Blastocystis colonization and associations with population parameters in Thai adults

Background Blastocystis is a unicellular eukaryote commonly found in the intestinal tract of humans and other animals. The prevalence of Blastocystis has been investigated in both developed and developing countries, yet its occurrence and distribution in rural locations has been less studied. Herein, we aimed to examine the distribution of Blastocystis colonization in Thai adults representing background populations along a rural/peri-urban gradient, as well as associations between colonization and personal characteristics. Methodology A total of 238 participants were recruited from rural and peri-urban areas situated in three provinces. The presence of Blastocystis in feces was evaluated using PCR and qPCR. Information on gender, age, region (province), rural/peri-urban location, and body mass index (BMI) was collected. Principal findings The overall rate of Blastocystis carriage was 67.2%. Univariate analysis revealed significant associations between Blastocystis carriage and region (p<0.05), location (p<0.001) and age group (p<0.05). Logistic regression analysis revealed that rural/peri-urban location and BMI were significantly associated with Blastocystis carriage. Nine subtypes (ST1-ST7, ST10 and ST23) were identified with ST3, ST7 and ST1 as the most abundant ones, in this order. The greatest diversity of subtypes, in terms of numbers, was found in the middle aged group (nine subtypes), while the least diversity was found in the young adult and obese (three subtypes each) groups. Conclusions This study increases the understanding of the epidemiology of Blastocystis colonization and its association with population parameters and characteristics in middle-income countries.

Nonetheless, few studies have provided data on the distribution of Blastocystis subtypes in humans across both rural and urban regions.Moreover, the prevalence of Blastocystis and certain subtypes in relation to BMI is poorly understood, especially in middle-income countries.The aim of the present study was to investigate the positivity rate and distribution of Blastocystis in Thai adults from the background population, focusing on identifying an urban-rural gradient and associations between BMI and colonization.

Ethics statement
The Human Ethics Committee of Mae Fah Luang University (Ethics registry number REH60103) and the Human Ethics Committee of Phramongkutklao College of Medicine (License number S053q/58) approved of the collection of human fecal samples.All enrolled participants were informed of the objectives and procedures involved in this study.All participants signed an informed consent form prior to the collection of fecal samples.

Study sites and sample collection
Two hundred and thirty-eight fecal samples were obtained once from adults with no gastrointestinal symptoms in three provinces: 95, 57, and 86 samples from Chiang Rai (CR), Phayao (PY), and Chachoengsao (CH), respectively (Fig 1).Forty-five samples were from a previous collection (GenBank accession numbers: OL351649-OL351797; 9), while the rest were newly collected.Three locations were sampled in the CR and the CH area, and one village was sampled in the PY area.The inclusion criteria were adult age, while gastrointestinal symptoms at the time of sampling, diagnosed gastrointestinal disease, and antibiotic treatment up to two months prior to sampling were the exclusion criteria.

Metadata
Metadata available for study included body mass index (BMI), age, gender, location (rural and/or peri-urban) and region (province)(S1 Data).BMI was calculated and assigned categories as follows: 'lean' (<25.0 kg/m 2 ), 'overweight' (25.0-29.9kg/m 2 ), and 'obese' (� 30.0 kg/ m 2 ).Participants were also categorized according to the Asia-Pacific guidelines recommendation for Asian populations, which calls for lower cut off points for 'overweight' (23.0-27.5 kg/ m 2 ), and 'obese' (> 27.5 kg/m 2 ) [53,54].Age was classified as 'young adult', 'adult', 'middleaged', and 'aged' for the age groups 19-24, 25-44, 45-64 and �65 years, respectively [55].The gender categories 'male' and 'female' were used.Finally, a location was considered as rural or peri-urban based on information on population size, type of area, occupation and availability of medical services and basic necessities [51,56], while the regional origin of all volunteers was divided into provinces (Chiang Rai, Phayao and Chachoengsao).All data collected from each subject remained confidential and were fully anonymized through the encryption of the identity of individuals.

DNA extraction
Total genomic DNA was extracted directly from human feces using QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's recommendations.The DNA was stored at -20˚C until analyzed.

Molecular detection and differentiation of Blastocystis
The samples were not screened for co-infections with other parasites, bacteria or viruses.
The nuclear SSU rDNA was used for screening and subtyping purposes.To achieve this, both quantitative polymerase chain reaction (qPCR) and polymerase chain reaction (PCR) were used.The former was employed for Blastocystis screening because of its high sensitivity, however the fragments obtained were of short length (~300 bp) and not always suitable for subtype identification.Hence, conventional PCR was used on all samples to obtain the barcoding region, which comprises a longer fragment and is widely used for subtyping.The approach was also used to compare the results with qPCR.
PCR.PCR amplification of DNA was performed using a nested PCR reaction.Briefly, the first PCR used the broad-specificity eukaryotic primer RD5 (5'-GGAAGCTTATCTGGTTGA TCCTGCCAGTA-3') and RD3 (5'-GGGATCCTGATCCTTC CGCAGGTTCACCTAC-3'), which amplifies a ~1,800 bp fragment of the SSU rDNA [58].The following reaction was used: initial denaturation for 3 min at 94˚C, 35 cycles with denaturation at 94˚C for 100 s, annealing at 65˚C for 100 s, extension at 72˚C for 100 s, and a final elongation step at 72˚C for 10 min.The first PCR product was used as template in the second PCR reaction (PCR product, ~660 bp) with the Blastocystis barcoding primers RD5F (5'-ATCTGGTTGATCCTGCCAGT-3') and BhRDr9R (5'GAGCTTTTTAACTGCAACAACG-3') [59].The PCR involved initial denaturing at 94˚C for 3 min, followed by 30 cycles including denaturing at 94˚C for 60 s, annealing at 60˚C for 60 s, and extension at 72˚C for 100 s, which was followed by an additional 10-min elongation at 72˚C.The reaction used EmeraldAmp GT PCR Master Mix (TaKaRa Bio USA, Inc.).

qPCR and PCR product purification and sequencing
All positive PCR and qPCR products were purified using the MEGAquick-spin Plus Total Fragment DNA Purification Kit (iNtRON Biotechnology, Inc., Korea) according to the manufacturer's instructions and sequenced using the reverse (PCR or qPCR) primer (Bionics Company, Korea).

Cloning
Seven PCR and fifty-five qPCR products for which mixed peak chromatograms were observed were cloned using the pLUG-Prime TA-cloning Vector Kit II (iNtRON, Korea) following the manufacturer's protocol.To maximize the chances of obtaining a clean sequence, two colonies per transformation were screened from each sample and sequenced by Sanger sequencing (Bionics Company, Korea).

Sequence analysis and subtype identification
The raw reads were edited and ambiguous bases at both ends were removed.The sequences were then used as queries to perform a BLAST search to check whether they were indeed Blastocystis.The sequences were then subtyped using the curated PubMLST Blastocystis database available online (https://pubmlst.org/organisms/blastocystis-spp).All sequences were submitted to GenBank under accession numbers PP728964-PP729011, PP729403-PP729456 and PP731515-PP731535.

Statistical analysis
Statistical analyses were carried out using the Stata13 [60](Stata Statistical Software: Release 13.College Station, TX: StataCorp LLC) software program to examine possible association of Blastocystis carriage and population parameters.The D'Agostino test was used to determine the distribution of the continuous parameters measured in this study.Age was not normally distributed; hence, five age groups were developed (see metadata section above).Data were also analyzed by region for regional variations according to gender, age, location, and BMI status using WHO and Asia-Pacific classification criteria (χ 2 test).Raw BMI data were used after logarithmic transformation.Pairwise comparison of means with Bonferroni correction was used to examine associations with the different population parameters.Blastocystis carriage and associations with other parameters were examined using univariate and multivariable analysis.For the univariate analysis, data (% proportions) were grouped according to Blastocystis carriage and associations with gender, region, location and age were tested.Pairwise comparison of means with Bonferroni correction was used to examine associations with combinations of different population parameters.For the multivariable level analysis, logistic regression models were developed to examine the associations between Blastocystis carriage and population parameters.Goodness of fit was established using Akaike information criterion (AIC) and Bayesian information criterion (BIC).

Results
A total of 238 human fecal samples were collected from 80 (33.6%) and 158 (66.4%) male and female individuals, respectively.The overall age range was 19-82 years.The age distribution was not normal and no transformation was approaching normality.Information on BMI was not available for the 86 participants from the CH province.The BMI range for the remaining study participants was 15.6-37.8 with a geometric mean of 23.87 (95% confidence interval (CI): 23.29-24.48).
Univariate analyses revealed no significant association between sex and regional distribution of the samples (χ 2 = 0.7605, p = 0.684).The distribution of samples by age group was not the same in each region (χ 2 = 19.6746,p<0.001).In CR, the samples from individuals younger than 25 years represented 28.4% of the total in the region.In PY and CH samples from individuals younger than 25 years represented 1.8% and 2.3% of the total in each area, respectively.There was no significant difference in regional BMI, when raw BMI data was used (t = 1.0912, df = 150, p = 0.2769).Using the WHO BMI status, no significant difference was identified between CR and PY (χ 2 = 0.233, p = 0.233).When the Asian BMI status was examined, the difference was significant (χ 2 = 36.0773,p<0.05), with proportionally more lean (LN) individuals identified in the CR province (43.2%) compared to PY (28.1%).The same applied to obese (OB) individuals, 21.1% in CR and 12.3% in CY.Overweight individuals had higher representation in the PY region (59.7%)compared with CR region (35.8%).The majority of the samples originated from rural areas (168, 70.6%).This was consistent across all regions, with PY having no samples from peri-urban locations.The associations of the different population parameters by region are summarized in Table 1.

Blastocystis prevalence and association with variables
Overall, by qPCR, 160 out of 238 samples (67.2%, 95%CI: 60.96-72.93)were positive for Blastocystis.When using PCR, the overall detection rate was 47.1% (112/238).Samples positive for PCR were also positive for qPCR.Fifty six PCR products were false positives as evidenced by Sanger sequencing (Fungi, Plantae), while only two qPCR products (Fungi) were false positive.False positives were not included in the calculations determining carriage and subsequent analysis.In terms of the 55 cloned products, 18  We tested the association between Blastocystis carriage and BMI, where univariate analysis identified a significant association between the two (t = -2.9821,df = 150, p<0.005).The power of the test (α = 0.05) was 0.8135 (   3.
Multivariable analysis.Multivariable analysis led to a logistic regression model (n = 152, LRχ2 = 19.41,p<0.001) that identified BMI (OR [odds ratio] for each point:    PLOS NEGLECTED TROPICAL DISEASES
According to gender, Blastocystis ST3, ST1, and ST7 were the most frequently found in males, while ST3, ST7 and ST1 were the most frequently found subtypes in females in that order.Blastocystis ST1-ST7, ST10 and ST23 were identified in lean and overweight, whereas in obese participants carriage was limited to ST1, ST3 and ST7 (Table 5).Finally, Blastocystis ST1, ST3 and ST7 were the most common subtypes across all age groups.The middle-aged (45-64 years) individuals exhibited the highest diversity of subtypes.Regarding region, Blastocystis ST1-ST7, ST10 and ST23 were observed in CR province, ST1-ST3 and ST7 in PY province and ST1, ST3, ST5, ST7 and ST10 in CH province (Fig 3).The top three Blastocystis subtypes in CR were ST3, ST1 and ST23, while ST3, ST7 and ST1 were the most abundant in PY and CH province, in this order.
When looking at the distribution of Blastocystis subtypes according to type of location, greater diversity was observed in the rural area (ST1-ST7, ST10 and ST23) compared with the peri-urban locations (ST1, ST3, ST5 and ST7) (Fig 4).Moreover, in the CH rural location ST1, ST3, ST7, ST10 were detected, whereas in the CH peri-urban areas ST1, ST3, ST5 and ST7 were found.Lastly, ST1-ST3 and ST7 were characterized in the rural location of PY.

Discussion
Blastocystis is one of the most frequently reported intestinal microeukaryotes in human fecal specimens, but it has also been identified in animals and the environment [2][3][4]61,62].Limited data on Blastocystis reported from humans indicates a higher prevalence in Africa and Asia [44,45,47,[63][64][65].Studies in Thailand have reported positivity rates ranging between 14.5% to 73.0% in adults [7,9,66] and 4.8% to 89% in children [63,[67][68][69][70].In the present study, the overall Blastocystis colonization rate was 67.2% (160/238) using molecular methods in line with our previous work in the area [9].The reported differences in prevalence across studies are likely due to the variable sensitivity of the methods used.This is also exemplified herein, whereby the detection rate using qPCR was higher than when employing conventional PCR.PLOS NEGLECTED TROPICAL DISEASES Moreover, variability of populations and the demographic factors investigated also likely contributed to the different results.
The prevalence of Blastocystis and associations with population parameters was examined.Blastocystis carriage was not associated to the gender of the participants in line with earlier observations [6,[71][72][73][74].In terms of BMI, there was no association between prevalence and BMI WHO classification in the univariate analysis.However, there was a significant association when considering the Asia-Pacific classification.Applying a logistic regression model  identified a significant association with BMI (both WHO and Asia-Pacific classification) confirming previous investigations [1,18,19,75].The higher prevalence of Blastocystis in obese rather than lean subjects is contrary to several previous studies.One could speculate that the low number of obese participants may have affected the result though this was not the case when BMI raw data were used in the analysis.Nonetheless, data is emerging that Blastocystis carriage rate in obese individuals might be higher than previously thought.Specifically, the organism was found in over 40% of three separate obese populations [76][77][78].
In the univariate analysis, a significant association between Blastocystis carriage and age was detected in line with previous studies [6,8,71,79].In testing the model, age groups were also examined and, given the variable age distribution in peri-urban and rural areas, age is a possible contributor to the association between location type and Blastocystis carriage, yet no interactions were identified and both AIC and BIC favored the model including only the type of location.
Blastocystis occurrence in rural areas was significantly higher than in peri-urban locations, aligning with previous findings in Turkey, Czech Republic, Iran and Qatar [8,64,74,80].Personal and community-level hygiene, geographic area, culture, and lifestyle of a population may play important roles in the distribution and prevalence of the organism.
Nine subtypes were identified including ST1-ST7, ST10 and ST23.The diversity of subtypes was highest in the rural areas; these are locations, where a variety of domesticated animals and wildlife are found in close proximity to humans.The richness of the identified subtypes might be due to the study area, climate, culture, food, and lifestyle.Overall, the top-three dominant subtypes were ST3, ST7 and ST1.While ST1 and ST3 are commonly identified worldwide in humans, ST7 seems to be geographically restricted in Asia [7,9,70].In vitro and mouse model studies, pointed towards ST7 being pathogenic [81][82][83][84].Nonetheless the pathogenicity of this subtype in humans remains unclear.Hence, the notable finding here is the presence of ST7 in gut-healthy participants, which aligns with results from our previous studies in the area [7,70] and argues against ST7 being pathogenic.On the contrary, it points towards ST7 being a common subtype in this gut-healthy population.
How did so many human hosts in our study acquire ST7?One explanation could be their proximity to avian hosts, well-known carriers of this subtype (85).Raising domestic fowl is routine practice in rural Thailand, while wildlife birds also abound.In turn, fowl could be shedding Blastocystis and contaminating soil and water.This would be in line with our previous work in the area, where we identified ST7 in both soil and water (9).
The presence of several ST10 and ST23 sequences is also of interest, as they are not typically found in humans.Although there are some reports of human-origin ST10 and ST23 sequences worldwide, this is uncommon [34,70,86], since these subtypes appear to be found more commonly in ruminants [20].The high number of ST23 sequences in this study raises some questions.One explanation is that the human specimens were contaminated with ruminant fecal matter or the individuals could have been exposed to a fecal/environmental source containing these subtypes simultaneously.Given that all ST23 sequences were found in the same closeknit community, points towards the latter scenario.Notably, both ST10 and ST23 were also detected in schoolchildren in other communities in northern Thailand [70].Nonetheless, this is not the first time that unusual subtypes are observed in a non-western country.In a recent study from Vietnam, more than half of the Blastocystis subtypes were not typical of humans [86].Further study of these communities should include temporal sampling to determine, whether these subtypes are specific to the community or transient passengers from the environment.
A single sequence of the geographically restricted ST4 was found.Subtype 4 has been found at very low frequencies (usually one or two samples per study) in Asia [7,66,[86][87][88][89][90].Given that PLOS NEGLECTED TROPICAL DISEASES this is a subtype commonly found in rodents throughout Asia [91][92][93][94], it is striking that ST4 is barely found in Asian human hosts.This calls into question the zoonotic potential or lack thereof of certain Blastocystis subtypes.
Our study has some limitations.The sample size is relatively modest with some categories being under-represented or absent.For instance, aged individuals from peri-urban locations were absent and BMI data was not available for the CH population.Screening for additional colonies or utilizing high throughput sequencing technologies would have shed light on the extent of mixed infections in this population.

Conclusions
This is a molecular epidemiological study of Blastocystis in Thai adults in rural and peri-urban areas that also examines associations with BMI, age and geographic location (rural/periurban).Living in a rural area and BMI are significantly associated with Blastocystis carriage.This study and others bring into question the factors that determine colonization by a specific subtype.For instance, even though birds worldwide carry ST7, we mainly see this subtype in humans living Asia.Likewise with ST4, while rodents worldwide carry it, mainly humans living in Europe seem to carry it.

Fig 4 .
Fig 4. Distribution of Blastocystis subtypes in humans based on sampling locations.A total of 238 volunteers participated in the study, of which 160 were Blastocystis-positive.ST7 was present in every location sampled and is in red font.Basemap traced from https://earthexplorer.usgs.gov/courtesy of the U.S. Geological Survey, terms of use https://www.usgs.gov/information-policies-and-instructions/copyrights-and-credits.https://doi.org/10.1371/journal.pntd.0012292.g004

Table 3 . Blastocystis detection associations with BMI and BMI status, univariate analysis.
For raw BMI data the power of the test is also included.

Table 4 . Logistic regression models of the associations between Blastocystis carriage and population parameters.
Goodness of fit was established using AIC and BIC.BMI was examined both as raw data and as BMI status using the WHO and Asian classification.