Dientamoeba fragilis diagnosis by fecal screening : relative effectiveness of traditional techniques and molecular methods

Introduction: Dientamoeba fragilis, an intestinal trichomonad, occurs in humans with and without gastrointestinal symptoms. Its presence was investigated in individuals referred to Milad Hospital, Tehran. Methodology: In a cross-sectional study, three time-separated fecal samples were collected from 200 participants from March through June 2011. Specimens were examined using traditional techniques for detecting D. fragilis and other gastrointestinal parasites: direct smear, culture, formalin-ether concentration, and iron-hematoxylin staining. The presence of D. fragilis was determined using PCR assays targeting 5.8S rRNA or small subunit ribosomal RNA. Results: Dientamoeba fragilis, Blastocystis sp., Giardia lamblia, Entamoeba coli, and Iodamoeba butschlii were detected by one or more traditional and molecular methods, with an overall prevalence of 56.5%. Dientamoeba was not detected by direct smear or formalin-ether concentration but was identified in 1% and 5% of cases by culture and iron-hematoxylin staining, respectively. PCR amplification of SSU rRNA and 5.8S rRNA genes diagnosed D. fragilis in 6% and 13.5%, respectively. Prevalence of D. fragilis was unrelated to participant gender, age, or gastrointestinal symptoms. Conclusions: This is the first report of molecular assays to screen for D. fragilis in Iran. The frequent finding of D. fragilis via fecal analysis indicated the need to include this parasite in routine stool examination in diagnostic laboratories. As the length of amplification target correlates to the sensitivity of PCR, this assay targeting the D. fragilis 5.8S rRNA gene seems optimal for parasite detection and is recommended in combination with conventional microscopy for diagnosing intestinal parasites.

Dientamoeba fragilis infections range from asymptomatic to causing acute or chronic disease in children and adults.The most common symptom of dientamoebiasis is diarrhea, followed by abdominal pain, fatigue, anorexia, and flatulence [8][9][10].Dientamoebiasis may occur at any age and has a cosmopolitan distribution.Prevalence of D. fragilis infection varies considerably and is influenced by factors including geographic location, population density, living conditions, and level of hygiene and sanitation [1].Data on the international prevalence of D. fragilis are limited.Worldwide, the prevalence has been reported to range from 0.4% to 71% [2,8,11,12], making it a more frequent cause of gastrointestinal infection than Giardia lamblia [13][14][15].The sensitivity of diagnostic techniques and the expertise of testing laboratories affect the reported prevalence rate of D. fragilis [14,16].Common methods such as direct smear and culture are challenging and require experience to distinguish D. fragilis from other gastrointestinal parasites [17].Accurate identification depends on detection of the trophozoites in permanently stained stool smears, since the nuclear structure cannot be demonstrated in unstained stool samples [18].The staining technique is generally laborious, time consuming, and relatively insensitive.The development of PCR has provided a highly sensitive and specific method for diagnosis of pathogenic protozoa.PCRbased assays using species-specific primers offer a convenient and reliable technique for the detection of D. fragilis [17,19].
Intestinal parasitic infections are a critical public health problem in Iran; however, research on D. fragilis has been limited.Its reported prevalence, as determined by the direct smear method, varies from 0.5 to 2.4% depending on area of the country [20][21][22].Using the iron-hematoxylin staining method, Jamali and Khademvatan [23] reported prevalence of 13.2%.As D. fragilis is a significant human pathogen, further research on its occurrence and effects is warranted [2,24].We therefore aimed to investigate D. fragilis infection in individuals referred to Milad Hospital in Tehran, comparing traditional and molecular methods of detection.

Sample collection
In a cross-sectional study, three fresh fecal specimens, separated by at least one day, were collected from each of the 200 participants referred to the clinical laboratory of Milad Hospital in Tehran, from March through July 2011.Participants provided informed consent and the study was approved by Ethics Committee under number IR.IUMS.FMD.REC 1390.1065.Fecal specimens were immediately submitted to the research laboratory of the Department of Parasitology and Mycology, School of Medicine, Iran University of Medical Sciences.All specimens were investigated for parasites by direct wet-mount microscopy, formalin-ether concentration, culture, modified iron-hematoxylin staining, and two PCR assays for D. fragilis.

Microscopic examination
Direct wet-mount microscopy and formalin-ether concentration methods Stool specimens were investigated microscopically for trophozoites forms of intestinal protozoan parasites using direct wet-mount in saline and iodine-solution (Lugol's iodine) [10].Formalin-ether concentration was conducted to identify ova and cysts or oocysts [25,26].

Sample preparation and culture
To a 10-20 g fecal sample, 50 mL of phosphate buffered saline (PBS) pH 7.4 was added and thoroughly mixed.The suspension was filtered through two layers of gauze and centrifuged at 800 × g for 5 min.Sediments were re-suspended in ~2 mL of PBS before combining with culture medium and fixing in either sodium acetate-acetic acid-formalin (SAF) or 80% ethanol [27][28][29].
For isolation of intestinal protozoa to be cultivated in an axenic medium, feces were cultured in a diphasic medium as described by Clark and Diamond [30]: slope of heat-inactivated horse serum (kindly provided by the Faculty of Veterinary Medicine, University of Tehran, Iran) overlaid with 5 mL of Ringer's solution and supplemented with ~1 mg rice starch (HSr+S).Penicillin-streptomycin (Sigma-Aldrich, Steinheim, Germany) was added to control the growth of human bacterial flora.A 300 µL sample of washed and unpreserved stool were added to culture tubes containing medium and rice starch and incubated in a vertical position at 35.5 °C.A drop of sediment from the tube was examined on a microscope slide three times at 48 hours intervals at 100× and 400× magnification.

Staining
The stool samples fixed in SAF were stained with modified iron-hematoxylin stain according to methods for identification of protozoa [9,27,31].Precise microscopic diagnosis of D. fragilis was based on morphological characters from permanent stained smears at 400× and 1000× magnification.All slides were examined by two independent examiners.

DNA extraction
One mL of stool preserved in 80% ethanol was centrifuged at 1000 × g for 5 minutes, and the sediment was re-suspended in PBS and washed twice in sterile PBS to remove ethanol.After washing, the sediment was re-suspended in 200 µL 2% polyvinylpolypyrrolidone (PVPP) (Sigma-Aldrich, Steinheim, Germany) in PBS, combined thoroughly, and stored at -20 °C for 24 hours [28].The samples were heated for 10 minutes at 100 °C before submitting to DNA extraction using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions, modified according to Verweij et al. [28].
The primers DF400 (5'-TATCGGAGGTGGTAATGACC-3') and DF1250 (5'-CATCTTCCTCCTGCTTAGACG-3') targeting SSU (18S) rRNA [27] in 20 µL final PCR reaction [10 µL of Taq DNA polymerase 2X-preMix (GeneOn, Ludwigshafen, Germany), 2 µL genomic DNA, and 0.4 µM of each PCR primer] with the reaction conditions of 3 minutes at 94 ºC followed by 30 cycles of 94 ºC for 1 minutes, 57 ºC for 1.5 minutes, and 72 ºC for 2 minutes and a final step of 7 minutes at 72 ºC.The PCR products were detected on ethidium bromide stained 1.5% agarose gels.All PCR reactions included a negative control containing sterile distilled water instead of DNA template and a positive control containing genomic DNA extracted from a stool specimen microscopically confirmed to be infected with D. fragilis.Some D. fragilis PCR-positive samples were confirmed by sequencing an 887 bp amplified SSU rRNA gene fragment in both directions (MWG-Biotech Company, Ebersberg, Germany).The sequence results were read by CHROMAS (Technelysium Pty Ltd., Queensland, Australia) and aligned using DNASIS MAX v. 3.0 (Hitachi, Yokohama, Japan).The final SSU rDNA sequencing results were compared with the Genbank database using the BLASTN program (http://www.ncbi.nlm.nih.gov/BLAST/).Phylogenetic analysis was performed in MEGA7 (www.megasoftware.net) using the neighbor-joining method, and the evolutionary distances were computed using the Kimura 2-parameter method and a bootstrap value of 1000.

Statistical analysis
Statistical analysis was performed using SPSS 16.0 software (SPSS Inc., Chicago, USA).A descriptive analysis was conducted to determine the prevalence of parasites by gender, age group, clinical symptoms, and reason for referral.Associations between qualitative variables were evaluated using the chi-square (χ 2 ) test to reveal statistically significant values (p-value < 0.05).

Participant enrollment
Two-hundred participants were enrolled in the study, 50.5% female and 49.5% male.The mean age was 27.6 ± 19.1 years, ranging from one to 79 years.Most participants lived in Tehran Province (86%), with  1).

Discussion
Gastrointestinal parasitic infections caused by helminths and protozoans are common worldwide and occur in most parts of Iran.Factors including method of sample fixation and examination may bias the diagnosis of D. fragilis and other protozoans in stool samples [32].Identification and differentiation of these parasites by common techniques such as direct smear and formalin-ether concentration has been reported to lack accuracy and to be laborious and time consuming compared to molecular assays [17,27,28,33].
In this study we found D. fragilis infections in ~13% of individuals referred to Milad Hospital in Tehran, with no correlation to gender, age, clinical symptoms, or reason for referral.Little information with respect to D. fragilis in Iran is available, and this is the first report of molecular diagnosis in the area.The obtained prevalence agreed with the reported prevalence of 0.4% to 71% worldwide, in which observed variations are primarily dependent on diagnostic method, the studied population, and the geographic region [2,12].The reported prevalence of D. fragilis in Iran varies from 0.5% [21,34] by direct microscopy to 2%-13.2% by iron-hematoxylin staining or trichrome staining in patients with intestinal symptoms [20,22,23].Ghazanchaei et al. [20] and Sarafraz et al. [22] used nested-PCR to confirm D. fragilis identified by permanent staining (2% and 2.4%, respectively).
The impact of diagnostic methods on the reported prevalence of D. fragilis was clearly seen in our study.Although three stool samples from each individual were collected at different times to increase the probability of detecting D. fragilis and other protozoa [8,27], D. fragilis was not detected by direct smear or formalinether concentration methods, similar to previous studies [22,35,36].Two D. fragilis-infected subjects (1%) were revealed by the culture method.It may be that the D. fragilis present were dead, or there may have been an over-growth of other protozoa in the stool samples that prevented D. fragilis replication [19].The prevalence of D. fragilis was 5% with iron-hematoxylin staining.Grendon et al. [32] suggested that accurate and reliable detection of D. fragilis requires permanently stained preparations of fixed or fresh unpreserved stool specimens.However, the accuracy of this technique is low, as the trophozoites of D. fragilis can be easily overlooked due to pale staining of their nuclei, which may resemble those of Entamoeba spp.We used ethanol-preserved stool samples for PCR to prevent DNA fragmentation [17,28].Prevalence varied from 6% when targeting the SSU (18S) rRNA gene to 13.5% with the 5.8S rRNA gene, likely reflecting the different size of amplicons of the 5.8S rRNA (98 bp) and the SSU rRNA (887 bp) genes.Verweij et al. [28] indicated that the amplification of large fragments can reduce the sensitivity of PCR for detecting D. fragilis directly from stool specimens.
In addition to conflicting reports of D. fragilis worldwide prevalence, the influence of gender and age on vulnerability to infection is unclear.Our data showed no significant differences in D. fragilis infection associated with gender or age.Nevertheless, the highest rate of D. fragilis (30%) was detected in participants 60-79 years.These results may be related to the limited study population, particularly of older participants, or might reflect a correlation of age with D. fragilis infection.A more comprehensive study with a broad age distribution is needed to resolve this issue.These limitations aside, this finding is similar to studies showing trends of higher infection rates in adults [9,14,37] and in contrast to some reports suggesting that children are common D. fragilis carriers [9,[38][39][40].Other studies have shown no influence of gender or age on rates of D. fragilis infection [10,27].As in most gastrointestinal infections, direct exposure to the parasite may play a crucial role.Therefore, it is probable that infection by D. fragilis is related to poor hygiene regardless of gender or age.
We found high overall prevalence of intestinal parasites (39%), including D. fragilis, Blastocystis sp., G. lamblia, E. coli, and I. butschlii and their coinfections.The most frequently detected parasite was Blastocystis (31.5%) followed by D. fragilis (13.5%).The majority of D. fragilis-positive individuals showed co-infection with other parasites, most frequently Blastocystis.Co-infection of D. fragilis with other enteric protozoa, especially Blastocystis, has been widely reported [10,26,41] and could support the hypothesis of direct transmission of D. fragilis through the fecal-oral route [2,6,19,42].Neither ova nor larvae of helminths were observed in the examined stool samples using the formalin-ether method, reflecting the decreasing incidence and prevalence of intestinal helminth infections in Iran during past two decades [43].
The presented data showed no significant relationship between infection with D. fragilis and clinical symptoms or reason for referral.Many studies have shown correlation of infection with D. fragilis and clinical symptoms [3,8,9,13,39], while others report no relationship between symptomatic infection and this parasite [44][45][46].This disparity is not surprising, as manifestations ranging from subclinical to severe gastrointestinal symptoms is typically observed in parasitic enteropathogen infections.This phenomenon is suggested to be related to genetic diversity in D. fragilis, resulting in a heterogeneous species [2,4,8,47,48].
Currently, two genotypes are described for D. fragilis, with genotype 1 being the most common [48].The investigation of genetic variation in D. fragilis SSU rRNA with respect to geographic area has shown that SSU rRNA gene variation is not sufficient to be used as an epidemiological marker [4,47,48].The SSU rRNA gene sequencing analysis of three D. fragilis isolates were similar and revealed 98% identity between our isolates and two corresponding published reference sequences for D. fragilis (accession nos.AY730405.1 and FJ649228.1).These results indicated low level of polymorphism, in agreement with recent studies [47,49].

Conclusions
This study demonstrated high prevalence of D. fragilis in Tehran via laboratory fecal analysis.Hence, clinical diagnostic laboratories should include screening for this parasite in routine stool examination.The PCR assay targeting the 5.8S rRNA gene detected a significantly greater number of D. fragilis-infected patients than did other analyses and is recommended as an effective tool for the accurate diagnosis of D. fragilis that should be employed in combination with microscopic methods to obtain a complete assessment of intestinal parasite infection.The use of these methods will prevent a high number of undiagnosed infections.Therefore, further studies applying this method to obtain accurate data on the prevalence of infection in specific age groups, symptomatic and asymptomatic individuals, other animals, and possibly a population-wide study, are required to ascertain epidemiology, pathogenicity, and transmission routes, as well as to identify reservoirs of D. fragilis.

Figure 2 .
Figure 2. Phylogenetic tree of D. fragilis genotypes constructed by neighbor-joining analysis, based on small subunit ribosomal DNA (SSU rDNA) sequences retrieved from this study (AB692771-3) compared with D. fragilis genotype 1 (AY730405.1),D. fragilis genotype 2 (U37461.1),and Tritrichomonas foetus (M81842.1)from Genbank.Bootstrap values obtained from 1000 replicates are indicated on branches in percentage.The length of the scale bar is equivalent to a sequence difference of 1%.The evolutionary distances were computed using the Kimura 2-parameter method and are expressed as the number of base substitutions per site.There were 805 positions in the final dataset.Evolutionary analyses were conducted in MEGA7.

Table 1 .
Characteristics and clinical features of participants positive and negative for D. fragilis Participants

Table 2 .
Number positive and prevalence (%) of intestinal parasites in participants referred to the clinical laboratory of Milad Hospital, Tehran, from March to July 2011.
a CI, Confidence Intervals; ND, not done; Infection percentages do not total 100% as some participants had multiple infections.

Table 3 .
Number of single and multiple infections in participants referred to the clinical laboratory of Milad Hospital Tehran, from March to July 2011.
a double infection of Dientamoeba with Blastocystis; b triple infection of Dientamoeba and Blastocystis with Giardia or E. coli or Iodamoeba; c double infection of Blastocystis with Dientamoeba or E. coli or Iodamoeba; d triple infection of Giardia, Dientamoeba and Blastocystis.