Positive control materials used in this study included axenic culture of ATCC strain 30459 E. histolytica HM-1:IMSS, purified Giardia lamblia cysts, and Cryptosporidium parvum oocysts (Waterborne Inc., New Orleans, LA). Control DNA was obtained from L3 larvae of S. stericoralis and Ancylostoma duodenale from coproculture. Control DNA for Necator americanus and A. lumbricoides was isolated from adult worms. Negative controls were water. For sensitivity testing, positive control materials were serially spiked into an identical stool sample from a healthy donor aliquoted into 200 mg volumes. We used an exogenous phocine herpes virus spiked into the lysis buffer for stool DNA extraction to provide a control for extraction and amplification.
One hundred twenty-nine fecal DNA specimens were obtained from preschool-age or younger children from the International Centre for Diarrhoeal Diseases and Research, Bangladesh (ICDDR,B). These samples were tested by saline wet mount microscopy per routine. Furthermore, 190 DNA specimens were obtained from the Leiden University Medical Center (LUMC) clinical microbiology laboratory and several epidemiological surveys in different countries from several collaborative projects with the Department of Parasitology at the LUMC. For this validation exercise at LUMC, positive samples were all previously extracted and tested for either protozoa or helminths by real-time PCR, thus we were not able to explicitly evaluate different DNA extraction methods in this work or test incoming specimens in real-time. All PCR-positive samples from LUMC were also positive by microscopy. Both positive and negative stool extracted DNA samples were included to establish sensitivity and specificity. Specifically, we sought to obtain at least 20 PCR-positive samples per organism with the exception of A. duodenale for which positive samples were limited. All protocols were approved by the Ethics Committees of ICDDR,B and the institutional review boards of University of Virginia (UVA) or Leiden University.
The DNA was extracted using a slightly modified QIAamp DNA Stool Mini Kit protocol (Qiagen Inc., Valencia, CA) as described previously for specimens from ICDDR,B.3
The DNA extraction for the specimens from LUMC was performed using regular QIAamp mini Kit spin columns with modifications, which included a pretreatment with PVPP (polyvinylpyrrolidone) and boiling.12
Protozoa-positive controls were extracted using an automated nucleic acid isolation system, QuickGene-810, with QuickGene DNA tissue kit S (Fujifilm, Tokyo, Japan). The manufacturer's protocol was modified to accommodate a larger input stool volume of 200 mg. Briefly, 1 mL of the tissue lysis buffer MDT was added to the stool, and the suspension was then pretreated by bead beating with a tube of 0.15 mm garnet beads (MO-BIO Laboratories, Inc, Carlsbad, CA) for 2 min followed by boiling for 7 min before extraction. Other modifications included the addition of 100 μL of EDT solution (Proteinase K) from the kit and a longer incubation time following the addition of EDT solution for 90 min. All DNA samples were stored at −80°C until use in PCR.
Primer and probe sequences3,4,13–17
and the real-time PCR protocols have been described previously (). Target genes included the 18S ribosomal RNA (rRNA) gene for E. histolytica
, G. lamblia
, and S. stercoralis
and COWP for Cryptosporidium
spp. (see GenBank accession nos. X64142, M54878, AF279916, and AF248743, respectively). Furthermore, the internal transcribed spacer (ITS) 1 target was used for A. lumbricoides
assay and the ITS2 for A. duodenale
and N. americanus
assays (GenBank accession nos. ALJ000895, AJ001594, and AJ001599, respectively).
Sequences for primers and probes used in the real-time polymerase chain reaction (PCR) and PCR-Luminex assays
Briefly, real-time PCR amplification reaction for the 3-plex protozoa was performed in 25 uL volume with 12.5 μL of iQ Supermix (Bio-Rad, Hercules, CA), which contains dNTPs, 6 mM MgCl2, 50 U/mL iTaq DNA, additional 2 mM MgCl2, 0.4 μM E. histolytica primers, 0.6 μM Giardia primers, 1.0 μM Cryptosporidium primers, 0.08 μM E. histolytica probe, 0.16 μM Giardia probe, 0.4 μM Cryptosporidium probe, and 4 μL of sample DNA. The Taqman probes used for this assay were labeled with Yakima Yellow for E. histolytica, FAM for Giardia, and Texas Red for Cryptosporidium and purchased from Eurogentec (Fremont, CA). The PCR cycling condition consisted of an initial 3 min 95°C cycle followed by 40 cycles of 30 sec at 95°C, 30 sec at 55°C, and 30 sec at 72°C with a final extension for 7 min at 72°C. Amplification reaction for the helminth real-time assay was performed in 25 μL volume consisting of 12.5 μL HotStarTaq Master Mix (Qiagen Inc.), which contains 1.5 mM MgCl2 and 200 μM each dNTP, additional 3.5 mM MgCl2 for a final concentration of 5 mM MgCl2, 0.1 mg/mL of BSA, 0.2 μM Ancylostoma primers, 0.1 μM Ancylostoma probe, 0.2 μM Necator primers, 0.05 μM Necator probe, 0.08 μM Ascaris primers, 0.05 μM Ascaris probe, 0.1 μM Strongyloides primers, 0.05 μM Strongyloides probe, 0.15 μM PhHV primers, 0.05 PhHV probe, and 5 μL of sample DNA. Taqman Fluorophores for the Taqman probes included Cy5, FAM, Quasar705, Texas Red, and Yakima Yellow for the PhHV, Necator, Strongyloides, Ancylostoma, and Ascaris probes, respectively. The helminth primers and probes were synthesized by Biolegio (Nijmegen, The Netherlands). Amplification steps were as follows: 15 min at 95°C; 50 cycles of 30 sec at 94°C, 30 sec at 60°C, and 1 min at 72°C; final extension for 10 min at 72°C. All real-time PCRs were performed either on the Bio-Rad iCycler or the Bio-Rad CFX platform (Bio-Rad, Hercules, CA). Real-time PCR Ct was analyzed by the iCycler software version 3.1 and CFX96 Real-Time System by CFX software version 1.1 (Bio-Rad, Hercules, CA).
Multiplex PCR for the Luminex assays were prepared exactly as the real-time mixture except without the TaqMan probes, and either the forward or the reverse primers were biotinylated on the 5′-end. Internal probes were amine modified at the 5′-end and included 12 carbon spacers. All probes and primers were purchased from Integrated DNA Technologies, Inc. (Coralville, IA). Probes were coupled to carboxylated Luminex beads with EDC (1-Ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride). Hybridization of beads to the PCR products was performed at 50°C for both the helminth and protozoa assays for 35 min using the Oligonucleotide Hybridization Protocol from Luminex Corporation.18
Detection of the amplicon was performed by the Bioplex 200 (Bio-Rad) or Luminex 100 (Luminex Corporation, Austin, TX) after addition of streptavidin PE as the reporter molecule. Negative control material (nuclease-free water) and positive controls were included with each run. All discrepant samples were retested by both modalities.
Analytical sensitivity for the Luminex 3-plex protozoa assay was determined by serially spiking C. parvum oocysts, G. lamblia cysts, and E. histolytica trophozoites into a single common stool specimen followed by DNA extraction with the Fujifilm method. For Ancylostoma, Ascaris, Necator, and Strongyloides, where cultured materials were not available, we used serial dilutions of control DNA to establish analytical sensitivity.
Luminex data were reported as mean fluorescence intensity (MFI). We calculated a corrected MFI (cMFI), which normalizes to background fluorescence as follows: cMFI = [MFI (sample) − MFI (background)]/MFI (background) to accommodate testing on different Luminex systems under both high and low photomultiplier tube voltages and different software (Star Station version 2.0, Applied Cytometry Systems, Sacramento, CA, and Bioplex software version 5.0, Bio-Rad, Hercules, CA).
The cycle threshold (Ct) and cMFI were compared using the Mann-Whitney test because all data sets were not normally distributed (SPSS Inc., Chicago, IL). To measure the degree of association between Ct values and cMFIs, Spearman's rank correlation test was used because the relationship between the variables was nonlinear. All P values were two-tailed. Data shown as mean + SEM unless otherwise indicated. We calculated 95% confidence intervals (CI) for sensitivity and specificity based on the sample size chosen for this study using the formula (CI = p ± 1.96 × [p(1 − p)/n]1/2)
where p = sensitivity or specificity and n = number of infected samples for sensitivity or uninfected samples for specificity determined by real-time PCR, the gold standard test.19