Blastocystis: Genetic diversity and molecular methods for diagnosis and epidemiology.

Blastocystis, an unusual anaerobic, single-celled stramenopile, is a remarkably successful intestinal parasite of a vast array of host species including humans. Fecal Deoxyribonucleic acid (DNA) analysis by nucleic-acid based methods in particular has led to significant advances in Blastocystis diagnostics and research over the past few years enabling accurate identification of carriers and molecular characterization by high discriminatory power. Moreover, Blastocystis comprises a multitude of subtypes (STs) (arguably species) many of which have been identified only recently and molecular epidemiological studies have revealed a significant difference in the distribution of STs across host species and geographical regions. Having a cosmopolitan distribution, the parasite is a common laboratory finding in the stools of individuals with and without intestinal symptoms across the entire globe and while the parasite remains extremely difficult to eradicate and isolate in culture, appropriate molecular tools are now available to resolve important questions such as whether the clinical outcome of colonization is linked to ST and whether Blastocystis is transmitted zoonotically. This review summarizes some of the recent advances in the molecular diagnosis of Blastocystis and gives an introduction to Blastocystis STs, including a recommendation of subtyping methodology based on recent data and method comparisons. A few suggestions for future directions and research areas are given in the light of relevant technological advances and the availability of mitochondrial and nuclear genomes.

INTRODUCTION

Over the past 10 years, screening of patients for intestinal parasites, especially protozoa, by molecular methods has started to gain a foothold in clinical microbiology laboratories. Although a variety of theoretical and practical limitations and obstacles can be identified,[1] these are counter-balanced by the many advantages of nucleic-acid based diagnostics; advantages that include higher diagnostic sensitivity, high accuracy (repeatability, reproducibility), and semi-or-fully automated processes, leading to higher cost-effectiveness and standardization.[23]

Conserved and variable regions within the 18S small subunit (SSU) ribosomal DNA (rDNA) constitute the basis for identifying phylogenetic relationships between species[4] and the SSU ribosomal RNA (rRNA) genes have been popular targets for diagnostic polymerase chain reaction (PCR) assays; multiple gene copy numbers ensure high diagnostic sensitivity, and SSU rDNA sequences are available for most known parasites.

Amplification of parasite-specific rDNA from genomic DNA extracted from feces (i.e., without any intermediate step such as culture or isolation of (oo) cysts) has not only opened new diagnostic avenues in the clinical microbiology laboratory but also enabled us to detect a multitude of novel ribosomal lineages, many of which are probably separate species, either in the absence of morphology data and cultures,[5678] or in situations where morphological differentiation is not possible.[910]

Molecular screening assays typically include those enabling immediate detection of and differentiation between Entamoeba histolytica and Entamoeba dispar, and those that target the most common and clinically relevant parasitic genera such as Giardia and Cryptosporidium.[21112] Species identification and genotyping of the latter genera have been used to clarify transmission patterns and to assist in outbreak investigations.

Much more common intestinal parasites include Blastocystis and Dientamoeba, the clinical significance of which remains to be fully understood, in part due to the absence of efficient treatment modalities.[13141516] It is clear that detection of these parasites by conventional methods is difficult for a variety of reasons, such as lack of a detectable cyst stage (Dientamoeba), fragility of vegetative stages, and general morphological inconspicuousness.[2171819]

The prevalence, intra-generic diversity and transmission modes of these two genera are only starting to be revealed following the introduction of molecular methods. For Dientamoeba, the design of conventional and real-time diagnostic PCR assays has proven relatively easy due to the fact that only one species has been found in humans (Dientamoeba fragilis), and there is relatively little genetic variation across the SSU rRNA gene.

GENETIC DIVERSITY OF BLASTOCYSTIS AND BLASTOCYSTIS SUBTYPING

The situation is quite different for Blastocystis, a remarkable, single-celled enteric parasitic stramenopile of humans and a vast array of non-human hosts. Based on SSU rDNA analysis, this genus comprises at least 17 lineages (arguably species), the so-called subtypes (STs).[20] ST1-ST4 account for probably more than 90% of human carriage,[2122] while the remaining human carriers are colonized by isolates belonging to ST5-ST9. All of these STs, apart from ST9 have also been found in various non-human hosts. Since 2009, no less than 8 novel STs (ST10-ST17) have been detected in non-human hosts,[20] primarily by amplification of Blastocystis DNA directly from fecal DNA template.[578]

Importantly, the ST system developed in 2007 includes Blastocystis exclusively from humans, other mammals and birds.[23] Hence Blastocystis from for instance reptiles, amphibia, and insects have not been included. Incidentally, ST5 has been found in an amphibian host, and conversely, a few isolates from mammalian samples representing two of the “novel” STs, namely ST15 and ST17, are phylogenetically closer related to reptilian or arthropod isolates than to STs usually found in mammalian hosts.[20]

Genetic variation within some STs (e.g., ST3) may amount to about 3% and experience has told us that the level of genetic difference between STs is at least 4-5%.[24] In the event that a Blastocystis barcode sequence does not match those in the 18S database and has less than 97-98% similarity to sequences available at the NCBI database, it is suggested that the complete SSU rRNA gene be sequenced using a mixture of general and specific primers (and preferentially from DNA from a cultured isolate since this makes amplification easier and sequence traces often become clearer). The complete sequence should be aligned with reference sequences and the alignment should be edited and submitted to phylogenetic analyses using a variety of models, including at least distance and Maximum Likelihood analysis.[520] Since a standard approach to Blastocystis subtyping and nomenclature has been proposed,[2325] it is appropriate to approach a standard methodology here also; for more detailed considerations and guidelines regarding ST analysis and nomenclature, please refer to Clark et al.[2024]

Subtyping has given us a crucial tool for testing the current working hypothesis that the clinical outcome of Blastocystis colonization is dependent on ST.[262728] Based on observations from multiple studies from various countries, no clear link between STs and clinical phenotypes can be made at present. However, at least two main confounders may play a role: Firstly, significant variation in the distribution of STs is seen across geographic ranges. A review of more than 3,000 observations from global surveys made it clear that while ST4 is almost as common as ST1 and ST3 in some countries in Europe, this ST is virtually absent in most other regions.[21] Conversely, ST6 is very rare in Europeans and Americans while human colonization by this ST appears rather common in some Asian and Middle Eastern countries, such as Egypt, Nepal, Japan and Malaysia. Although ST3 is the most common ST reported to date, a few individual surveys have seen a strong preponderance of ST1 or a “tie” between ST1 and ST3.[2930313233] Secondly, different methodologies are used for subtyping, and differences in results from ST surveys may stem from differences in subtyping methodologies.[25] Subtyping has been approached in mainly two ways: By screening fecal samples by PCR using either (1) ST-specific primers (sequence-tagged site (STS) primers)[34] or (2) genus-specific primers with subsequent sequencing for ST identification; one such method is barcoding covering the 600 5’-most bases of the SSU rDNA.[35] These two methods, STS and barcoding have different advantages and limitations, which will be discussed in the following section.

Barcoding appears robust for genetic characterization combining the use of a forward primer of broad, eukaryotic specificity (RD5[36]), and a genus-specific reverse primer (BhRDr[35]). Sequence traces from PCR products obtained by barcoding may be difficult to decipher in cases of mixed ST carriage. The foremost advantage is that barcode fasta files can be queried individually or by bulk submission at the “Blastocystis ST (18S) and Multi Locus Sequence Typing (MLST) Multi Locus Sequence Typing Databases”[3738] at www.pubmlst.org/blastocystis and identified not only to ST level, but to 18S allele level; 18S allele analysis offers higher resolution than subtyping alone.[38] This is especially attractive for new colleagues in the field since this approach bypasses the need for phylogenetic analyses based on reference sequences and hence the risk of annotating sequences to STs to which they do not belong, which has been a problem previously.[2439] It should be noted that barcoding PCR should not be used on fecal DNA template as a strictly diagnostic tool since the RD5/BhRDr primer pair typically amplifies common fungal DNA in the absence of Blastocystis with no obvious difference in PCR product size.[24]

By using the STS method, the need for sequencing PCR products is circumvented, and detection of mixed ST carriage is straight-forward. However, although highly specific, some of the primer pairs appear rather insensitive, especially those targeting ST4, for which the common 18S allele (allele 42) appears to go undetected.[2538] Moreover, STS primer pairs are only available for ST1-ST7; hence, human carriage due to other STs remains undetected. Moreover, STS PCR products are rather long (~300-700 bp), which is a clear disadvantage since these primers are sometimes used diagnostically. Finally, there is evidence that sequence variation in the STS primer regions exists.[25]

A list of studies employing the two different methods is available in Table 1. It appears that ST4 accounts for 1% and 17% of Blastocystis carriage detected by the STS method and non-STS methods, respectively. Although potentially coincidental, it should be noted that ST4 is absent mainly in those countries where surveys have been carried out using the STS method, typically Asian and Middle Eastern countries. Therefore, the use of barcoding in these countries is especially warranted to confirm/refute the absence of ST4 in these particular regions.

Table 1

Subtype distribution by country of sample origin (modified after[21])

Although numerous authors have avidly pointed toward a large overlap in STs shared by humans and animals,[7337576777879] the hypothesis of zoonotic transmission of especially Blastocystis ST1 and ST3 has to a large extent been unsubstantiated by recent studies looking at intra-ST genetic diversity, such as MLST[38] and 18S allele analysis.[203880] However, it is still highly likely that some STs are passed on zoonotically; for instance there are very few reports of human ST8 infections and only two 18S alleles have been identified so far. ST8 is probably mainly a ST of non-human primates (NHP), and human carriage may very well be a result of exposure to NHP manure.[35388081]

Intra-ST diversity appears to vary dramatically from ST to ST. ST4 SSU rDNA sequences from humans are strikingly homogenous compared to those stemming from ST3.[38] A total of 132 ST3 and ST4 samples from human and non-human primates were studied by multilocus sequence typing, and while 58 sequence types were found among the 81 ST3 samples, only five sequence types were found among 50 ST4s. Intra-ST diversity of ST1 and ST2 are currently being undertaken by MLST analysis.

Molecular screening tools for human Blastocystis carriage

The extensive genetic diversity of Blastocystis was clearly and independently substantiated by Clark and Böhm-Gloning et al., in 1997.[2647] Pairwise genetic distances among some of the STs amount to at least 14.8%,[5] and this situation has made the design of genus-specific primers applicable to fecal DNA template amplification much more challenging than for most other parasites encountered in the clinical microbiology laboratory.

The first diagnostic PCR for detection of Blastocystis was introduced in 2006,[42] i.e. before the introduction of the consensus terminology. This conventional PCR was based on sequences present in GenBank at that time and targets ~ 300 bp toward the 3′-end of the SSU rRNA gene, but later observations suggest that it may exhibit preferential amplification of ST3 over others due to ST-specific polymorphisms in both primer annealing regions (unpublished observations).

Only two “pan-Blastocystis” real-time PCR assays have been published so far: A SYBR-green-based real-time PCR assay allowing subtyping by direct sequencing of SSU rRNA gene PCR products was introduced in 2011.[46] Despite the fact that the assay was designed to be applicable to fecal DNA, no inhibition control was incorporated. Moreover, in silico amplification shows that PCR products as large as 345 may be amplified from Blastocystis-specific template in human fecal samples (unpublished observations); this is a relatively large product for a real-time PCR assay, potentially impairing test sensitivity. The assay had a specificity of only 95%, which despite the high prevalence of Blastocystis gives a positive predictive value only around 91%, given a prevalence of 35%.

It may initially seem appropriate to evaluate the PCR assay specificity against other intestinal parasites such as Entamoeba, Giardia, Cryptosporidium, and helminths;[46] however, Blastocystis is genetically not even remotely related to any of these organisms. Neither are fungi, but specificity testing of diagnostic Blastocystis PCR assays ought include species of fungal genera such as Candida, Geotrichum and Saccharomyces,[82] all of which are common components of human fecal samples, and potentially quite abundant in samples that have been on their way to the laboratory for a couple of days at ambient temperature.

The other real-time PCR method was published by Stensvold et al.,[82] and was based on TaqMan technology including a specific probe and an internal process control to identify cases of PCR inhibition. The assay was evaluated against a large panel of fungal and bacterial DNAs and proved 100% specific. ST data were generated for positive samples from the test panel, and interestingly, samples positive for ST3 had lower cycle threshold (Ct) values than samples positive for ST1, indicating that ST1 colonization is generally lighter than ST3 colonization (P = 0.022). So far, this method has not been used for screening/surveys, but among the potentially interesting results using this method are whether Ct value ranges are associated with ST and whether the subjective experience of symptoms is related to infection intensity as measured by Ct-values.

It was previously shown that conventional PCR was not significantly more sensitive than short-term xenic in-vitro culture (XIVC);[18] however, XIVC had a sensitivity of only 79% and 52%, respectively, compared to the real-time PCRs developed.[4682] An estimate of the number of copies of rDNA in one Blastocystis cell is not available, although, it may lie somewhere between 10 and 100 based on currently available data.[8283]

It is already clear that the prevalence of Blastocystis even in Western countries is much higher than expected only few years ago, when prevalence figures were generated using traditional parasitological methods. Accurate identification of carriers and non-carriers by real-time PCR is key to understanding fundamental questions about transmission, treatment efficacy, longevity of colonization, and thereby overall clinical significance.

Mixed subtypes infection

It was recently argued that sequencing-based subtyping methods likely underestimate mixed ST infections.[84] However, a review of the data collected from various surveys across the globe (especially those using the STS primers that enable direct detection of mixed ST infections) strongly indicates that the level of mixed ST infections is rather low compared to the prevalence of the parasite, probably <10% of all cases. Roughly, the extent of mixed ST infection detected by the STS method corresponds to that seen by PCR and sequencing (e.g., barcoding).[34]

The impact of sampling and sample processing on Blastocystis detection and subtyping

Methodological issues other than those pertaining to the choice of subtyping method may influence our ability to accurately identify the distribution of Blastocystis STs in any given population. For instance, sometimes Blastocystis is subtyped directly from fecal DNA template, at other times from DNA template prepared from cultured isolates; there is evidence that XIVC may favor one ST over another in cases of mixed ST colonization,[75] and it is also possible that different strains (e.g. some animal vs. human strains) may have different requirements in terms of growth conditions, including temperature and nutrients; however, in some labs, short-term XVIC appears not to lead to differential growth of STs.[20]

Storage conditions before sample processing may be another variable potentially influencing our ability to detect Blastocystis. Samples are often collected in locations far from the lab and may be kept under varying conditions such as on cold/freeze storage, at ambient temperature, or in preservatives such as lysis buffer or ethanol.[2252]

DNA extraction can be performed in many ways and differences in quality may obviously impact our ability to amplify and sequence Blastocystis; here, the size of the PCR product is also important since short PCR products may be amplifiable even from low quality DNA template while longer reads (eg. 600 bp used for barcoding) may be more difficult to obtain (unpublished observations). Furthermore, the amount of competing DNA template in DNA extracted directly from feces may influence PCR results.

Although clearly advantageous, it may prove much more difficult to standardize sampling and sample processing than standardizing subtyping methods. Suffice to say that direct screening of fecal DNA templates by PCR (preferably real-time PCR or a conventional PCR amplifying a “short” PCR product) with subsequent barcoding of positive samples may be considered “state-of-the-art” due to high diagnostic sensitivity and the added benefit of automatically assigning sequences to a ST number and even 18S allele when using www.pubmlst.org/Blastocystis, and this should be feasible in most laboratories.

New technologies and perspectives

There is no doubt that the future use of relevant technologies will facilitate and improve epidemiological surveys plus molecular analyses and detection of Blastocystis. So far, TaqMan array technology has been used for instance to generate a “one-size-fits-all” array for the detection of viral, bacterial, and parasitic enteropathogens.[85] It is not unlikely that this technology could prove useful in terms of facilitating the simultaneous detection and differentiation of Blastocystis STs. Another alternative that could potentially be looked into is the design and applicability of a real-time PCR assay using molecular beacons or scorpion probes for ST/allele analysis by single nucleotide polymorphism (SNP) detection; use of molecular beacons has proven useful for instance in screening for hotspot resistance mutations in Aspergillus fumigatus.[86] ST-specific SNP analysis by pyro sequencing has been developed, but never applied to large-scale studies.[87]

In this era of “omics,” it is possible that Blastocystis data can be extracted from pools of data already available and obtained for instance by metagenomic approaches. For instance, we are currently identifying and analyzing Blastocystis-specific mitochondrion-like organelle sequences (ST1-ST4) retrieved from metagenomic species originally generated in a study of the intestinal microbiomes of 124 Danish and Spanish individuals conducted by the Meta HIT Consortium.[88] After comparing the respective prevalence figures of STs identified by this method with data obtained, e.g. by barcoding in Denmark and Spain, this approach appears valid and useful for extracting ST data on Blastocystis. Moreover, it enables a characterization of the bacterial flora that accompanies Blastocystis colonization (unpublished observations).

While efforts to generate data on the extent (prevalence) and nature (ST) of Blastocystis carriage in various cohorts (mainly patients with intestinal disease versus healthy individuals) go on, it is also important to identify the methods relevant for investigating the function of Blastocystis genes.[89] The genome of Blastocystis ST7 was obtained by traditional Sanger sequencing and published in 2011,[83] and more genomes from other, more common, STs may be published as early as 2013-2014, thanks to faster and less expensive new generation sequencing methods. Comparative genomic and transcriptomic analysis should assist us in identifying effector proteins of potential interest that can be detected, monitored and correlated with health and disease phenotypes. Finally, analyzing Blastocystis and other common micro-eukaryotes in an ecological context is highly warranted for a complete understanding of their function, life cycle, and clinical significance.[8990]