), Clade 2 exhibits genetic heterogeneity. It could be speculated that the introduction and expansion of Clade 1 in humans happened once, in Europe and relatively recently. Analysis of MLO genomes of ST4‐positive faecal DNAs representing both clades confirmed the genetic homogeneity of Clade 1 [41]. Very interesting is the fact that Betts et al. identified Blastocystis in all 38 water voles sampled in their study, observing a clear predominance of ST4 [131], and there were examples of both clades of ST4 among the sequences.

Efforts to disclose whether some Blastocystis subtypes are more commonly seen in individuals with GI symptoms than in gut‐healthy individuals have not produced any major breakthroughs. In one of our studies, it appeared that those harbouring zoonotic subtypes such as ST6, ST7 and ST8 might be more prone to experiencing symptoms than those who did not harbour these subtypes [117]. To this end, we published a case of ST8, where clinical and microbiological resolution was reached after treatment with TMP‐STX [87].

In one of our other studies, we found almost exclusively ST4 in patients with acute diarrhoea sampled in Denmark [92], which at first glance could suggest a link between ST4 and this clinical condition; however, the positivity rate in the study was low compared with that of other groups of individuals studied in Denmark, and instead of reflecting an association between ST4 and diarrhoea, this observation may indicate that ST4 might be more resilient to any gut microbiota disturbances experienced during periods of diarrhoea than other subtypes. It should be noted that all the ST4 sequences identified in that study clustered in Clade 1 (data shown in the article).

To this end it should be mentioned that the use of traditional PCR followed by Sanger sequencing, which is the backbone of the barcoding method [46, 47, 48] and which many teams have used to characterise Blastocystis, may have led to an underestimation of mixed ST colonisation. Indeed, our use of metabarcoding has disclosed that mixed subtype colonisation appears to be a quite common phenomenon [14], and we recently concluded that the combination of real‐time PCR with metabarcoding would be beneficial for epidemiological and surveillance studies [108, 132]. We also prioritised the development of subtype‐specific primers to map the extent of mixed subtype colonisation [133], which is an approach that can be taken by those who do not have access to ‘omics’ technologies.

Over the past few years, we and others have produced data indicating strong links between Blastocystis and gut microbiota signatures and a role for Blastocystis as an indicator of gastrointestinal health. This is dealt with in more detail in Section 4.

Dientamoeba

Aspects of Biology, Life Cycle, Diagnosis and Epidemiology

Although described by Jepps and Dobell as early as in 1918 [134] and although being an extremely common coloniser of humans, the number of articles listed in PubMed rendered when searching for ‘Dientamoeba’ is only about 429 at the time of writing (January 2023), which is about five times less the number of articles available on Blastocystis.

Despite its name, Dientamoeba is not an amoeba, but a flagellate that lost its flagellum [135]. Only one species, D. fragilis, is known. It is genetically related to Giardia, and shares a most common ancestor with Histomonas meleagridis, a protozoon that causes ‘histomoniasis’ in poultry. The observation of a cyst stage was reported relatively recently [136] but remains to be confirmed by additional teams. Indeed, microscopy of faecal concentrates from D. fragilis carriers would very likely reveal cysts of Dientamoeba, should these exist; the question remains as to whether anyone would recognise these. With regards to the sister taxon H. meleagridis, no cyst stage has been identified. Conspicuously, rather than surviving outside the host on cyst from, it would appear that H. meleagridis is transmitted by Heterakis gallinarum, a nematode of poultry [137]. Given the apparent lack of a cyst stage for Dientamoeba and given the genetic similarity to H. meleagridis, it is hypothesised that this parasite takes advantage of a transmission mode similar to that of H. meleagridis. A few years back, we found evidence of DNA of D. fragilis inside eggs of Enterobius vermicularis that we had surface sterilised with hypochlorite in order to remove the risk of contamination [138]. Although this does not prove that D. fragilis can be transmitted by pinworm, it suggests that further investigation of pinworm as vehicle for D. fragilis would appear relevant. In a registry‐based retrospective cohort study of 9945 patients tested for D. fragilis at the SSI between 2008 and 2011 we identified that mebendazole (MB) exposure was associated with increased risk of testing positive for D. fragilis [139], and as MB is practically only used to treat pinworm infections in Denmark, this could be interpreted as indirect evidence of an association between pinworm and D. fragilis.

We published the first data on D. fragilis in Denmark in 2007 [140]. We found a positive rate of 12% among patients with suspected enteroparasitic disease in the Copenhagen metropolitan area. As the international scientific literature at that time suggested that D. fragilis should be considered an intestinal pathogen in humans [141, 142], routine testing for D. fragilis was initiated as part of the general parasitological workup for patients with diarrhoea or other intestinal symptoms, such as abdominal pain. The diagnosis of D. fragilis typically relies on examination of permanently stained fixed faecal smears or DNA‐based detection using for instance real‐time PCR. In 2007, Verweij et al. published a real‐time PCR assay for specific detection of D. fragilis [143]; this assay was implemented in the Laboratory of Parasitology, SSI shortly after.

Six years later, we reviewed data from the routine testing for D. fragilis at SSI and identified a positivity rate of 43% across the more than 22,000 samples tested over a 4‐year period [144], with the positivity rate peaking at 71% in children aged ~7 years.

After (i) dismissing any statistically significant microbiological or clinical effects of MZ treatment of children with abdominal pain and D. fragilis colonisation [145], (ii) demonstrating that practically all children in institution in Denmark would be or become colonised by D. fragilis over a period of a year [146] and (iii) showing that the organism was more common in the background population than in patients with GI symptoms [90], it was decided to exclude D. fragilis testing from the routine parasitological workup panel at SSI. This is a clear example showing the importance of having the level of parastisim in the background population inform decisions on which parasites to test for in the CML.

Indeed, we investigated the presence of D. fragilis‐specific DNA in faecal samples obtained from a cohort of 142 0‐ to 6‐year‐old children that we had already tested for a diverse range of gastrointestinal pathogens [147]. Among the 108 children who had submitted two or more samples and thereby included in a longitudinal analysis, 32 tested D. fragilis‐negative on the first sample but positive later, and the last sample from each of the 108 children was positive [146]. Risk factors associated with testing D. fragilis‐positive including being > 3 years old and having a history of recent travel abroad.

Colonisation by D. fragilis appears to be very common in Denmark and other European countries. However, in countries far away from this region, such as Australia where comparable diagnostic tools are used for detection and where there may not be large differences in testing strategies or populations tested (e.g., in terms of age), the colonisation rate appears to be significantly lower (Table 4). It is therefore possible that the clinical and public health significance of D. fragilis colonisation differs according to geographical region. Indeed, it should be investigated whether the overall prevalence of pinworm infections mirrors the overall prevalence of D. fragilis colonisation; for instance, it may be so that the prevalence of pinworm in Australia is much lower than that seen in Europe; however, data on this remain scarce.

In a study aiming to investigate whether symptom relief could be obtained in D. fragilis‐positive individuals with IBS using either MZ or tetracycline, we noticed microbiological responses in 15 of 25 individuals (60%), all by MZ; a clinical response was observed in 7 of 22 patients (32%), all by MZ. Meanwhile, some test individuals were insufficiently treated by MZ [153].

We were involved in a patient case where GI symptom relief was achieved upon eradication of D. fragilis, which proved successful using PM but not MZ [154]. In another case already mentioned (Table 3), D. fragilis was eradicated only after administration of PM 500 mg + MZ 750 mg thrice daily/10 days.

A cluster of D. fragilis infections associated with peripheral eosinophilia (PE) in a family that had been sharing a meal of shashimi (raw fish) was presented by Gray et al. [155]. To my knowledge, Cystoisopora is the only gut protozoon that can trigger eosinophilia [156], an immune response typically induced by helminth infections and allergic reactions. We thought that the PE might reflect to exposure to live anisakids that might have been present in the fish consumed by the family [157].

Genetic Diversity in and Host Specificity of D. fragilis

What was not done in the case report by Halkjaer et al. [154] mentioned above was typing of the D. fragilis strain identified. D. fragilis comprises two genotypes, 1 and 2, of which Genotype 1 is by far the most prominent; not only in Denmark, but also in other countries [158, 159, 160, 161, 162].

In 2000, Johnson and Clark demonstrated cryptic genetic diversity in D. fragilis by applying restriction‐fragment length polymorphism (RFLP) analysis to SSU rDNA amplicons, establishing the existence of the two genotypes [163]. A few years later, Peek et al. used both PCR and RFLP and Sanger sequencing to screen D. fragilis‐positive individuals sampled in the Netherlands for genotypes (referred to as ‘haplotypes’ in that study), finding only Genotype 1 [164]. Similar work was performed by Stark et al. and our own group a couple of years later [158, 165] in Australia and Denmark, respectively. In the study by Stark et al., only Genotype 1 was identified, and in our own laboratory, we also only identified Genotype 1 (unpublished observations [in the abstract published in the Proceedings from the 5th European Congress on Tropical Medicine and International Health 24–28 May 2007 Amsterdam, the Netherlands by Stensvold et al. [158], there were no data from lab analyses (the work was ongoing at the time). However, the results are available in the MSc thesis by co‐author Kenneth Dinesen that can be made available upon request]). It is not known whether these two genotypes differ in terms of clinical significance. The genetic differences between the two genotypes across the entire SSU rRNA gene is ~3.5%.

Windsor et al. attempted sequencing of the Internal Transcribed Spacer (ITS) region of D. fragilis, but concluded that the value of this approach was limited because of intra‐strain genetic heterogeneity [166]. ‘C‐profiling’ was developed as a means of extracting useful data from sequenced ITS clones [167], but, to the author's knowledge, this method has only been used once ever since and on a very limited material [168], and therefore, the applicability and epidemiological relevance of this method remains uncertain.

In 2012, we noticed that two D. fragilis genotype 1 housekeeping genes had been amplified and sequenced, namely actin and elongation factor 1 alpha, both of which sequences were present in the NCBI database. Based on these sequences, we developed specific primers for PCR and sequencing and screened D. fragilis‐positive genomic DNAs extracted from stool for genetic variation across these two loci plus the SSU rRNA gene. Our data indicated that genetic analysis of these three D. fragilis housekeeping genes enabled clear distinction between the two known genotypes, and phylogenetic analysis of translated, concatenated sequences confirmed the phylogenetic position of D. fragilis. Meanwhile, integration of housekeeping genes in multi‐locus sequencing tools for D. fragilis would possibly have limited epidemiological and clinical values because of no further added genetic resolution [52]. In a multi‐centre study a few years later, this conclusion was further corroborated [162], when we applied six new genetic markers to D. fragilis‐positive samples from individuals from Italy, Denmark, Australia and Brazil. Here, only one of 111 samples was positive for Genotype 2, while the rest exhibited the genotype 1 profile; importantly, no further genetic resolution could be identified as potential markers for further discrimination.

Since then, the genetic diversity within D. fragilis has been studied only to a limited extent. David et al. identified only Genotype 1 in a study carried out in Brazil [169]. Oliveira‐Arbex et al. identified a D. fragilis positivity rate of 10.3% among 156 asymptomatic children in day care centres in São Paulo State, Brazil. The 16 positive samples were typed, identifying 14 Genotype 1 and 2 Genotype 2 sequences [161].

The data published to date bear witness of a low level of polymorphism and, as also mentioned by Caccio [170], are compatible with a clonal population structure of D. fragilis, which is in stark contrast to for instance Blastocystis. For H. meleagridis, the sister taxon of D. fragilis, two genotypes have been identified, and also here, Genotype 1 appears to be predominating, although with some intra‐genotype genetic variation [171].

To date, there are at least 52 SSU rDNA sequences (18S) available (Table 5) + quite a few sequences from the study by Windsor et al. [166], who sequenced the part of the ribosomal operon containing the ITS regions 1 and 2.

As Genotype 2 appears to be rare in studies involving human faecal samples, it could be speculated that this genotype would have a reservoir in one or more non‐human hosts, and screening faecal samples from non‐human hosts for D. fragilis therefore appears relevant.

The number of studies that involved testing of faecal material from non‐human hosts is still limited, and an overview of these studies is provided in Table 6. It should be noted that there is a handful of reports on findings of D. fragilis in rats and non‐human primates [176, 177, 178, 179]; however, as these data are based on morphology only, we have chosen not to include the data in Table 6. At the time of writing (January 2023), there are D. fragilis‐specific DNA sequences only from humans, pigs and budgerigars in the NCBI database.

As seen in Table 6, only genotype 1 has been identified in studies of non‐human hosts that involved genotyping. Hence, any reservoir for genotype 2 remains to be identified. The situation is somewhat similar to that of the microsporidium Enterocytozoon bieneusi genotype C, which has been linked to two conspicuous outbreaks in Scandinavia [2, 6] and observed in immunosuppressed individuals [180, 181], but which has otherwise only rarely been reported of in humans, and which has only been detected in a couple of instances in non‐human hosts (e.g., in two gorilla samples in Rwanda [182] and a handful of mice in Spain and in central Europe [183, 184]), despite quite extensive screening of various animals. However, E. bieneusi genotype C was identified in quite a few of the Swedish wastewater samples [33] tested for Amoebozoa and Blastocystis (unpublished data). If more DNA were available, the Swedish wastewater samples could be screened for D. fragilis genotype 2 with specific primers.

In the event that D. fragilis is indeed transmitted by pinworm, the question arises which transport organisms that might be used by D. fragilis when colonising non‐human hosts, which are not natural hosts of pinworm. It may be so that other nematodes might serve as vectors for D. fragilis.

Entamoeba

Entamoeba is a genus comprising aflagellated endobiotic taxa, most of which parasitise the guts of vertebrates. Within the Archamoebae, Entamoeba forms a sister taxon to the genus Pelomyxa, which comprises large, free‐living, flagellate, multinucleate amoebae [185] that can be found in anaerobic or microaerobic bottom sediments of stagnant freshwater ponds or slow‐moving streams. As a result of adaptation to the anaerobic gut environment, Entamoebas do not have mitochondria, but mitosomes, which are considered mitochondrial remnants in a way similar to what is seen for Blastocystis, in which they are called MLOs, but which lack a detectable organellar genome [186].

Entamoebas are motile, can phagocytose bacteria, and, at least in humans, they typically colonise the parts of the gastrointestinal tract that have most bacteria; that is, the oral cavity and the colon. A couple of species have invasive properties and have been linked to disease in humans and non‐human hosts (E. histolytica, Entamoeba nuttalli and Entamoeba invadens), although asymptomatic carriage may be common. A couple may be found primarily or only in the environment (e.g., Entamoeba marina, Entamoeba moshkovskii) (Table 7).

Morphologically, organisms belonging to the genus Entamoeba are easily recognised by their nuclear features. Parasitologists will be familiar with the ‘ring‐and‐dot’ appearance of one to several nuclei that can be observed in cysts. Most Entamoebas produce cysts, an exception being E. gingivalis. Mature cysts typically have either one, four or eight nuclei, depending on species (Table 7). Each nucleus has a karyosome, which may be centrally located (e.g., E. histolytica) or eccentric (e.g., E. coli), and peripheral chromatin is present, which may be distributed in a fine, homogenous distribution (e.g., E. histolytica) or in lumps (e.g., E. coli).

While it may be straightforward to identify an organism as belonging to the genus of ‘Entamoeba’, the differentiation of species within Entamoeba is associated with a lot more difficulty. Firstly, humans can host several established species of Entamoeba (Table 7), some of which are species complexes and some of which cannot be differentiated based on cyst morphology. For instance, E. histolytica, E. dispar, E. moshkovskii, E. nuttalli and E. bangladeshi all produce cysts of similar size and with four nuclei. E. nuttalli has only been reported once in a human [199], and there are still only scarce reports of E. bangladeshi [200, 201]. As E. nuttalli is and E. bangladeshi might be pathogenic (the clinical significance of E. bangladeshi remains unresolved and there are only six SSU rDNA sequences in GenBank), these two species will not be subject to separate discussions in this thesis. Secondly, surveys using only morphological identification have been challenged by the fact that cysts of each species of Entamoeba might exhibit a continuum of morphological features, especially features such as cyst size and number of nuclei, depending on the maturity of the cysts.

Given this situation, it may not be a surprise that Entamoeba taxonomy has been subject to two types of ‘error’ in the naming of species, which also pertain to the situation for other CLIPPs [187]: (1) reliance on a character that does not reflect underlying genetic divergence, leading to overestimation of diversity and the naming of invalid examples include relying on the host as a species‐specific character when in fact some Entamoeba species have quite a broad host range; and (2) perceived morphological simplicity means that genetic divergence is not always reflected in morphological differences, which leads to underestimation of diversity and assigning the same species name to quite different organisms, a situation that we refer to as ‘cryptic genetic diversity’ (see below). To this end, and as a clear example, both of these considerations also led to dismissal of the species name ‘Blastocystis hominis’ back in 2007 [26].

Nevertheless, over the past 20 years, the genetic universe of Entamoeba has been unfolding bit by bit (Table 7, Figure 4), and seven of the works included in this thesis [13, 33, 43, 44, 49, 50, 53] have contributed to what today is known about genetic diversity in Entamoeba hosted by humans and other larger mammals, such as cattle and pigs, and in waste water. State‐of‐the‐art terminology of Entamoeba involves the use of ‘species’, ‘subtype’, ‘ribosomal lineage’ and ‘conditional lineage’, depending on the amount of genetic diversity and the type of information available (sequence length, morphology, host, etc.; Tables 2, 7 and 8).

The use of NGS technology has assisted greatly in terms of mapping genetic diversity within Entamoeba [13, 33, 116, 203, 204, 205], but it is also clear that experience is needed in terms of interpreting such data. Quite a few sequences are quite short and deposited in GenBank as ‘Entamoeba sp.’, even in cases where it would be possible to assign both a species name and even a subtype name.

In the following, a selection of ‘non‐pathogenic’ Entamoeba species will be accounted for with focus on genetic diversity and host specificity. There will be examples of survey data for each species and these data are included mainly to provide a small impression of the global prevalence of the species.

Entamoeba coli

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba coli

Octonucleate cysts of Entamoeba have been identified in faeces from human and non‐human primates as well as from ungulates and rodents. Among these, E. coli is a common finding in human faecal samples, with survey positivity rates reaching almost 40% in countries such as Burkina Faso and Venezuela [206, 207, 208]. In Denmark, the E. coli colonisation rate may reach at least 14.6%, depending on the population studied [90, 105, 209, 210]. If not defeated by E. gingivalis, E. coli may very well be the most common species of Entamoeba to colonise humans, and most of the 26 wastewater samples analysed by metabarcoding were positive [33].

Although cysts of E. coli are generally larger than those of E. histolytica, Dobell and Jepps [211] claimed that it is impossible to use the size of the cyst alone for the differentiation of these two species. Hence, the number and structure of the nuclei in mature cysts may be the only hallmark to separate these two species morphologically.

Entamoeba coli‐specific primers were developed for PCR‐based detection and molecular characterisation by Stensvold et al. [49] and later used by for instance Chihi et al. to detect and differentiate E. coli [212]. A different approach to specific molecular detection of E. coli was taken by Matey et al. who developed a nested PCR for specific detection [213]. This methodology was used by Matsumura et al. in a survey of healthy Indonesian school children, where a positivity rate of 44% was reported [214].

Genetic Diversity and Host Specificity of Entamoeba coli

A bit more than ten years back, the extensive use of ribosomal gene sequencing from a diverse set of E. coli‐positive samples enabled us to identify two different lineages within E. coli, and this led to the first published evidence of cryptic genetic diversity in E. coli with a suggestion to divide the species into subtypes (ST), ST1 and ST2 [49]. Unfortunately, data on cyst morphology were not available for study to identify any differences in cyst size between ST1 and ST2 that could explain the early findings of Dobell [215] and Matthews [216] that suggested a bimodal cyst size distribution among E. coli cysts.

When the data were published (2011), it appeared that ST1 was genetically homogenous compared with ST2, and that only sequences from humans were in ST1, whereas ST2 had sequences from both human and non‐human primates. Recently, we revisited the genetic diversity of E. coli and investigated two different regions of the SSU rRNA gene to minimise the risk of losing important information [44].

Our phylogenetic analyses this time pointed towards not only two but three subtypes (Figure 5). This study confirmed the genetic homogeneity of ST1 and disclosed further heterogeneity in ST2. However, a non‐human primate E. coli sequence appeared in the ST1 clade, namely one from a Mandrillus leucophaeus (FR686410), and AB749457 was found in a macaque in China, which means that ST1 is not exclusive to humans. The other NHP sequences clustered with ST2.

To date, a good 200 E. coli‐specific sequences have been deposited in GenBank, and these are all SSU rDNA sequences (Table 9). Across the six near‐complete E. coli SSU rDNA sequences in GenBank available, up to at least ~12% genetic difference can be appreciated.

When reviewing GenBank data for the development of this thesis, I took note of a 527 bp‐long sequence deposited in 2020, namely MW026738 from a human sampled in the Amazonas regions, Brazil. Phylogenetic analysis of this sequence could indicate that it might represent a new subtype of E. coli, which in that case would be ST4 (Figure 6). If possible, a longer sequence should be produced for this strain to enable a more robust phylogenetic analysis. But even then, it might not be possible to tell with much confidence whether the sequence would represent a basal branch of ST2 or indeed a new clade. This leaves us in the same position as we were with what is now E. coli ST3; originally we had sequence S2702 that appeared as a deep branch at the base of ST1, and that is what we called it [49]. Even though it was a complete sequence it was not enough for us to call it a new subtype until we had additional sequences from multiple samples that clustered together with S2702 [44].

When scrutinising the relatively large genetic variety within ST2, subclades seem to form within the subtype, and there are examples of specific variants that have been found across the globe in independent studies. One such example is made up by the sequences MW819961, MK559462 and AB749456. These are all from monkeys and share SNPs not seen in other ST2 sequences (data not shown).

In their study of protist parasites in human‐habituated mountain gorillas (Gorilla beringei beringei), humans and livestock from Bwindi Impenetrable National Park, Uganda, Nolan et al. identified a few sequences (KY658155–KY658157, KY658172 and KY658177–KY658179) that they deposited as ‘Entamoeba sp.’ in GenBank in 2017 [217]. When included in a phylogenetic analysis, these sequences cluster with E. coli ST2 (Figure 5) [44].

The ST3 clade holds only a limited number of sequences, which are all from humans, and all of which stem from either South America, Africa, West Bank, or Iraq, including two from human samples from the Uganda study mentioned above [44, 217]. The ST3 data from Iraq have not been included in any publication so far. The vast majority of ST2 sequences identified to date are also from South America or Africa, while ST1 appears to have limited geographical restriction [44]. Studying the genetic diversity of Entamoeba in humans in Brazil, Calegar et al. found evidence of all three subtypes, including ST3 (MW026736); their data were in support of extensive levels of intra‐subtype genetic diversity in E. coli [218].

The species name ‘E. coli’ has been applied mainly to octonucleated cysts found in faeces from human and non‐human primates, and among the 214 sequences included in Table 9, 183 E. coli sequences are from humans, and 21 from NHPs; the one remaining sequence, for which information on host species is available is from a rodent (see below). However, as early as in 1928, Kessel reported the finding of E. coli in a pig [219], and there are recent reports on E. coli commonly observed in pigs sampled in Colombia [220, 221]. Unfortunately, no molecular data were included for E. coli in the two Colombian studies, and so it remains to be confirmed whether the parasites referred to as E. coli were in fact E. coli, another octonucleated‐cyst producer such as Entamoeba RL7 or even other species. E. coli cysts with a diameter as small as 11 μm have been reported [211], and those could be difficult to separate from other Entamoebas in pigs, such as E. polecki or E. suis, in case the nuclei cannot be clearly discerned.

In non‐human primates, E. coli is a frequent finding, and both apes and monkeys appear to be natural hosts. Positivity rates of 20%–70% have been reported in sun‐tailed monkeys [222], macaques [223, 224], baboons [225], gorillas [226, 227] and chimpanzees [228].

Experiments seeking to infect rodents with E. coli produced discrepant results. Both Kessel [229] and Regendanz [230] reported successful experimental infection of rodents, whereas Neal [231] failed to be able to establish E. coli in mice and rats; meanwhile, experimental infection with E muris, which is closely related octonucleate‐cyst producer was achieved. It is possible that different subtypes of E. coli have been used, and it may be possible that one or more may be able to colonise/infect rodents. Indeed, Ponce‐Gordo et al. deposited a SSU rDNA sequence in the NCBI database referred to as ‘E. muris’ from a rat (FN396613); however, in our phylogenetic analysis, FN396613 clusters with E. coli ST2 [44], so a plausible theory would be that ST2 can be hosted by rodents, whereas ST1 cannot. Indeed, we identified E. coli ST2 in a chinchilla [49], which could support the hypothesis of ST2 being able to colonise rodents.

In our recent study of Swedish wastewater samples, both ST1 and ST2 were found, with ST1 being more common [33].

Apart from E. muris, a couple of other ribosomal lineages clustering with E. coli, namely Entamoeba RL7 and Entamoeba RL11. RL11 was found in a field vole by Jacob et al. [28], while RL7 was identified in both humans and langurs [28, 49]. Given the relatively consistent pattern between phylogenetic topology and number of nuclei in mature cysts, it would appear reasonable to assume that both RL7 and RL11 are octonucleate cyst producers; this has been confirmed for RL7, whereas no morphological data are yet available for RL11.

On the basis of host and cyst morphology, other octonucleate cyst‐producing species have been described, namely Entamoeba cavie, observed in laboratory guinea pigs; Entamoeba cuniculi, a parasite of the rabbit and Entamoeba gallinarum from chicken and turkeys [232].

The host specificity, distribution and genetic diversity of E. coli subtypes call for further investigations to delineate the role of this species in human health and disease and to identify routes of transmission.

Entamoeba dispar

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba dispar

The species E. dispar was introduced by Emilie Brumpt in 1925 after recognising that an organism that could not morphologically be differentiated from E. histolytica did not give rise to intestinal symptoms in colonised individuals [187]; moreover, it did not produce invasive disease in cats, which at that time were typically used in studies to prove the invasiveness of E. histolytica. His observations were largely ignored over the next many decades, possibly due exactly to the fact that these two species could not be told apart.

It was only about 50 years later that isoenzyme analysis and molecular methods assisted in providing evidence in support of Brumpt's observations [20, 21, 233, 234, 235, 236]. Sequencing of the near‐complete SSU rRNA genes of E. histolytica and E. dispar revealed a genetic difference between the two of ~1.5%. After several studies focussing on the distribution of and ways to separate the two species, Verweij et al. developed a real‐time PCR for simultaneous detection of E. histolytica, Giardia lamblia, and Cryptosporidium parvum in faecal samples using multiplex real‐time PCR in 2004 [237], which heralded a new area in the diagnostic parasitology laboratory. This assay took advantage of the fact that E. histolytica and E. dispar could be separated based on SSU rRNA genes, and so the 22 bp‐long E. histolytica‐specific TaqMan probe was developed so that it annealed to a relatively polymorphic region (no less than six mismatches) separating E. histolytica from E. dispar. This type of strategy and methodology informed the development of numerous molecular analyses now in place in most modern parasitology laboratories [11].

E. dispar appears to be quite a common parasite of cosmopolitan distribution. In a survey of 199 healthy school children sampled in Nigeria, we identified a positivity rate of 18.6% [238]. Among human immunodeficiency virus (HIV)‐infected patients followed up in Denmark, the positivity rate using the same method (real‐time PCR) was 10.4% [105]; however, in a random subset of samples tested in our laboratory, we only saw a positivity rate of 2/889 (0.2%) [239], indicating that HIV‐infected individuals might be a population particularly prone to developing E. dispar colonisation compared with non‐HIV infected individuals, at least in this country. To this end, we identified E. dispar in only one of 41 Syrian asylum seekers in Denmark [209]. We found E. dispar‐specific DNA in 11 of 26 (42.3%) wastewater samples from Sweden [33].

Among 175 human patients with intestinal symptoms sampled in Egypt, the positive rate was 40.5% by PCR [240], and data from a study from South Africa indicated the presence of E. dispar in 14.7% of 170 patients tested [241]; however, in both studies, the diagnosis of E. dispar was done only by gel inspection of PCR products with no confirmatory DNA sequencing.

There is one DNA sequence in the NCBI database (KX357142; Table 10) deposited by a team in India that is from a pus sample from a human. In the metadata, the title ‘Amoebic Liver Abscess caused by Entamoeba dispar and Staphylococcus aureus’ has been provided, but to my knowledge, no such article has been published to date. There are additional sporadic sources of information that could indicate that some strains of E. dispar could be pathogenic [242, 243, 244].

Genetic Diversity and Host Specificity of Entamoeba dispar

To my knowledge, the species has been identified mainly, if not only, in primates. In non‐human primates, positivity rates from surveys have reached at least 18.1% in mixed study populations of apes and monkeys [245]. Pomajbikova et al. found a positive rate of 16% specifically in Pan troglodytes schweinfurthii [228]. There are a limited number of sequences (OP453103–OP453107) in the NCBI database for which ‘human and dogs’ have been entered in the ‘host’ field, and so it remains unknown to most whether these sequences are from humans and/or from dogs in the absence of an accompanying article. Data that could indicate an even much higher prevalence among NHPs are those published by Dong et al. in their impressive study from China [246]; however, the exact positivity rates are difficult to decipher in that article. It could be speculated that NHPs may constitute a reservoir for human carriage of E. dispar.

To date, at least 134 SSU rDNA sequences of E. dispar are available in GenBank (Table 10). Although the within‐species diversity of E. dispar appears to be limited (~0.5%) and confined to a handful of SNPs, these are organised in a way that could fuel a hypotheses of the existence of at least two separate subtypes (Figure 7). There is currently, however, limited data, so it may be premature to hypothesise on differences in geographic distribution and cryptic host specificity.

Entamoeba gingivalis

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba gingivalis

Being a digestive‐tract Entamoeba species with no evidence of a cyst stage, E. gingivalis is a parasite of the oral cavity passed on by direct buccal contact and/or by saliva or contaminated mouth utensils or food [247, 248]. As the name implies, this amoeba lives in the gingival areas around the teeth. The textbooks of Levine [232] and Noble and Noble [249] concur on E. gingivalis being a harmless commensal, although ‘often present in diseased gums’.

The parasite appears to be a quite common finding and may be the most common Entamoeba species colonising humans. PCR‐based testing of saliva and dental plaques revealed colonisation in 1 of every 5 women sampled in Iran [250]. Also in Iran, PCR testing revealed an overall positivity of rate of 11.7% in randomly selected adolescents, with colonisation being statistically significantly linked to (i) a gingival index that indicated severe inflammation and (ii) having decayed, missing and filled teeth [248]. In neighbouring Turkey, and also using PCR, Yaseen et al. identified colonisation rates of 88.9%, 84.9% and 47.9% in patients with periodontitis, patients with gingivitis and in healthy individuals, respectively [251]. Until recently, there was a lack of data from the ‘background population’. However, in our recent survey of Tanzanians with non‐oral/non‐dental diseases (N = 52), 31% of the study individuals tested positive for E. gingivalis‐specific DNA extracted from oral washings [43].

Recently, Keeler et al. identified E. gingivalis as a likely host of human‐associated redondoviruses, which have a high prevalence in healthy humans, but the abundance of which is increased in patients with periodontitis, acute illness and severe Coronavirus disease 2019 [252]. We and others have investigated associations between bacterial communities and the presence/absence of E. gingivalis in mouthwash samples, subgingival plaque or other types of samples from the oral cavity. In our laboratory, we found that, despite higher microbial diversity in E. gingivalis carriers, the top‐10 most common bacterial genera were almost similar; only E. gingivalis carriers were more likely to be colonised by Aggregatibacter [43], which has been associated with periodontal disease. Moreover, Neisseria spp. were enriched in carriers relative to non‐carriers. These observations confirmed those of Koller et al. [253], who speculated that E. gingivalis might promote oral cavity colonisation by phagocytosis‐resistant bacteria [253].

Genetic Diversity and Host Specificity of Entamoeba gingivalis

At least two subtypes, ST1 and ST2, have been observed. In 2018, García et al. introduced evidence of a ‘new subtype’ of E. gingivalis, which was named ‘E. gingivalis ST2, kamaktli variant’ (e.g., KX027297) [254], and which only shared 89% similarity with another E. gingivalis sequence (KX027298); the latter sequence was referred to a ST1 and had high similarity (99.58%) to D28490, originally deposited back in 1995 by Yamamoto et al. [255]. This new variant was identified in patients with dento‐oral diseases in Mexico. The D28490 sequence had been obtained from an ATCC (American Type Culture Collection) strain from a subgingival space of adult with periodontal disease (geographical data lacking). The authors suggested that this new variant could in fact be the SSU rDNA sequence of one or more of the species Entamoeba pyogenes, Entamoeba canibuccalis, Entamoeba equibucalis, and Entamoeba suisginvalis, but this is not possible to confirm, as these species names were introduced in the absence of sequence data.

Interestingly, in August 2022, rDNA (18S‐ITS1) sequences of E. gingivalis were published from a study of individuals sampled in Austria. Here, the research team found evidence of what in GenBank is proposed as a new subtype, ST3. However, the publication linked to the GenBank accession numbers (Table 11) is not yet available. Nevertheless, I include a phylogenetic analysis that is indeed in support of the new subtype (Figure 8). As the sequences from Austria only included the very 3′‐end part of the SSU rRNA gene along with the ITS1 region (when García et al. launched the idea of the existence of two subtypes, they based it on complete SSU rDNA + ITS analysis), it is currently unknown, how much ST3 differs genetically from the two other subtypes across the SSU rRNA gene alone.

The intra‐subtype diversity within the three subtypes identified today remains relatively unexplored given the few studies that involve molecular characterisation of E. gingivalis (probably only a dozen of articles according to PubMed). In our study of E. gingivalis sequences from the Tanzanian mouthwash samples, which all belonged to ST1, a single SNP was identified and this SNP was also picked up by García et al. [254]. An overview of the sequences obtained in our Tanzanian study is provided in Figure 9. As seen, some of the consensus sequences for test individuals for whom two consecutive sequences were available (reflecting two different time points of sampling), were not completely identical (e.g., the sequences for the individuals with the IDs ‘P1’ and ‘C53’). However, it should be borne in mind here that the consensus sequences were generated from PCR products sequenced by ILLUMINA sequencing technology and that for some samples, the number of sequence reads available might have been quite small. This might have given rise to one or two SNPs across the consensus sequence that would be due to sequencing error rather than ‘true’ polymorphism.

In terms of relative distribution, data are still scarce, but ST1 appears to be the predominant version of E. gingivalis [256]. Nevertheless, ST2 has been found in both Mexico [254], Turkey (Orsten et al., unpublished GenBank entries OP456215 and OP422447), and in DNA from bronchioalveolar lavage samples in Denmark (Stensvold et al. unpublished observations) suggesting a potentially global distribution.

The level of genetic diversity within E. gingivalis is reminiscent of the level observed within E. coli (between 10% and 15%), and it is highly likely that it would be useful to consider the three subtypes as three distinct species.

A PubMed search on E. gingivalis in dogs renders two articles [197, 257], while none when cats are searched as hosts. For now, the host spectrum of E. gingivalis appears to be extremely narrow. To date, 50 SSU rDNA sequences have been deposited in GenBank (Table 11), all being from humans.

Future research might reveal whether the two subtypes differ in terms of clinical significance and epidemiological features.

Entamoeba hartmanni

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba hartmanni

Entamoeba hartmanni is a quadrinucleate cyst‐producing amoeba, with cysts resembling those of, for example, E. histolytica, apart from the fact that they tend be smaller, with a size of approximately 8 μm in diameter [232], and with the peripheral chromatin of the nuclei tending towards being arranged in lumps [232]. For many years, this species was confused with E. histolytica (considered small‐race E. histolytica [258, 259, 260]), and so, data from surveys based on morphology only should be interpreted with caution. Generally, however, E. hartmanni appears to be a less common than E. coli; not only in humans but also in non‐human primates.

Matsumura used nested PCR followed by species‐specific amplification of species of Entamoeba, identifying a positivity rate of 31% in healthy school children in Indonesia [214] and an association between E. hartmanni colonisation and loose stools; it should be mentioned, that no sequence data were available for this study, so it remains unknown which subtype(s) (see below) might have been involved. Entamoeba hartmanni generally appears to be less common than E. coli. In Denmark, positivity rates in different study populations based on traditional microscopy have been found to range between 0.8% and 7.3% [90, 209], and in our metabarcoding analysis of Swedish wastewater samples, E. hartmanni was one of the least common parasites detected, with a positivity rate of 15% [33].

Genetic Diversity and Host Specificity of Entamoeba hartmanni

When we did our first study on molecular characterisation of E. hartmanni from humans and NHPs [49], we did not obtain data that could indicate the existence of more than one species; neither did other authors after us [13, 33, 218]. In our most recent study, however, which involved characterisation of a part of the gene that was different from the one previously investigated, we found evidence pointing towards the existence of three subtypes (Figure 10) [44].

At the time of writing, there are at least 84 SSU rDNA sequences of E. hartmanni in the NCBI database, 64 of which are from humans (Table 12). There are 10 and 5 sequences from NHPs and domestic dogs, respectively, and for five sequences, no information on host is available. The five sequences from dogs were published only in October 2022, but no accompanying article is available to date. The data stem from Iraq (accession OP688358–OP688382), and although Sanger sequencing is indicated as the sequencing method, we do currently not know which type of DNA amplification (including any pre‐DNA extraction procedures) that was used to produce the data.

DNA‐based evidence of E. hartmanni was moreover found in pigs by our team in a recent study of intestinal parasites shared between humans and pigs [13], and indeed, the E. hartmanni sequences obtained from pig samples were identical to E. hartmanni sequences produced from human stool samples. More specifically, all 10 pigs that were found positive had E. hartmanni similar to FR686375, a ST1 sequence from a human. This was the first report of a finding of E. hartmanni in pig faeces, thus expanding the host spectrum for this species. In 1928, Kessel [219] reported Entamoeba cysts with 1–4 nuclei in pigs, and the cysts were between 5 and 12 μm, which would include the E. hartmanni cyst size range.

Studies of NHPs have identified both apes and monkeys as what would appear to be natural hosts. Positivity rates of 11%, 27%, and 51% have been found in Macaca cyclopis in Taiwan [262], in Gorilla gorilla beringei (mountain gorillas) sampled in Rwanda [227], and in Pan troglodytes schweinfurthii in Tanzania [228], respectively; molecular methods were included in the studies from Taiwan and Tanzania.

In the previously mentioned study on protist parasites in human‐habituated mountain gorillas in Uganda, Nolan et al. [217] deposited a number of ~500 bp‐long sequences in GenBank that cluster together with E. hartmanni (Figure 11) [44]; however, as these are currently referred to as ‘Entamoeba sp.’ in GenBank, these sequences were not included in Table 12.

Hence, based on DNA data, three major groups of hosts of E. hartmanni have now been identified, namely primates, pigs, and dogs. The last group may be a bit surprising, as Entamoeba colonising the distal part of the digestive tract is rarely seen in canids. Further sampling of canines and other carnivores is warranted for confirmation.

Entamoeba moshkovskii

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba moshkovskii

Since its description in 1941, Entamoeba moshkovskii has been identified mostly in the environment. Reports on a potential link between E. moshkovskii colonisation and the development of GI symptoms such as dysentery exist [263, 264]; however, as the species has been known for more than 80 years, and as its pathogenicity still remains to be demonstrated, I have included the species in this account of CLIPPs.

We have not identified E. moshkovskii in any human sample by any of the methods in use in our lab for the almost 20 years that I have been working in the Laboratory of Parasitology at SSI; nor do I remember finding it in samples from any non‐human host (NHPs, pigs, cattle, muskox, deer and rodents). Contrasting my own personal experience are the many reports on human E. moshkovskii colonisation/infection from ‘warmer’ parts of the world. López et al. reported a positivity rate of 25.4% among children sampled in a district in Cundinamarca, Colombia [265], and even higher rates have been identified in Australia [264]. The ubiquity of the species in sewage samples from Scandinavia coupled with absence of colonisation of mammals in this region is intriguing when compared with the situation in other climate regions, where mammalian colonisation appears to be common, such as India and Bangladesh [198, 263, 266, 267], and where there are examples of studies not identifying Entamoeba in sewage samples [268]. What are the factors driving mammalian colonisation? Also, while most Entamoeba species seen in humans have also been documented extensively in NHPs, to my knowledge there is no DNA sequences from E. moshkovskii found in NHPs available in GenBank.

As the morphology of E. moshskovskii is overlapping with that of other quadrinucleate Entamoebas infecting and colonising humans, quite a few DNA‐based methods have been developed to facilitate differentiation between these [16, 269, 270, 271]. We and others have found NGS‐based methods useful to detect and differentiate E. moshkovskii from other Entamoebas present in complex matrices [16, 33, 205].

Genetic Diversity and Host Specificity of Entamoeba moshkovskii

In the book ‘Protozoan Parasites of Domestic Animals and of Man’, Levine wrote as follows, ‘This species occurs in sewage. It is not a parasite of animals, but of the municipal digestive tract’ [232], indicating that this species would be free living only. Indeed, in our study of Swedish wastewater samples, we identified in E. moshkovskii in all of the 26 sample analysed [33], which not only documents the presence of E. moshkovskii even in Northern Europe but also points to its ubiquity in ‘the municipal digestive tract’. Nevertheless humans, cattle and turtles have been identified as hosts [28], and there are E. moshkovskii SSU rDNA sequences in the NCBI database from pigs (e.g., MW926950), a snake (MN536488), green June beetle (MN536495) and from American cockroaches (e.g., MN536492) (Figure 12). A number of near‐complete SSU rDNA sequences have been deposited in GenBank from human stool in Iraq, which sequences are allegedly similar to sequences found in faeces from dogs in Iraq; these sequences are included in Figure 12.

It therefore seems likely that this Entamoeba is both free living and parasitic. Studies demonstrating E. moshkovskii in culture from humans and other hosts would be helpful for confirmation.

At the time of writing, 267 SSU rDNA sequences of E. moshkovskii are available in NCBI's nucleotide database (Table 13). Only 26 of these represent sequences longer than 1.45 kb, thereby covering ~80% of the gene or more (Figure 12), and the vast majority of these are from either the United States or Iraq.

In our study of wastewater samples, we identified two genetic variants of E. moshkovskii differing by 6.7%, which indicates substantial genetic diversity within the species [33], and one of which was novel. One variant (MN498050) was found across all the samples tested, and the consensus sequence shared 99.83% similarity with sequences such as MN536502 from freshwater sediment from the United States and MN536492 from a cockroach sampled in the United States. The novel variant (MN498051) shared only 94.07% similarity with strains such as KP722601 isolated from human stool in Iraq.

Of course, one of the important questions here is whether environmental strains are genetically different from the strains colonising humans, or at least whether human (and other mammalian) strains are in the same clade(s). Based on our analyses, two major clades exist, and both clades contain strains of both environmental and animal (including human) host origin; however, sequences from humans belonging to the lower clade in Figure 12 are yet to be identified, and sequences of human and porcine origin cluster primarily with the ‘Laredo’ strain (data not shown). These observations could easily indicate cryptic host specificity in E. moshskovskii.

Entamoeba polecki

Aspects of Biology, Diagnosis and Epidemiology of Entamoeba polecki

Prowazek was the first to describe Entamoeba polecki, and 2 years later, in 1914, Swellengrebel introduced the species Entamoeba chattoni, which he had observed in NHPs [272].

Entamoeba polecki was considered synonymous with E. suis by Levine among others [232]. Meanwhile, in 2006, Clark et al. provided evidence of a uni‐nucleate cyst producer observed in pig faeces that was genetically much more related to E. gingivalis than to E. polecki, indicating that suids may host at least two species of uni‐nucleate cyst producers namely E. suis (Hartmann, 1913) and E. polecki [202].

This species is much less commonly reported than other species of Entamoeba of similar size. Whether this is because of factual differences in distribution or because of morphological confusion remains unclear. Advanced morphological experience is required to tell the cysts of E. polecki apart from those of Entamoebas of similar size, and DNA‐based detection and differentiation appear more and more relevant, as more and more laboratories turn to molecular diagnostics. It is doubtful, however, that even modern laboratories screen for E. polecki specifically, but these may be picked up by metabarcoding methods, for instance, although we failed to detect E. polecki by metabarcoding in three wastewater samples that were identified positive by E. polecki‐specific PCR and sequencing.

Surveys involving the reporting of E. polecki in humans are very scarce. Based on microscopy of faecal concentrates, Desowitz and Barnish identified a positivity rate of 19% among 184 children sampled in Papua New Guinea [273], whereas Park et al. reported a positive rate of 1.1% among children sampled in Bat Dambang in Cambodia [274].

Genetic Diversity and Host Specificity of Entamoeba polecki

Nearly 100 years after the description of E. polecki, Verweij et al. proposed a new terminology for E. polecki [275], which involved the recognition of four subtypes (ST1–ST4) within the species (Figure 13); this initiative appeared relevant because of the overlap in morphology and host specificity (see Table 14). All of the four subtypes have been observed in humans. E. polecki ST2 was previously known as E. chattoni, and E. polecki ST3 has been referred to as E. struthionis [49, 53, 275]. The four subtypes differ genetically by up to at least 5% across the SSU rRNA gene.

In 2018, we studied SSU rDNA sequences from 18 stool samples that had been diagnosed with uninucleate Entamoeba cysts at the Public Health Agency of Sweden [53]. These sequences were obtained by applying a single‐round PCR to genomic DNA extracted directly from stool, using a primer pair (UNINUC_400F and UNINUC_1050R) designed to detect and differentiate subtypes of E. polecki by sequencing of the PCR products. Using the same PCR method, we also obtained sequences from six pig faecal samples and three environmental samples (wastewater). Mainly ST4 was seen in the samples from humans, but ST2 and ST3 were seen in a couple of instances. The six pig samples mostly had mixes of ST1 and ST3. Notably, ST4 was not found in any of the three wastewater samples; here, ST1–ST3 were found, with one sample being positive for two subtypes (ST1, ST3).

The host range of the four subtypes of E. polecki is summarised in Table 14 and overview of E. polecki‐specific SSU rDNA sequences currently available in the NCBI database is available in Table 15. Briefly, ST1 and ST3 have been observed mainly in humans and pigs; ST2 mainly in humans and NHP; and ST4 mainly in humans.

Interestingly, two ‘Entamoeba polecki’ sequences were recently deposited as OP919601 and OP753638, originating from a dog and a cat, respectively, sampled in India (Table 15). Moreover, a sequence from a pig sampled in Germany was deposited in 2019 as MK801450 (Table 15). The latter sequence is likely a chimaera (a sequence artefact [chimaeras arise during PCR amplification, usually when there are two distinct subtypes in the DNA sample and when there is incomplete replication of a DNA strand during a cycle [27]]), as it would appear to match AJ566411 (ST3) at both ends but AF149913 (ST1) in the middle, and the hypothesis of it being a chimaera is supported by the fact that it sits on a long branch in Figure 13, looking like a deep branch of ST3 that could potentially distort the position of the OP sequences. The MK801450 sequence had been produced in a study that used metagenomics to study parasite distribution in pig faeces [116]. The analysis shown in Figure 13 was therefore repeated—this time without MK801450 (Figure 14).

Based on phylogenetic analysis of 562 nt positions (Figure 14), one might argue that a hypothesis of OP919601 representing a new subtype could be supported. However, in the analysis, the bootstrap value for ST1 could be higher, and therefore, similar to the situation accounted for in 2.3.1., the position of this sequence will remain unresolved until longer sequences are produced and/or more sequences accumulate that share 99%–100% similarity with OP919601. A somewhat similar situation is seen for MW718195; here, however, the deposited sequence is very short (259 bp).

Of note, two of the sequences in Table 15 were deposited in GenBank as Entamoeba polecki, (AB845670, AB845671); however, these are not Entamoeba‐specific sequences.

The geographical distribution of E. polecki is not well known, but positive samples have been observed from humans with recent traveling in Europe (Italy, Spain), Africa (Eritrea, Somalia, Ethiopia and Kenya) and Asia (Afghanistan) [53]. Although there is a couple of reports on observations of E. polecki in humans in the Americas, no DNA data have yet been made available to corroborate the findings. The one DNA sequence from America available to date is from a pig (AF149913) and represents ST1.

It has not been possible to identify any cryptic host specificity within any of the subtypes so far based on SSU rDNA analysis; analysis of other genes appears relevant to investigate this further.

Entamoeba in Ungulates

The number of published studies on Entamoebas in ungulates that used molecular methods for detection and differentiation is relatively low. In 2010, we published data on Entamoeba in ruminants obtained by conventional PCR and Sanger sequencing [50]. Prior to the molecular work, cysts had been isolated from faecal samples by sucrose gradient centrifugation, and all cysts were uni‐nucleate. Near‐complete SSU rDNA sequences were obtained for six samples from cattle (n = 2), sheep (n = 2), reindeer (n = 1) and roe deer (n = 1). We argued that the sequences obtained from cattle, sheep and reindeer might be considered different genetic variants (genotypes) of Entamoeba bovis, and these data were the first DNA data made available for this species. Prior to our study, only morphological data were available for E. bovis. Meanwhile, the taxonomical status of the sequence obtained from the reindeer was unclear, as it might be considered a separate species based on the phylogenetic analysis.

There are currently 376 SSU rDNA sequences of E. bovis for which information on host species is available (data available on GitHub https://github.com/Entamoeba/DMSc‐Thesis/blob/main/Entamoeba%20bovis.xlsx). Most sequences are from China (n = 290), but a few sequences are available from Australia, Brazil, Japan, and Sweden. The host species for E. bovis identified so far are summarised in Table 16. As seen, none of the sequences have been observed in non‐artiodactyl hosts.

In our recent study of pigs [13], Entamoeba polecki was commonly seen, and, as expected, no subtypes other than ST1 and ST3 were observed. A minor proportion of the pigs (3%) were positive for E. hartmanni. No evidence of E. suis or other Entamoebas was obtained.

Our data could indicate that synanthropic and wild ungulates are commonly colonised by species of Entamoeba; however, of the hosts sampled to date, only pigs appear to be able to host Entamoebas that can colonise humans, namely E. hartmanni and E. polecki.

Quite a few other species of Entamoeba have been described based on samples from ungulates; however, in the absence of sequence data, the validity of species names such as Entamoeba ovis, Entamoeba dilimani, Entamoeba bubalus, Entamoeba equibucalis, Entamoeba suigingivalis, Entamoeba gedoelsti and Entamoeba caprae remain unconfirmed.

Endolimax

Together with Entamoeba and Iodamoeba, Endolimax belongs to the Archamoebae, a group of anaerobic free‐living or endobiotic protists that constitutes the major anaerobic lineage of the supergroup Amoebozoa [284]. The genus of Endolimax was described by Kuenen and Swellengrebel in 1917 [285]. Several species of Endolimax have been described in a range of host groups, including mammals, birds, reptiles, amphibians and insects; for a comprehensive list of the species reported, see the review by Poulsen and Stensvold [31].

First described by Wenyon and O'Conner in 1917 [286], Endolimax nana is the smallest of the Archamoebae commonly infecting mankind, which is somehow reflected in its name. The mature cyst stage typically measures about 8–10 μm [249] and contains four nuclei that do not have peripheral chromatin, which make them relatively easy to recognise on light microscopy. Indeed, the parasite is usually detected by microscopy of faecal concentrates obtained by the traditional FECT. Specific primers have been developed and published by our groups for molecular detection and characterisation [31, 32, 212].

Aspects of Biology, Life Cycle, Diagnosis, and Epidemiology of Endolimax nana

Positivity rates of about 30%–40% are not unusual in surveys carried out in some parts of the world, including Mexico [287], Colombia [288] and Brazil [289], but are typically somewhat lower. Meanwhile, in a survey of 3374 children sampled in rural Côte d'Ivoire in the beginning of the millennium, 82.6% of the children tested positive for the parasite [290].

Based on a metanalysis, we recently estimated that 13.4% of the global gut‐healthy population might be positive for E. nana, while only 3.4% of patients with GI symptoms may carry the parasite [31]. In Denmark, positivity rates of up to 7.5% have been seen, depending on study population [10, 90, 105, 209].

Using metabarcoding, we showed the presence of E. nana‐specific DNA in 10 of 26 sewage samples from Sweden [33].

Apart from E. nana, Endolimax piscium is the only species of Endolimax for which SSU rDNA data is available to date (Table 17). Endolimax piscium was described in 2014 by Constenla et al. [291] as a cause of granulomatous disease in cultured Senegalese sole.

Genetic Diversity and Host Specificity of Endolimax nana

Phylogenetically, Endolimax clusters closely together with Iodamoeba within the Mastigamoebidae B group of the Archamoebae. By combining PCR and TA cloning and Sanger sequencing and PacBio sequencing of pooled PCR products, we were recently able to describe the existence of two ribosomal lineages of E. nana [32] (Figure 15). These two lineages differ genetically by up to 16%. Surveys identifying the relative positivity rates of these two lineages among Endolimax‐positive individuals are still to be carried out with a view to identifying any differences in the epidemiology and clinical significance of the parasite.

It is clear from our phylogenetic analyses that Endolimax and Iodamoeba are sister taxa and cluster within Mastigamoebidae B, which confirms the findings of Pánek et al. [284]. However, despite the use of near‐complete SSU rDNA sequences, it is still not possible to identify whether E. piscium is congeneric with E. nana. Near‐complete SSUrDNA sequences from more species of Endolimax should be obtained and included in phylogenetic analysis to allow for more elaborate taxonomic inferences.

As of the time of writing (January 2022), there are 34 Endolimax‐specific sequences in the NCBI database from sewage, wastewater, stool, pig faeces and fish muscle. The majority of these represent sequences of only ~100–700 bp, and these sequences are not all covering the same part of the SSU rRNA gene. Only 14 of the 34 sequences have more than 1500 bp, but, again, they do not overlap completely in terms of gene coverage. Consequently, the data currently available for phylogenetic inferences are still very limited.

Based on publicly available DNA sequence evidence (Table 17), two species have been identified as hosts of E. nana, namely Homo sapiens and Sus scrofa domesticus [280]. There are also reports of DNA‐based detection of E. nana in non‐human primates [292]. Microscopy studies have revealed high and moderate colonisation rates in macaques [293] and cercopithecids [294], but contrary to the situation for Iodamoeba (see below), no E. nana‐specific sequences from non‐human primates have yet made it into the NCBI database. More molecular studies are needed to delineate the host range for E. nana, but these preliminary data could indicate that the host range for both E. nana and I. bütschlii is quite similar, involving at least human and non‐human primates as well as pigs.

Iodamoeba

Aspects of Biology, Life Cycle, Diagnosis, and Epidemiology of Iodamoeba bütschlii

Iodamoeba is a genus of intestinal parasitic protists found in humans, non‐human primates and other animals. The genus was described by Dobell in 1919 [215], who also gave the name Iodamoeba bütschlii to the human parasite; since then, Iodamoeba found in humans has been assigned to this species. Other species names introduced over the years include Iodamoeba kuenenu, Iodamoeba suis and Iodamoeba williamsi [249]; however, it remains unknown, to which extent one or more of these are in fact Iodamoeba bütschlii.

The parasite has a typical Amoebozoan life cycle involving a cyst stage, which is the resting and infective stage, and a trophozoite stage, which is the feeding stage. Trophozoite stages may be difficult to tell apart from those of Endolimax.

The name ‘Iodamoeba’ reflects the noticeable iodophilic glycogen mass present in Iodamoeba cysts often erroneously referred to as a vacuole. Cysts are irregularly shaped, vary in diameter with a mean of approximately 10 μm [215] and usually have a single vesicular nucleus with a large spherical karyosome and very little peripheral chromatin.

Among the genera of CLIPPs included in this thesis, Iodamoeba is probably among the least common in humans, if not the least common one. Positivity rates reported in surveys of GI parasites in humans sampled across the globe typically range between 0.1% and 2.0%. Of note, however, positivity rates between 7.5% and 15% were reported among 381 apparently healthy subjects from Camiri, Boyuibe, Gutierrez in Bolivia [295] and in surveys from Colombia [296], Venezuela [208] and Peru [297, 298], suggesting a relatively high transmission rate in the peri‐equatorial part of South America. Observing Iodamoeba in a stool sample in the absence of other CLIPPs is a rare phenomenon (unpublished observations).

Genetic Diversity of Iodamoeba bütschlii

It was only in 2012, that we published the first SSU rDNA sequences of Iodamoeba [29], and so Iodamoeba was the last genus of intestinal protozoa known to parasitise on humans to have its ribosomal DNA sequenced. The reason why this had taken so long has to do with the fact that sequencing Iodamoeba‐specific DNA straight off PCR products with no post‐PCR steps such as cloning is of limited use (see Section 1) for this genus—a situation similar to that of Endolimax. The SSU rDNA sequences of Iodamoeba are longer than 2.5 kbp and thereby remarkably longer than those of for instance Entamoeba, Dientamoeba and Blastocystis, which are about 1.8 kbp.

Using DNA extracted from purified cysts confirmed as Iodamoeba by microscopy, low‐specific primers, post‐PCR cloning steps and sequence assembly techniques, we were successful in obtaining a few sequences reflecting Iodamoeba‐specific SSU rRNA genes, six of which were longer than 1000 bp; three sequences were even longer than 2000 bp and thus covered what would be expected to be almost 80% of the gene.

We identified two ribosomal lineages (RL1 and RL2) that—given the remarkable genetic difference of ~30%—easily could be considered separate species—or even genera (Figure 15). The existence of these two lineages has since then been corroborated by independent studies [14, 33, 116], some of which involved metagenomics or metabarcoding. Moreover, there is potential evidence of yet another ribosomal lineage in the article by Hamad et al. [299], who sampled gorillas in Cameroon and deposited Iodamoeba sequences in GenBank (the JX‐ sequences in Table 18) that are all identical apart from two unique bp substitutions in two of the sequences. They appear to be more related to RL1 (94.5% similarity) than to RL2 (90.0% similarity) and may represent a new ribosomal lineage; however, these sequences do not align well with other Iodamoeba sequences, and no data on morphology were provided, so it is not known whether Iodamoeba cysts or trophozoites were in fact present in these samples. Not until longer sequences are provided (the sequences in question were only 697 bp long, representing less than a third of the gene), preferably along with morphology data from microscopy of faecal concentrates, it remains premature to conclude on this. The same situation applies to the seven sequences also referred to as ‘uncultured Iodamoeba’ from cattle and sheep (Table 18), which data, however, were not accompanied by an article to my knowledge.

It should also be mentioned, that in our recent Endolimax study [32], we produced a sequence, ‘SW04’ (OK483224) (Figure 15) from one of the Swedish wastewater samples, which clustered with Iodamoeba RL1 and RL2 with relatively high bootstrap support, sitting on a long branch, which indicates a high degree of genetic divergence. Whether this sequence also represents a novel Iodamoeba lineage is uncertain as of yet; no morphology data were available for this sample either, which could have informed conclusions.

Together with its sister taxon, Endolimax nana (see above), I. bütschlii forms part of the Mastigamoebidae B group [185]. Apart from Endolimax and Iodamoeba, this group holds exclusively species of Mastigamoeba, which are mostly free‐living amoeboflagellates.

Host Specificity of Iodamoeba bütschlii

Apart from humans, I. bütschlii has been reported in non‐human primates [29, 299, 300, 301] and pigs/wild boars [13, 29, 116, 301, 302, 303], for which hosts also cysts specific to Iodamoeba have been observed [301]. Moreover, there are sequences in the NCBI database that have been obtained from samples from sheep and cattle (Table 18), but these data would warrant confirmation. There is limited evidence that other natural hosts could be rodents, camels and birds [29], but to my knowledge, there are no DNA sequences available for Iodamoeba from these hosts yet to confirm the findings.

Maybe not surprisingly, I. bütschlii (both RL1 and RL2) was found in the Swedish wastewater samples [33]. There are moreover two Iodamoeba sequences in the NCBI database isolated from waste water in Australia (MH623069 and MH623073; Table 18) belonging to RL1.

Positivity rates in NHP sampled in zoos range typically between 5% and 8% [304, 305]. However, a staggering 42.96% of 443 cynomolgus macaques bred in China and sampled in Italy tested positive in a relatively recent survey [293].

While both RL1 and RL2 have been found in humans, to date, only RL2 has been found in suid hosts. RL1 has been found in a macaque [29]; meanwhile, the many Iodamoeba sequences produced from samples from gorillas by Hamad et al. [299] may represent a new ribosomal lineage.

Summary of CLIPPs Rarely Reported in Humans

A number of non‐pathogenic Metamonad species have been reported only sporadically in surveys of humans (Table 1). Most of these organisms have life cycles that involve both a trophozoite and a cyst stage in their respective life cycles; meanwhile, Pentatrichomonas and Trichomonas appear to lack the cyst stage; a situation that may be similar to that of Dientamoeba.

The most common of these may be Chilomastix mesnili, which is reported with a frequency of at least up to 8.7% [295, 306, 307] in surveys from across the globe. With some reports for NHPs [293, 294, 308, 309], this species appears to be exclusive to primates. There are currently 12 SSU rDNA sequences in GenBank from two sources ([310] plus Čepička, unpublished [KC960584–KC960590]), which by alignment could appear to reflect at least two distinct subtypes—if not species, differing by ~9%; all sequences are from humans.

For Enteromonas, Retortamonas and Pentatrichomonas, survey data are so limited that it makes little sense to review the data or to speculate on the frequency by which these may show up in stool samples. However, in a recent survey of 127 faecal metagenomes from individuals sampled in Cameroon, Tanzania, Peru, Italy or the United States, Enteromonas‐specific DNA was identified in hunter‐gatherer populations (6.2%–50%) and in Cameroonian fishers (15.8%), but not in individuals from Western countries [311]. The metagenomes included were not queried for Retortamonas‐ and Pentatrichomonas‐specific data, and only a handful of samples displayed evidence of C. mesnili. To date, only three Enteromonas SSU rDNA sequences have been deposited in GenBank all of which were from the study by Kolisko et al. [312], and only one of which reflects Enteromonas hominis (EF551180); the remaining two sequences are from a turtle (EF551179) and a tortoise (EF551178), respectively. A prevalence of 2.4% was reported for Retortamonas intestinalis in the survey from Bolivia by Cancrini et al. [295], and in Libya, the positivity rate across 350 stool samples from children and neonates admitted to hospital was 3.43% [313]. Most of the 18S data in GenBank on Retortamonas stem from two Hendarto et al. performed a study of vertebrate hosts (including humans), identifying a positivity rate of 4/290 (1.4%) in humans and 9/31 (29.0%) in water buffalos [314]; these hosts were the ones for which sampling was most extensive. The other larger set of SSU rDNA sequences were provided by Čepička in 2013. Of the 114 Retortamonas‐specific DNA sequences currently available in GenBank, most are from humans, cattle, water buffalos, pigs, goats, rats and insects. Studies have agreed that Retortamonas from insects cluster separately from Retortamonas from vertebrate hosts. Interestingly, Retortamonads from vertebrates appear to cluster with diplomonads (which include Giardia and Enteromonas), whereas the ones from insects cluster with Chilomastix [314]. That Retortomonadida are polyphyletic have been confirmed by at least two research teams [314, 315].

With regards to Pentatrichomonas, survey data are also scarce, and some of the data published should be interpreted carefully, as reporting would rely on the detection of trophozoites in faecal material either by culture or by permanent staining methods, as this genus probably does not produce cysts, or amplification of specific DNA. Based on PCR, Li et al. identified a positivity rate of Pentatrichomonas hominis of 7.8% across 500 pig samples in China [316], and In Korea, Li et al. identified a positivity rate of 31.4% among 315 pet dogs [317]. A group of monkeys in China was also observed to be commonly colonised based on PCR data [318]. Meanwhile, positivity rates in humans in studies using reliable methods are up to 4.0% in northern China [318].

There are 180 18S sequences specific to P. hominis in GenBank with information on host source at the time of writing, with most sequences being from primates, including humans, and canine hosts. High positivity rates have been identified in, for example, raccoon dogs [319]; however, it should be taken into consideration that surveys using nested PCR for detection might be at risk of overestimating the extent of factual colonisation rates due to issues with increased contamination risks and if specificity is not confirmed by Sanger sequencing. Of the genera mentioned in the present section, Pentatrichomonas is the one that is represented in GenBank with far most DNA sequences, including sequences reflecting coding genes. Analyses of nuclear SSU rRNA genes and flanking regions have not enabled distinction between Pentatrichomonas isolated from human and non‐human hosts [320].

All local general and specific DNA‐based attempts at SSI to screen for Enteromonas and Pentatrichomonas across samples from healthy individuals and individuals with GI symptoms in Denmark have failed to reveal any positive samples (Kaul and Stensvold, unpublished).

Trichomonas tenax is a parasite of the oral cavity of humans and other animals, especially dogs; other hosts include cat and horse [321]. It may co‐colonise the oral cavity with E. gingivalis [251], and it may have a role in the development of periodontal disease [251, 256, 321]. Interestingly, a sequence named ‘Trichomonas tenax’ (JX943581) is 99.73% similar to a sequence named ‘Trichomonas canistomae’ (AY247748), and as other sequences with same coverage (99%) named ‘Trichomonas tenax’ are less similar (e.g., 99.38% for D49495), this could indicate that T. tenax and T. canistomae might be the same species. However, the data in GenBank for these organisms are scarce, so it is premature to do more in‐depth analysis of this.

Conclusively, the organisms briefly and superficially reviewed in this section may not be as rare in humans as could be anticipated judged from available literature, as issues pertaining to lack of a faecal cyst stage (Pentatrichomonas), lack of reference data for developing primers for DNA‐based screening and for mapping of DNA from, for example, metagenomics studies (Enteromonas and Retortamonas) plus potentially large variation in prevalence according to geography and intrageneric diversity may have hampered attempts to develop a fuller picture of the colonisation rates in surveys of intestinal parasites.

CLIPPs as Markers of Gastrointestinal Health?

Based on what was then recent data from our own work and products of international collaboration [80, 90, 93, 94, 133, 322, 323], we published a comment in Journal of Clinical Microbiology asking the following ‘Blastocystis in health and Are we moving from a clinical to a public health perspective?’ [89].

This question embodied the acknowledgement of the fact that Blastocystis was generally more common in healthy individuals than in patients with irritable bowel syndrome, inflammatory bowel disease and acute diarrhoea [90, 92, 323]. We had data suggesting that almost every third adult individual in Denmark is colonised with Blastocystis [90]. Although we had published two case reports describing successful eradication of Blastocystis followed by symptom resolution (one of which involved a rare subtype, ST8) [87, 88], we were unable to identify—based on a literature review—any medical treatment that could consistently lead to eradication of the parasite [83], and our general understanding is that Blastocystis colonisation might not typically be linked to the development of symptoms.

The healthy human gut is characterised by high alpha diversity and a predominance of obligate anaerobes [324, 325]. We hypothesised that Blastocystis might be linked to certain gut microbiota features, a hypothesis that was tested for the first time by our team using metagenomics data generated by the MetaHIT Consortium. We showed that Blastocystis was significantly more common in healthy individuals than in patients with Crohn's disease, that it was linked to the Ruminococcus and Prevotella enterotypes, while inversely linked with the Bacteroides enterotype, and that there was a strong trend towards Blastocystis positivity being associated with low‐to‐normal body mass index [94]. Interestingly, the latter finding was later corroborated by a large metanalysis of 12 sets of metagenomics data by an independent research team [95].

Not only Blastocystis, but also Entamoeba has been linked to a lower relative abundance of Bacteroidetes and generally linked to eubiosis [326, 327], and the fact that the relative abundance of Proteobacteria was the phylum distinguishing Blastocystis carriers from non‐carriers to the largest extent led us to speculate that also Blastocystis might generally be linked to eubiosis [45]. In our study led by Krogsgaard [93], we showed that bacterial alpha diversity of Blastocystis carriers was significantly higher than in non‐carriers, a finding that was later corroborated using samples processed by metabarcoding in our own lab [45], and an observation that we also made for Dientamoeba carriers versus non‐carriers in a different study [328]. Fuelling the hypothesis I had stipulated a few years earlier [329], this experience and these findings led us question the clinical relevance of Blastocystis, but certainly prompted interest in continuing investigating Blastocystis and other CLIPPs as potential markers not only of gut health specifically, but also of overall public health.

The ‘Old Friends hypothesis’ embodies the theory that the development of the immune system relies on input from three sources, collectively referred to as the ‘old friends’: (i) the commensal microbiotas transmitted by the mother and other family members; (ii) organisms from the natural environment that modulate and diversify the commensal microbiotas; and (iii) the ‘old’ infections that could persist in small, isolated hunter‐gatherer groups as relatively harmless subclinical infections or carrier states (colonisation). These categories of organisms have to be tolerated and hence play a role in the development and regulation of the immune system [330]. Comparing diversity patterns of intestinal eukaryotes between individuals with a westernised life style (USA) and individuals with an agrarian life cycle (Malawai), Parfrey et al. published data exemplifying the ‘defaunation’ of the human gut [189]. Interestingly, the authors noticed that individuals with non‐western diets and life styles had microeukaryotic diversity patterns much more similar to those of non‐human mammals compared with those with a western life style and diet. It has been hypothesised that the absence of exposure to parasites that used to commonly colonise and infect humans could result in the development of autoimmune diseases, such as IBD [331]. The hypothesis is indirectly supported by the fact that autoimmune diseases appear to be much more common in the Western world and it may also be a rare phenomenon in non‐human mammals, for example. The question arises by which mechanisms parasites can mature the development of the host immune system. Some parasites may be in lifelong or lengthy direct contact with the immune system (e.g., Toxoplasma gondii and some parasitic nematodes), while others may not (e.g., gut parasites such as CLIPPs); however, the latter may be influencing the immune system indirectly, perhaps by selecting for gut bacteria that do not interfere with the immune system in an inappropriate way.

The gut microbiota of patients with IBD differs from that of healthy individuals [332]. Patients with IBD experience a shift from strictly anaerobic bacteria towards facultative anaerobes such as the Enterobacteriaceae, indicating a role of oxygen in intestinal dysbiosis [333]. Reduced microbial diversity, increased Bacteroidetes and Enterobacteriaceae, and decreased Firmicutes proportions have all been observed in patients with IBD [334, 335], and the gut microbiota profiles of IBD patients to a large degree contrasts with that of Blastocystis‐positive individuals. In our studies, we have generally found a lower prevalence of Blastocystis in patients with IBD [94, 126, 323]. It is likely that a parasite such as Blastocystis do not thrive in microaerophilic environments [96], and we have argued that colonisation by some intestinal parasites can be predicted with quite a high degree of accuracy merely by studying the composition of gut bacteria [96]. This might also indicate that Blastocystis may be merely an indicator organism rather than a gut microbiota manipulator.

Nevertheless, some micro‐eukaryotes are known to exert a beneficial effects on the host. Parfrey et al. provided examples of mutualistic relationships between flagellates residing in the hindgut of termites and cockroaches where they break down cellulose [336]. These insects use parabasalid and oxymonad symbionts to break down cellulose and release energy, and these flagellates can constitute 15%–30% of the body weight of the termite. Looking at larger animals, ciliates are known to colonise the intestinal tracts of a wide range of ruminant and non‐ruminant herbivores. Although not essential for feed degradation and survival of the host, ciliates may contribute to overall gut function by adding degradative complexity, by their ability to scavenge oxygen or by their grazing behaviour, which helps to shape and regulate prokaryotic populations [337]. Mishra et al. recently showed that the camel rumen eukaryotes (mainly ciliates) are highly dynamic and depend on the type of diet given to the animal [338], with two different types of feed selecting for two different types of micro‐eukaryotes (Ciliophora vs. fungi); these observations could fuel investigations into CLIPPs as contributors to host metabolism and gut ecology homeostasis and potentially as markers of dietary intake of the host.

Over the past decade, attempts with faecal microbiota transplantation have been successful in terms of treating recurrent infection with Clostridioides difficile [339]. As high microbiota diversity is an attractive asset of FMT donor stool, it is likely that a number of the donors used for obtaining FMT material are positive for Blastocystis and/or other CLIPPs. Although one study did not identify a difference in clinical improvement between recipients of Blastocystis‐positive FMT material and those receiving Blastocystis‐negative FMT material [340], the impact of the presence of CLIPPs in FMT donor stool should be investigated and it should be explored to which extent CLIPPs can be ‘transplanted’ successfully from donor to recipient with FMT solutions.

Discussion and Future Perspectives

Before I started my interest in parasitology more than 20 years ago, very little was known regarding genetic diversity within Entamoeba species isolated from humans, and there was no evidence of cryptic genetic diversity in E. nana and I. bütschlii; in fact, there were no DNA data on I. bütschlii at all, a parasite colonising many millions of people as well as non‐human primates and pigs, and there was only one DNA sequence of E. nana. Indeed, the whole field of molecular detection and differentiation of intestinal parsaites was in its very infancy. I introduced the idea of testing for D. fragilis in our laboratory, and we were the first ones to produce data on this parasite in Denmark; little did we know that this parasite would turn out to be a more or less obligate finding in children in Denmark. When I defended my PhD thesis on Blastocystis in 2008, only nine subtypes were known; no nuclear or mitochondrial genomic data had been published, and there was limited knowledge on host‐specific differences within Blastocystis subtypes. Finally, we did not know that CLIPPs would typically be more common findings in gut‐healthy individuals than in patients with GI symptoms, and, very importantly for the clinical aspect of CLIPPs, we had no idea that some of these parasites were markers of a healthy bacterial gut microbiota.

Prior to the introduction of DNA‐based methods, it was customary to generate species names based on host range and morphology. We now know that both are unreliable, because host ranges can be broad, and identical morphology can hide substantial genetic differences. For some CLIPP genera, it may be so that we have only scratched the surface of universes of genetic diversity, the implications of which are still to be revealed (e.g., Endolimax and Iodamoeba); for one or two, we may have come a long way already (e.g., Dientamoeba).

A lot of the work included in this thesis (also including articles that were not shortlisted but were listed as supporting articles [p. 10–17]) has been central to informing our attempts to develop useful terminologies applicable by peers to similar data. Standardisation of terminologies is crucial to broaden the global understanding of the genetic diversity and host specificity of parasites in order to be able to delineate parasite transmission patterns and clinical and epidemiological differences. This was clear to us already in 2006, when we developed the subtype terminology for Blastocystis [26]. This terminology was quickly adapted by the community, and research into the distribution an overall epidemiology of Blastocystis subtypes has really taken off since then; however, we have had to keep an eye on the way new subtypes were introduced in order to reduce the risk of confusion [27]. Similar initiatives have been carried out for Entamoeba [187]. Given the complexity of the genetic diversity of some of these parasites, it is important that terminology is clear, robust and practical and expert groups should revisit guidelines and standards on a continuous basis, the way it is currently done in the Blastocystis COST action (https://www.cost.eu/actions/CA21105/).

PCR coupled with Sanger sequencing for molecular characterisation and time‐consuming and expensive primer walking for the mapping of Blastocystis mitochondrion‐like genomes have been instrumental to the production of a lot of the data that have gone into this work. However, with more modern technologies, such as metabarcoding and Nanopore sequencing, data will be produced much faster and in greater quantity; longer reads can be produced with Nanopore sequencing and we will have an opportunity to obtain a more comprehensive view of pro‐ and eukaryotic diversity when metabarcoding is used. We also have the opportunity to study links between parasites and other organisms, as exemplified in Section 4. In this process, however, one should take care to try and remember the relevance of including data on morphology where possible for reasons that should be clear from previous sections.

In terms of any use of metagenomics, it should also be obvious from the work included in this thesis that efficient use of metagenomics data relies on complete reference sequence databases. For many of the CLIPPs included in the present thesis only DNA sequences reflecting the SSU rRNA genes are available, and it might appear that not even these have yet been characterised fully. Data from WGS are still pending for most CLIPPs. At the time of writing, full genome data are available for five species of Entamoeba (E. histolytica, E. dispar, E. moshkovskii, E. nuttalli and E. invadens), and for some subtypes of Blastocystis.

Even for our metabarcoding assay, there might be issues, despite the fact that this assay is targeting 18S genes only, which is probably the most commonly characterised gene across all living organism. The issue pertains to the fact that for each DNA sample processed by the assay, a varying number of sequence reads cannot be mapped to a reference taxon and therefore remain unannotated. Although these sequences can easily be extracted for each sample, the process of querying the sequences for new genetic variants is time consuming and of limited feasibility, with a rather limited the risk‐reward trade‐off, probably. Finally, this particular assay has varying degrees of sensitivity depending the parasite in question [13, 33], so in its current version it cannot be used as a one‐fits‐all (but maybe a one‐fits‐most) assay.

There is evidence that we only started to scratch the surface of eukaryotic diversity in complex matters such as faecal samples. Chouari et al. [341] used 18S sequencing to investigate eukaryotic diversity in wastewater, and of 1519 analysed sequences, 160 operational taxonomic units (OTU) were identified. Altogether 56.9% of the phylotypes were assigned to novel phylogenetic molecular species, exhibiting < 97% sequence similarity with their nearest affiliated representative within public databases. Similarly, Matsunaga et al. [342] observed that 60% of their 18S rRNA gene clones obtained from DNA extracted from municipal wastewater had < 97% sequence identity to described eukaryotes. In both studies, data on Blastocystis and Amoebozoa were observed. These studies confirmed not only the vast DNA data gap in the eukaryotic tree of life, but also the relevance of using sewage as study material for investigations into eukaryotic diversity involving CLIPPs.

One of the most remarkable findings of our studies is the demonstration of vast cryptic genetic diversity in some of the species. For comparison, E. dispar only differs from E. histolytica by < 2%, and from E. bangladeshi and E. ecuadoriensis by 5%–6% across SSU rRNA genes. Meanwhile, the genetic distance between E. coli subtypes of 12%–13% equals the distance between quadrinucleated E. dispar and uninucleated E. bovis, two species that clearly differ in both host spectrum and cyst morphology. Things become even more conspicuous when looking at the ribosomal lineages of I. bütschlii: The extent of genetic diversity across the two currently acknowledged lineages differs by as much as a good 30%. This figure might be difficult for us to grasp, when we remind ourselves that human and ovine or suid nuclear SSU rRNA genes differ only by a handful bp or two. The implications of this are yet not clear. It should be investigated whether this amount of genetic diversity is reflected in the remaining genome, in which case gene prediction analyses might provide us with an opportunity to study phenotypic differences among the lineages and whether there would be reason to hypothesise that the lineages differ in the impact on host ecology and host health overall. Single‐cell genomics/transcriptomics may prove a way forward in this respect, potentially combined with the cyst isolation procedure that we used in the study of Iodamoeba [29], or on parasites in culture.

For some protozoa, at least two ‘sets’ of ribosomal genes exist, an asexual set and a sexual set; this was exemplified recently in our study on a case of Plasmodium cynomologi in a Danish tourist [343], where our metabarcoding assay picked up both types of SSU rDNA sequences. These two sets of sequences differ substantially. However, there is currently no evidence that something similar should exist within the Amoebozoa. Also, the evidence of links between lineages and geography would not support the hypothesis of two sets of ribosomal RNA genes in some of the Amoebozoa. Nevertheless, SSU rRNA genes are organised differently among the CLIPPs: In Entamoeba, rRNA genes are located exclusively on extrachromosomal plasmids (circular DNA) [344], whereas they are organised on different chromosomes in for instance Blastocystis (linear DNA) [54, 345]. This has implications for our interpretation of rRNA data and for what to expect in terms of intra‐strain diversity.

Our research has shown that Entamoeba and Blastocystis are cosmopolitan parasites, possibly reaching every ‘corner’ of the world, including remote and frigid areas such as Greenland (unpublished observations). It is likely that the study of these parasites can further inform studies of the evolution and migration of host species (because of co‐evolution). Indeed, it is also interesting that humans and several other larger species of mammals share some if not all of the CLIPP genera dealt with in this work and are indeed very common—in some instances almost obligatory—hosts; however, one major group of animals appears to be only accidental hosts of most of these, namely the carnivores. For instance, surveys of intestinal parasites in wild and synanthropic carnivores have revealed positivity rates going towards zero [106, 123, 346, 347, 348, 349]. This may have to do with the theory that most Amoebozoa and Blastocystis are parasites lodged in the colon and maybe particularly in the caecum [249]. The latter anatomical structure is typically more or less lacking in animals that are predominantly carnivorous such as felines and canids. To this end, diet may play a major role here, as herbi‐ and omnivores would eat relatively much fibre compared with carnivores, and the metabolism of fibre, which predominantly takes place in the caecum and adjacent parts of the intestine, results in the generation of short‐chain fatty acids (SCFA) that are important for reducing oxygen levels and maintaining eubiosis in the gut [96]. SCFAs, which include butyrate and propionate, have several functions essential to colonic health and immune function and are known to regulate cells of both the innate and adaptive immune systems. With respect to intestinal homoeostasis, significant reductions in the abundance of bacteria involved in butyrate and propionate metabolism have been identified as markers of dysbiosis in ulcerative colitis [115]. It should also be explored to which extent CLIPPs would rely on SCFAs as a source of energy.

Another observation suggesting that carnivores are not natural hosts of these CLIPPs is that those few animals that do test positive for Blastocystis, for instance, appear not to be colonised by one or few select subtypes the way we usually see it for natural hosts (e.g., suids being positive for typically ST1 and ST5, and bovids being positive for typically ST10 and ST14). The situation is similar for, for example, lemurs; contrary to NHPs, these are not common hosts of Blastocystis, but whenever Blastocystis would be found, no particular subtype would appear to predominate [51, 350]. This is an example of how molecular characterisation of parasites can assist in identifying natural and accidental hosts.

Contrary to cattle, pigs share a number of CLIPPs with primates. Endolimax nana, E. hartmanni, E. polecki and I. bütschlii are all Amoebozoa shared by both pigs and primates. For Blastocystis, Subtypes 1–3 can be seen in both pigs and primates, and ST5, which is particularly common in pigs, appears common in apes, although not in human primates. Finally, Dientamoeba, a common coloniser of humans, has been observed in pigs by some research teams [172, 351]. The large overlap in micro‐eukaryotic fauna may reflect the fact that pigs and humans are omnivorous while cattle are herbivorous and/or that pigs are genetically more related to primates than cattle (pigs and human share 98% of the DNA, while cattle and humans share about 80%), or at least that the gut microbiota of primates might be more similar to that of pigs than of any other non‐primate host. A somewhat distant but still reminiscent scenario is seen for Taenia, where humans can serve as intermediate host of Taenia solium but not of Taenia saginata.

Taking the faecal‐oral transmission for CLIPPs into account, it is striking that some of these—Blastocystis and *Dientamoeba—*are very common in a country such as Denmark, where hygiene practices are relatively high. If indeed pinworm is a suitable vector for Dientamoeba, then the relatively high occurrence of pinworm infection in Denmark (data not shown) might explain the common occurrence of Dientamoeba carriage in this country. For Blastocystis, the explanation may be less straightforward. However, given the fact that this organism belongs to a separate phylum—even kingdom—it might be of little use to assume that Blastocystis is limited to entirely the same way of transmission as cyst‐forming protozoa. There may be characteristics of Blastocystis that enable it to exist and persist in the environment to an extent that we are not aware of and that enables the organism to be transmitted in a way that differs from what is seen for members of the Archamoebae. By all means, the colonisation pressure of Blastocystis must be enormous.

When evaluating observations from DNA‐based surveys of species/ribosomal lineages of CLIPPs there are a couple of things to bear in mind. Ideally, full‐length SSU rDNA sequences should be obtained, but to date, most researchers have used ‘partial sequences’; that is, sequences reflecting only part of the SSU rRNA gene, typically sequence fragments that could be covered by conventional bidirectional Sanger sequencing (up to about 800 bp). Within the field, there have only been few attempts towards advocating for a standardisation of primers used for amplification and sequencing, and so there are numerous examples of differences in the SSU rRNA gene regions covered. For example, Calegar et al. [218] used the Entamoeba primers developed by Verweij et al. [275], which cover a region different to that explored by our group, namely the 5′‐end of the SSU rRNA gene. Meanwhile, in our recent study on the diversity of E. coli and E. hartmanni, we used low‐specificity primers to amplify the middle section of the E. hartmanni SSU rRNA gene [13]. This part of the gene apparently holds more genetic information and enable better resolution than the 5′‐end of the gene. It should also be investigated whether the ‘Entam’ primers developed more than 20 years ago (i.e., at a time when the NCBI database held a very limited number of Entamoeba‐specific DNA sequences) for genus‐specific amplification are oligos targeting DNA sequences that are conserved among all species of Entamoeba; at least for the reverse primer (‘Entam2’), there is a single mismatch compared with E. suis (DQ286372) and E. gingivalis DNA sequences (KX027297, D28490). At least one mismatch to most if not all Entamoebas is also seen in the ‘TN14’ reverse primer developed by Matey et al. [213], which was also recently used in the study by Mulinge et al. [352]. The genus‐specific primers published by our group in 2011 (‘ENTAGEN_F’ and ‘ENTAGEN_R’ [49]) still appear to be useful as we have not been able to identify sequence variation in Entamoeba in the primer annealing regions. Also for Blastocystis, various regions of the SSU rRNA gene have been studied (e.g., ‘Scicluna region’ [46], ‘Santin region’ [353] and ‘Stensvold region’ [66]).

Nested PCR approaches have been used numerous times to detect and differentiate CLIPPs. Sometimes, the inner PCR has been designed as a species‐specific PCR, and the advantage of that would be that the work and expense of sequencing could be obviated, as the species diagnosis would be carried out based on the size of the PCR product alone or maybe by RFLP. However, this methodology is also associated with drawbacks. The primers will only amplify what is already known, and any new variants of the species in question might therefore not be detected, in case this variation exists in the primer annealing region. Secondly, if PCR products are not sequenced, false‐positivity is a possibility, depending on the quality of assay validation. Thirdly, in a study yet unpublished, DNA sequences of E. hartmanni were deposited in 2022 in the NCBI database (OP688358‐OP688362), with the domestic dog listed as the host. In this very instance, sequence data are in fact available. Nevertheless, it should also be noted that if this data is a result of nested PCR, the very significance of the findings should be interpreted with caution, given the fact that dogs are coprophagous. Parasites may be able to pass through the digestive tract of hosts that are not natural hosts and still be detected by highly sensitive DNA‐based methods such as nested PCR in faeces even if present in only very small numbers. PCR does not distinguish between DNA from live and dead organisms. The best way to clearly demonstrate dogs as ‘new’ hosts of E. hartmanni would therefore be to establish the parasite in culture of a dog faecal sample (which would strongly suggest a live isolate that had in fact been colonising the dog's intestine), or at least produce a sequence from only a single‐round PCR, which would typically be a less sensitive method than nested PCR.

While at the time of writing, it still appears likely that CLIPPs per se do not inflict disease on humans to any major extent, the roles of CLIPPs as ‘Trojan horses’ of for instance viruses should be explored further. It has been known for long that parasites can host bacteria (e.g., Acanthamoeba can host Legionella [354]), and moreover, there is recent evidence suggesting many CLIPP species as hosts for pecoviruses, hudisaviruses, Kirkoviridae and Redondoviridae, among others [252, 355]. Wider use of metagenomics is expected to enable the disclosure of such relationships.

DNA from intestinal parasites that are common and that exhibit high degrees of host specificity can be used as strong indicative evidence of host in cases where faecal material is available for analysis by, for example, metabarcoding, and where DNA data are insufficient for host identification (humans and some other larger mammals differ very little across the 18S gene).

In clinical microbiology laboratories the reporting of CLIPPs has been ‘good practice’, also in settings where the general consensus has been not to treat. The reporting of CLIPPs has been relevant especially to raise awareness of faecal‐oral exposure, which could prompt further investigations for intestinal pathogens. The problem here is that colonisation by CLIPPs can be lengthy [80], and, unless there is a recent reference sample that was negative for CLIPPs, the reporting of CLIPPs in stool may have limited value. Paraclinical findings, such as Charcot–Leyden crystals (break down products of eosinophils and basophils) and/or blood cells in stool, which may be revealed by microscopy of faecal concentrates, or better, by direct microscopy or permanent staining of fixed faeces, might be more relevant information; however, the practice of reporting these findings is probably declining.

Summary of outstanding questions and Update reference DNA sequence databases with genomic data (near‐complete 18S sequences or even entire ribosomal operons, genomes of organelles where possible and nuclear genomes).Try to interpret what the extreme genetic diversity seen in Iodamoeba and Endolimax tells us and investigate whether the extensive genetic diversity observed across their SSU rRNA genes is reflected elsewhere in their genomes. What impact does their level of intrageneric diversity have on exisiting species concept(s)?Expand on the knowledge of the genetic diversity and the host spectrum of CLIPPs by sampling more and different hosts.Expand on the knowledge of the genetic diversity and the host spectrum of CLIPPs by sampling more and different hosts.Obtain ‘helicopter views’ of microeukaryotic diversity in human and non‐human faecal samples using metagenomics and OTU annotation together with bioinformatics tools to identify hitherto unexplored diversity. Findings could be coupled with metadata and used to identify OTU communities linked to demographic features, life styles (including diets), hosts, diseases, etc. OTUs of specific interest could be characterised by rDNA full‐operon analysis using, for example, Nanopore sequencing.Explore the metabolism of CLIPPs by genomic in vitro predictions or by wet lab experiments, including the ability of CLIPPs to ‘predate’ on the host–gut microbiome, including investigations into the enzymes and metabolites released by CLIPPs.Explore the role of CLIPP colonisation on host microbiomes. Animal or advanced in vitro models that can mimic in vivo models are warranted. It should also be investigated for how long one can be colonised with species of Amoebozoa.Explore the role of CLIPPs as hosts and transmitters of bacteria and viruses.

Conflicts of Interest

The author declares no conflicts of interest.