Investigation of neglected protists Blastocystis sp. and Dientamoeba fragilis in immunocompetent and immunodeficient diarrheal patients using both conventional and molecular methods

Introduction The clinical significance of Blastocystis sp. and Dientamoeba fragilis in patients with gastrointestinal symptoms is a controversial issue. Since the pathogenicity of these protists has not been fully elucidated, testing for these organisms is not routinely pursued by most laboratories and clinicians. Thus, the prevalence of these organisms and the subtypes of Blastocystis sp. in human patients in Turkey are not well characterized. This study aimed to determine the prevalence of Blastocystis sp. and D. fragilis in the diarrheic stool samples of immunodeficient and immunocompetent patients using conventional and molecular methods and to identify Blastocystis sp. subtypes using next generation sequencing. Material and methods Individual stool specimens were collected from 245 immunodeficient and 193 immunocompetent diarrheic patients between March 2017 and December 2019 at the Gazi University Training and Research Hospital in Ankara, Turkey. Samples were screened for Blastocystis sp. and D. fragilis by conventional and molecular methods. Molecular detection of both protists was achieved by separate qPCRs targeting a partial fragment of the SSU rRNA gene. Next generation sequencing was used to identify Blastocystis sp. subtypes. Results The prevalence of Blastocystis sp. and D. fragilis was 16.7% and 11.9%, respectively as measured by qPCR. The prevalence of Blastocystis sp. and D. fragilis was lower in immunodeficient patients (12.7% and 10.6%, respectively) compared to immunocompetent patients (21.8% and 13.5%, respectively). Five Blastocystis sp. subtypes were identified and the following subtype distribution was observed: ST3 54.4% (n = 37), ST2 16.2% (n = 11), ST1 4.4% (n = 3), ST6 2.9% (n = 2), ST4 1.5% (n = 1), ST2/ST3 11.8% (n = 8) and ST1/ST3 8.8% (n = 6). There was no statistically significant difference in the distribution of Blastocystis sp. subtypes between immunocompetent and immunodeficient patients. Conclusion and recommendation Our findings demonstrated that Blastocystis sp. and D. fragilis are commonly present in immunocompetent and immunodeficient patients with diarrhea. This study is the first to use next generation sequencing to address the presence of Blastocystis sp. mixed subtypes and intra-subtype variability in clinical samples in Turkey.


Introduction
The clinical significance of Blastocystis sp. and Dientamoeba fragilis in patients with gastrointestinal symptoms is a controversial issue. Since the pathogenicity of these protists has not been fully elucidated, testing for these organisms is not routinely pursued by most laboratories and clinicians. Thus, the prevalence of these organisms and the subtypes of Blastocystis sp. in human patients in Turkey are not well characterized. This study aimed to determine the prevalence of Blastocystis sp. and D. fragilis in the diarrheic stool samples of immunodeficient and immunocompetent patients using conventional and molecular methods and to identify Blastocystis sp. subtypes using next generation sequencing.

Material and methods
Individual stool specimens were collected from 245 immunodeficient and 193 immunocompetent diarrheic patients between March 2017 and December 2019 at the Gazi University Training and Research Hospital in Ankara, Turkey. Samples were screened for Blastocystis sp. and D. fragilis by conventional and molecular methods. Molecular detection of both protists was achieved by separate qPCRs targeting a partial fragment of the SSU rRNA gene. Next generation sequencing was used to identify Blastocystis sp. subtypes.
The most common parasitological examination methods used in clinical laboratories to detect Blastocystis sp. and D. fragilis are based on microscopy: direct smear (Native-Lugol examination), formalin-ethyl acetate concentration technique (FECT), and permanent staining. However, these methods are known to be insufficient for the definitive diagnosis of these two protists [56][57][58][59]. For Blastocystis sp., culture from stool samples are significantly more sensitive than direct microscopic examination for the detection, but stool cultures can be time consuming making them not practical for diagnosis when a quick turnaround is needed [56,60]. Molecular techniques are progressively replacing microscopy for diagnosis of intestinal parasites, and they are the first-line diagnostic method in laboratories particularly in industrialized countries [61]. Molecular epidemiology studies of Blastocystis sp. and D. fragilis have clearly demonstrated that molecular screening methods are needed in accurately detecting the presence of these protists in stool samples [12,56,57,62,63]. The use of molecular methods to improve detection of Blastocystis sp. and D. fragilis in stool samples is crucial as it is challenging to identify these parasites by microscopy. By improving detection, we also improve our understanding of their epidemiology.
For Blastocystis sp., it is also key to identify subtypes and intra-subtype variability to understand its public health significance and pathogenicity. Several methods have been used to describe Blastocystis sp. mixed subtype infections. Next generation sequencing (NGS) provides a powerful tool for Blastocystis sp. detection that allows: subtype identification, detection of mixed subtypes within a sample, detection of low-abundance subtypes, and intra-subtype variations [8,30,64]. The aim of this study was to detect the prevalence of Blastocystis sp. and D. fragilis in immunocompetent and immunosuppressed patients with diarrhea by conventional and molecular methods. We also use NGS to characterize the prevalence of Blastocystis sp. subtypes and mixed subtypes in these patients. This is the first study to use next-generation sequencing technology to investigate Blastocystis sp. subtypes in Turkey.

Ethics statement
All study procedures, informed consent forms, and epidemiological questionnaires involved in the study were approved by the Ethics Commission of Gazi University (09.05.2017/05). Written informed consent was obtained from the participants.

Study population and collection of samples
Individual stool samples from 438 outpatients were collected between March 2017 and December 2019 at the Gazi University Training and Research Hospital in Ankara, Turkey. The inclusion criteria for the study was to be adult patients (18 years and older) with diarrhea. Diarrhea was defined according to the Bristol Stool Form Scale. Stool samples were examined macroscopically and compared to the Bristol stool chart, and stools consistent with type 6 (fluffy pieces with ragged edges, a mushy stool) and type 7 (watery, no solid pieces, entirely liquid) were considered diarrhea [65]. Patients included in the immunodeficiency patient group were those patients treated at hematology, oncology, rheumatology, nephrology, and bone marrow transplantation units. Patients included in the immunocompetent patient group where those treated at gastroenterology and other clinics with the complaint of diarrhea and with no known immunodeficiency. Exclusions to participate in the study included patients who have inflammatory bowel disease, irritable bowel syndrome, urticaria, under antibiotic treatment, or those that had a colonoscopy in the last three months. Additionally, any patients found positive to Adenovirus and Rotavirus (RIDA QUICK Rotavirus/Adenovirus Combi test, R-biopharm, Germany) or Salmonella/Shigella (culture) were also excluded from the study. Of samples which met the criteria for inclusion in this study, 245 (56.3%) were immunodeficient diarrheal patients and 193 (43.7%) were immunocompetent diarrheal patients. Among immunodeficient patients, four were from rheumatology (immunosupresive treatment recipients), nine were from nephrology, 56 received bone marrow transplant, 70 were from medical oncology, and 106 were from hematology clinics (S1 Table). One hundred thirty-six (55.52%) patients with immunodeficiency were female and 109 (44.48%) were male. The age of the immunodeficient patients ranged from 18 to 85 years with a median age of 55 years. Immunocompetent patients consisted of 99 (51.30%) females and 94 (48.70%) males. The ages of the immunocompetent patients ranged from 1 to 84 years, and the median age was 41 years (p<0,001). All collected stool samples were independently screened for the presence of enteric parasites by conventional (Native-Lugol examination, FECT, trichrome, and acidfast staining) and molecular methods for the detection of Blastocystis sp. and D. fragilis. A flow chart of parasite detection methods used in this study is depicted in Fig 1. Blastocystis sp. and D. fragilis positive stool samples were screened by ELISA test for E. histolytica, G. duodenalis and Cryptosporidium spp. according to manufacturer's recommendations (E. HISTOLYTICA II, TECHLAB, USA; GIARDIA II, TECHLAB, USA; CRYPTOSPORIDIUM II, TECHLAB, USA, respectively).

Microscopy
Fresh stool samples were immediately examined microscopically by preparing saline and iodine wet mounts to test for the presence of enteric parasites. Saline and iodine wet mounts were prepared by mixing a small volume of stool with a drop of physiological saline or Lugol's iodine (diluted 1:5 with distilled water) on a glass microscope slide and placing a coverslip over the mixture [66]. Entire coverslips were examined systematically at 10X and 40X magnification under a light microscope (CX31, Olympus, Japan).
The formalin-ethyl acetate concentration technique was conducted for all stool samples as previously described [67]. Briefly, for each sample, 3 ml of ethyl-acetate solution were added to 10 ml of filtered stool suspension and the tubes were vigorously shaken and centrifuged at 500 × g for 10 minutes. After centrifugation, the supernatant was discarded and the pellet was placed on a microscope slide, covered with a coverslip, and examined microscopically as described above. Additionally, trichrome and Kinyoun's acid-fast stained smears were prepared from all stool samples, after parasite concentration, and examined using a 100X immersion objective by screening a minimum of 200-300 fields [66].

Blastocystis sp. culture
Fresh stool specimens were cultured in 2 ml Eppendorf tubes containing Jones' medium with 10% horse serum and incubated at 37˚C for 48-72 hours then examined microscopically to detect Blastocystis sp. [60].

DNA extraction
Genomic DNA was isolated from all stool samples. First, approximately 200 mg of each fecal sample was lysed using a tissue homogenizer (Qiagen TissueLyser LT, Hilden, Germany) for 15 minutes by adding 200 mg of acid-washed glass beads prior to DNA extraction then processed according to the manufacturer's recommendation using a QiaAmp DNA Stool mini-kit (Qiagen, Hilden, Germany). DNA was eluted in 100 μl elution buffer provided with the kit. DNA extraction was performed weekly. DNA was stored at -20˚C until molecular analyses were performed.

Molecular detection of Blastocystis sp. and D. fragilis
All stool samples were tested by quantitative polymerase chain reaction (qPCR) for both Blastocystis sp. and D. fragilis every two weeks. Molecular detection of Blastocystis sp. was achieved

PLOS NEGLECTED TROPICAL DISEASES
by a qPCR method to specifically amplify a 118-bp fragment of the SSU rRNA gene of the parasite [68]. Amplification reactions (25 μl) contained 12.5 μl of PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA), 0.5 μM of the primer pair Blasto_FWD_F5/Blasto_R_F2, 0.3 μM of probe (S2 Table), and 2 μl of template DNA. Cycling parameters were 95˚C for 3 min, and 40 cycles of denaturation at 95˚C for 15 s followed by annealing and extension at 57˚C for 1 min.
Detection of D. fragilis was achieved by a qPCR protocol amplifying a 78-bp fragment of the SSU rRNA gene of the parasite [69]. Reaction mixes (25 μl) consisted of 12.5 μl PCR of Master Mix (Thermo Fisher Scientific), 0.5 μM of the primer pair DF3/DF4, 0.3 μM of probe (S2 Table), and 2 μl of template DNA. Cycling parameters were the same as for Blastocystis sp.: 95˚C for 3 min, and 40 cycles of denaturation at 95˚C for 15 s followed by annealing and extension at 57˚C for 1 min. For all Blastocystis sp. and D. fragilis negative qPCR results, these DNAs were diluted tenfold with 1xTE buffer and then qPCR was repeated.
Positive (target DNA previously identified as Blastocystis sp. or D. fragilis by DNA sequencing) and negative (sterile water) controls were included in each run. Amplification reaction was carried out in a Rotor-Gene 6000 real-time cycler (Rotor-Gene Q, Germantown, MD, USA).

Molecular characterization of Blastocystis sp. subtypes
To detect Blastocystis sp., a ca. 500 base pair fragment of the Blastocystis sp. SSU rRNA gene, which contains a variable region suitable for subtyping, was amplified by PCR [70]. PCR products were analyzed using a QIAxcel (Qiagen, Valencia, CA, USA). All positive samples were used to conduct next generation amplicon sequencing and libraries were prepared as previously described [64]. Briefly, all positive samples were amplified by PCR using primers ILMN_Blast505_532F and ILMN_Blast998_1017R. These primers amplify a region of the SSU rRNA gene and are identical to Blast505_532F/Blast998_1017R [70], except for containing the Illumina overhang adapter sequences on the 5 0 end. PCR conditions were as follows: 95˚C for 4 min, 35 cycles of 95˚C for 30 s, 54˚C for 30s, and 72˚C for 30 s, and a final elongation step at 72˚C for 5 min. Each 25 μl PCR reaction contained 2.5 μl template DNA, 12.5 μl 2X KAPA HiFi HotStart ReadyMix (KAPABioSystems, Cape Town, South Africa), 2.5 μl BSA (0.1 g/10 ml),1 μM of each primer, 7.75 μl H 2 O, and 1.25 ul of BSA (0.1 g/10 ml). Final libraries were quantified using the Quant-iT dsDNA Broad-Range Assay Kit (ThermoFisher, Waltham, MA, USA) on a SpectraMax iD5 (Molecular Devices, San Jose, CA, USA) prior to normalization. A final pooled library concentration of 8 pM with 20% PhiX control was sequenced using Illumina MiSeq 600 cycle v3 chemistry (Illumina, San Diego, CA, USA). Paired-end reads were processed and analyzed with an in-house pipeline that uses the BBTools package v38.22 [71], VSEARCH v2.8.0 [72], and BLAST+ 2.7.1. After removing singletons, clustering, and the assignment of centroid sequences to operational taxonomic units (OTU) was performed within each sample at a 98% identity threshold. Only those OTUs with a minimum of 100 sequences were retained. All OTUs were assigned a Blastocystis sp. subtype based on the best match by BLAST search in the GenBank database. The nucleotide sequences for unique sequences obtained in this study have been deposited in GenBank under the accession numbers MW728054-MW728093.

Data analysis
Between-group differences in baseline characteristics and parasite prevalence were calculated using Pearson chi-square test, Fisher's exact test, and Kruskal-Wallis rank test in STATISTICA 12.0 (StatSoft, Tulsa, OK, USA). Proportion confidence limits were carried out using software available at http://openepi.com/Proportion/Proportion.htm. Cohen's kappa index for intertest agreement was calculated using software available at (http://openepi.com/DiagnosticTest/ DiagnosticTest.htm). Kappa considered values < 0 as indicative of no agreement,0-0.20 as slight agreement, 0.21-0.40 as fair agreement, 0.41-0.60 as moderate agreement, 0.61-0.80 as substantial agreement, and 0.81-1 as almost perfect agreement [73]. Medians and means of cycle threshold (Ct) values were calculated and a two-tailed Student t test for comparison of medians was carried out using STATISTICA version 12.0. Odds ratios (OR) and their 95% confidence intervals (CI) were calculated by univariate and multivariable analyses using logistic regression models to assess the association between potential risk factors. Exact logistic regression was used to calculate univariate odds ratios to avoid division by zero (https://stats. blue/Stats_Suite/logistic_regression_calculator.html). The qPCR was considered the reference test to compare results obtained by conventional methods for detection of Blastocystis sp. and D. fragilis. The statistical significance level was considered p <0.05 for all analyses.

Comparison of methods used in the diagnosis of Blastocystis sp. and D. fragilis
The diagnostic performance of saline, Lugol's iodine, FECT, trichrome stain, and culture were compared with qPCR ( Table 2). Both Blastocystis sp. and D. fragilis were detected more frequently by qPCR than by any other detection method.
When comparing conventional methods with Blastocystis sp. qPCR, kappa value was detected as the highest only in the culture method (k = 0.84, perfect agreement). Dientamoeba fragilis apart from qPCR could only be detected in trichrome staining and the kappa value was found very low (k = 0.27, fair) ( Table 2).

Blastocystis sp. subtypes identified using NGS
Out of the 73 positive qPCR samples, only 68 isolates were positive when PCR was done to prepare NGS library. The remaining five samples were not sequenced using the MiSeq platform. Clustering yielded 40 unique Blastocystis sp. OTUs across the 68 Blastocystis sp.-positive samples (Table 5). Five Blastocystis sp. subtypes (ST1, ST2, ST3, ST4, and ST6) were found. ST4 was detected only in an immunocompetent patient.

Blastocystis sp. intra-subtype variability
Forty unique OTUs were detected among the five Blastocystis sp. subtypes identified in this study. A high degree of intra-subtype diversity was observed for ST1 and ST2 with eight unique OTUs among the nine ST1-positive samples and 20 unique OTUs among the 19 ST2-positive samples (Table 5). ST1 and ST2 Blastocystis sp. positive samples frequently contained multiple unique OTUs. In fact, up to three unique OTUs were detected in a single ST2 sample (Table 5). Interestingly, ST3 displayed low intra-subtype diversity relative to the number of positive isolates, with only nine unique OTUs among 51 ST3-positive samples. ST4 had one unique OTU in only one ST4-positive sample. ST6 had two unique OTUs among two ST6-positive samples.

Discussion
In the present study, Blastocystis sp. and D. fragilis were investigated in immunodeficient and immunocompetent diarrheal patients using conventional and molecular methods. This study demonstrates that the successful diagnosis of Blastocystis sp. and D. fragilis infections depends on the detection method. Using conventional methods such as direct smear, it is possible to overlook protists, especially when few organisms are present. In fact, D. fragilis was not detected in any samples by direct smear. In the diagnosis of Blastocystis sp., the sensitivity of direct smear, FECT, and trichrome smear was significantly lower than qPCR, while culture had a similar diagnostic accuracy when compared to qPCR. Our findings are in agreement with other studies that have also shown low sensitivity of microscopic methods to detect Blastocystis sp. and D. fragilis [56,74]. Studies that compared direct microscopy, culture and qPCR for Blastocystis sp. detection have reported that qPCR was the most sensitive method [75,76]. Similarly, a study comparing just culture and qPCR to detect Blastocystis sp. reported a higher sensitivity of qPCR [67]. However, the effectiveness of the culture method for Blastocystis sp. diagnosis has also been demonstrated by other researchers [77,78] and the mini-culture method is a practical method especially for diagnostic laboratories with a limited budget [79]. In this study, the overall prevalence of Blastocystis sp. was 16.7% (n = 73/438) by qPCR. The prevalence of Blastocystis sp. has been reported to vary widely among studies ranging from 0.54% to 88.8% [20, 26, 47, [80][81][82]. Prevalence variations could be related to many factors such as studies conducted in different geographical regions, different populations (socio-economic level, immune status, age. . .), or use of different diagnostic methods for detection [34]. In Turkey, Blastocystis sp. is the most common gastrointestinal parasite reported and prevalence ranges between 0.5% to 37.9% (S5 Table) [18,21,[83][84][85][86][87][88][89][90]. Blastocystis sp. carriage was detected in limited studies from the same province in Turkey as 14.2-14.6% in school children [91,92] and 15.5% in adults [75]. Both the present and the other studies from Turkey indicate no apparent differences between symptomatic and asymptomatic cases, supporting the commensal nature of Blastocystis sp. [75,91,92].
Blastocystis sp. was observed in 20.7% and 11.4% of immunocompetent and immunodeficient diarrheal patients examined in this study, respectively. One limitation of this study is a lack of a healthy comparison group (no diarrhea) as all samples were collected from hospital patients presenting with diarrhea. Thus, no comparison can be made of Blastocysits prevalence between symptomatic and asymptomatic individuals. The prevalence of Blastocystis sp. observed in immunodeficient patients in this study is similar to prevalence reported in cancer patients in Turkey (6.5-10.8%) (S5 Table) [84,93]. However, studies in some neighbour countries have reported higher Blastocystis sp. prevalence in patients with cancer ranging from 22.3 Table 6. Distribution of Blastocystis sp. subtypes in the immunodeficient (n = 28) and immunocompetent (n = 40) patient groups.

PLOS NEGLECTED TROPICAL DISEASES
to 27.5% [14,94]. In a prospective study in France using qPCR, similar prevalence of Blastocystis sp. was found in immunodeficient patients (16%;15/94) and immunocompetent patients (13%;12/92) [76]. In our study, immunocompetent group had a significantly higher Blastocystis sp. prevalence than immunocompromised group. It is possible that immunocompromised patients that are under constant medical monitoring tend also to avoid contact with the external environment and behave carefully making them less likely to contract parasitic infections. We did not observe significant differences in Blastocystis sp. prevalence between males and females. However, a previous study detected Blastocystis sp. more frequently in males than in females in immunocompromised patients in Iran [95,96]. We found significant differences in Blastocystis sp. prevalence between immunocompetent males (22.3%) and immunodeficient males (11.0%). Additionally, a significantly higher Blastocystis sp. infection rate was observed in the age range of 64-74 years in immunocompetent patients (44.4%) than in immunocompromised patients (8.3%). This differs from results of a study conducted in Turkey in patients with gastrointestinal symptoms that found highest prevalence of Blastocystis sp. in the 20-29 age group (28.9%) [75]. Unfortunately, there was no information regarding to the immune statuses of those patients. An age-related epidemiological pattern was also reported in a study in France in which higher prevalence was observed in patients between 15 and 49 years of age (22.2%) than the patients over 50 years of age (16.6%) [26]. In the same study, the prevalence of Blastocystis sp. in immunocompromised subjects (12.4%) was significantly lower than in immunocompetent patients (24.2%). The patients were further divided into immunocompromised subgroups which were HIV, solid organ transplants, immunosuppressive therapy, solid cancer, and bone marrow transplants. The results of these subgroup analyses validated that the prevalence of Blastocystis sp. was considerably lower in subjects getting immunosuppressive treatment (8.4%) and bone marrow transplant (7.7%) but not significantly lower in the other subgroups [26]. The patients recieving bone marrow transplant had lower prevalence (7.1%) than the other immunodeficient patient subgroups (S1 Table).
Because qPCR allows quantification, it has been reported that the use of qPCR in largescale surveys could assist in identifying whether the development of symptoms is related to infection intensity by simple analysis of Ct values [34]. We found a significantly lower median Ct value in immunocompromised females (Ct = 21.4) than in immunocompetent females (Ct = 26.5) ( Table 4) that appears to indicate a higher burden of Blastocystis sp. in females with immunodeficiency. A multicenter case-control study in The Netherlands designed to clarify the clinical importance of qPCR in patients with gastroenteritis found a higher mean (SD) Ct value of Blastocystis sp. and D. fragilis in cases as compared to controls [6]. Interestingly, the hypothesis of this study was that during an episode of diarrhea, "a flush out effect" could decrease the load of Blastocystis sp.. However, no information was given about the immune status of the cases.
Dientamoeba fragilis prevalence has been reported to range from 0% to 82.9% in studies using conventional or molecular methods [12,43,97]. Differences have been attributed to variations associated to geographical region, socioeconomic status, and diagnostic methods [12,51]. In Turkey, the prevalence of D. fragilis ranges from 0% to 18.3% [48, 49, 97-109] (S6 Table). The overall prevalence of D. fragilis found in this study (11.9%) falls within the previously reported range for Turkey. A limited number of studies exist on D. fragilis carriage in Turkey. However, D. fragilis was found in 3.4% of school children by trichrome staining in one study [91]. In another study, D. fragilis was detected in 12.04% of outpatients with gastrointestinal symptoms such as diarrhea, abdominal pain and nausea [48]. In the second study, real-time PCR was used and diarrhea was statistically more significant in patients with the presence of D. fragilis. In this study, although D. fragilis was detected more frequently (13.5%) in the immunocompetent patients with diarrhea than immunodeficient patients with diarrhea (10.6%), these differences were not statistically significant. In Turkey, using qPCR, two studies reported a prevalence for D. fragilis in patients with gastrointestinal symptoms, with unknown immune status, of 10.7% and 12.4% [48,109]. There are only a few articles about the prevalence of D. fragilis in immunodeficient patients. In a study from Iran, D. fragilis was found 1.2% of 190 patients including primary immunodeficiency patients, cancer patients and organ transplants recipients [14]. A study in HIV positive men with diarrhea in Australia reported a prevalence of 0.3% for D. fragilis [44]. Dientamoeba fragilis was detected only by microscopy in both studies. In the Netherlands, using the multiplex real-time PCR method D. fragilis was detected in 25.8% and 37.6% of patients with gastrointestinal complaints and without complaints, respectively [6]. In our study, no statistically significant differences were found between the two groups in terms of prevalence, gender, age, and seasonal variables for D. fragilis (Table 3). Case-control studies with a large number of samples are needed to determine the relationship between D. fragilis and symptomatology.
In this study, co-infection with D. fragilis was detected in 12 (17.6%) of the of 73 Blastocystis sp. positives. This coinfection rate was lower than previously reported in other studies [48, 95,110]. To investigate the relationship between D. fragilis colonization and specific gastrointestinal symptoms, and sociodemographic characteristic, a cross-sectional study that included 490 fecal specimens were collected from outpatient with gastrointestinal symptoms and tested with qPCR in Turkey [48]. Their results suggested that D. fragilis is a pathogenic parasite and that the most common clinical symptom found in infected patients is diarrhea [48]. They found that 23.7% were co-infected with Blastocystis sp. [48]. A study in Italy that included 756 patients suspected of harboring intestinal parasites were subjected to multiplex RT-PCRs to detect parasites [95]. The prevalance of Blastocytis was 34% and co-infection with D. fragilis was detected in 24% of patients [95]. Our study showed lower co-infection than both of those studies. The common occurrence of co-infection of these two parasites may indicate that there are shared sources of transmission.
Low infection rates were found for G. duodenalis and Cryptosporidium spp. in this study (<1%). This situation may be explained in part by the study location which is located in the center of Ankara, the second biggest city and the capital city of Turkey. Ankara is known also as a civil servant city, far from the places where agriculture and animal husbandry are made. The majority of the population lives in apartments and uses carboys or purified water. Healthcare workers like nurses and doctors also inform immunocompromised patients about protection from infections. Similar results were obtained in a study conducted in another university hospital in Ankara. The period of the study was 2003-2012 and G. duodenalis and Crptosporidium spp. prevalences were 1.3% and 0.003% respectively. Out of the 85,707 fecal samples examined, 3,681 (4.2%) were positive to parasites [108]. Another study also reported low prevalence of G. duodenalis (0.61%) and Cryptosporidium spp. (0%) from one of university hospitals in Izmir city (the third biggest city of Turkey) in the western region of Turkey [48]. To date, several methods based on specific primers for determining subtypes of Blastocystis sp. have been developed [60,70,[111][112][113][114][115]. However, only a few of these methods can be used to determine mixed subtypes [111,113,115,116]. The disadvantages of STS-PCR [113,116] and ST-specific nested PCR assay [115] are being time-consuming, having a high risk of contamination, and being able to identify only certain STs. Recently, NGS was used for identification of Blastocystis sp. subtypes with the advantage of allowing identification of all currently known subtypes, the ability to evaluate mixed infections, and the ability to detect intra-subtype diversity within a single sample quickly and with only a small amount of input material conserving both time and valuable sample material. [8,30]. Furthermore, the NGS method provides improved sensitivity for identifying subtypes with zoonotic potential that could be in low proportions within a specimen. This is the first study adopting an NGS method to investigate the genetic diversity of Blastocystis sp. in Turkey, as previous studies conducted in the country were based on STS-PCR and Sanger sequencing [21, [83][84][85][86][87]. NGS was used to study Blastocystis sp. genetic diversity in a rural human population from Mexico where 3 subtypes were reported, ST1-ST3, with the following frequencies: ST3 (67.7%), (ST2 11.3%), ST1 (7.3%), ST1/ST3 (7.3%), ST2/ST3 (4.0%), and ST1/ST2/ST3 (2.4%) [8]. Both the present study and the study in Mexico report ST1, ST2, and ST3, with ST3 being the most prevalent subtype. However, we have a higher proportion of mixed infections than the study in Mexico. Another difference between the two studies is that we identified two additional subtypes in patients in Turkey, ST4 and ST6. Previous molecular studies in Turkey have shown that ST1-ST4 represent 87.5% of Blastocystis sp. positive human samples, ST5-ST7 represent 3.5%, and mixed subtype infections represent 9% (S5 and S7 Tables).
Globally, ST1-ST4 have been identified as the most common subtypes in humans. Thus, the findings of our study are consistent with other studies carried out in Turkey and worldwide [10,75,83,87,117,118]. ST2 was the second most common ST in our study. This is in agreement with other studies in humans that also report ST2 as the second most common ST after ST3 [8,81,[119][120][121]. In Turkey, Dogruman-Al, et al. (2008) reported ST2 as the second most common subtype after ST3 with a statistically significant association to asymptomatic patients, suggesting that ST2 may be a non-pathogenic subtype of Blastocystis sp. in this population [117]. Another study in Turkey reported a relationship between ST1 and abdominal pain [122]. In our study, ST4, which is generally found across Europe, was detected in only one immunocompetent patient. In Turkey, reports in humans of ST4 are sporadic (S4 Table) [86,118,123]. In our study, ST6 was detected in one immunocompetent and one immunodeficient patient. There are other reports of ST6 in humans from Turkey, Poland and South America [30,31,33,75,87,123,124]. Because ST6 is mostly identified in birds, its presence in humans may indicate the potential for zoonotic transmission [31]. In the present study, mixed subtype infections of Blastocystis sp. represented 20.6% of the total subtype identified. The incidence of mixed subtypes in Turkey has been reported to range from 3.2% to 30.5% with the STS-PCR method (S7 Table). We found no statistically significant differences between immunocompetent and immunodeficient patients for both Blastocystis sp. positivity and mixed subtype distribution ( Table 6). The importance of the mixed subtype infection of Blastocystis sp. is not clear yet for symptomatology or pathogenicity, and additional studies that include methodology to detect mixed subtype infections are needed.
Some researchers have suggested that intra-subtype variability could have a role in the transmission and pathogenicity of Blastocystis sp. [8,29,30,64,125]. This study showed intrasubtype variations only for ST2 and ST1. Other subtypes (ST3, ST4, and ST6) did not have within sample intra-subtype variations. However, result for ST4 and ST6 should be taken with caution because ST4 and ST6 were only identified in 1, and 2 patients, respectively. Intra-subtype variability observed for ST1-ST3 is consistent with intra-subtype variations reported in previous studies in Iran and Mexico [8,126]. In the study in Mexico, intra-subtype variability was also more common in ST2 and ST1 than in ST3 [8]. However, there is still limited data and more studies reporting intra-subtype variability are needed to understand the potential role of this variability in pathogenicity, zoonotic potential, and transmission.
This study provides valuable information about Blastocystis sp. and D. fragilis in humans. However, because no healthy controls were available for comparison, conclusions about the role of these parasites in health and disease could not be drawn. In the future, further case-control studies implementing high-resolution molecular tools or functional genomic analysis are necessary to understand the role of intra-subtype variation of Blastocystis sp. in pathogenicity or symptomatology and the role of neglected protists in health and disease.