Material and methods

Strains of Acanthamoeba

The ATCC’s available strains were purchased from its correspondent in France, LGC Prochem. Tubes were labelled as: A comandoni (ref 30135); A palestinensis (ref 50708); A sp (ref 50655), A palestinensis (ref 30870); A griffini (ref 30731); A castellani (ref 50373); A lenticulata (ref 30841); A astronyxis (ref 30137); A tubiashi (ref 30867) and A hatchetti. Acanthamoeba were cultured and maintained according to the specific procedures indicated by the ATCC for each strain. The viability of parasites, the number of trophozoits, the cyst load and the cyst morphology were assessed daily by direct microscopic examination. Quantified viable cultures were processed for DNA extraction.

AK samples

Samples were obtained in accordance with the Declaration of Helsinki (http://www.wma.net/e/policy/) from patients with suspected AK and from patients presenting with keratitis requiring microbiological diagnosis by ophthalmologists under a slit lamp by deep scraping of the cornea with sterile stainless steel blades20 and sent to the laboratory for conventional culture and real-time PCR. Aliquots were frozen and tested blind after the addition of 200 μl of sterile phosphate buffer solution (PBS). The bacterial, fungal or viral strains from clinical samples were isolated and characterised by routine classic culture tests.

Primers and probe

The sequences were obtained from the 41 591 bp DNA Acanthamoeba castellanii mitochondrion gene (gi|562028|gb|U12386.1|ACU12386, complete genome Accession U12386, U12386.1GI:562028; http://www.ncbi.nlm.nih.gov). The forward primer was 5′GCAGTCGCGGTAATA CGA; the reverse primer was: 5′ACCACCTACGCACCCTTTACA; and the TaqMan probe was: 6-FAM-AGTGT TATTCGCATTGACTGGGTGTAA-TAMRA.

DNA extraction

Five microlitres of a whole virus preparation of Seal Herpesvirus (phHV) (a gift from van Doornum, Department of Virology Erasmus MC, Rotterdam, The Netherlands) was added to 200 μl of each cyst suspension before extraction (final concentration of 1000 to 2000 viral particles/ml) to monitor the extraction processes and to assess the potential PCR inhibitors.202122 The suspensions were mixed with 200 μl of ATL buffer (Qiagen, Courtaboeuf, France) and 40 μl of Proteinase K (Qiagen) and incubated at 56°C for 10 min. Proteolysis was stopped in a dry plate at 95°C for 10 min, and the DNA was extracted with the MagNA Pure Nucleic Acid isolation kit (MagNA Pure LC, Roche, Mannheim, Germany) and eluted in 100 μl.22 The extract were cooled immediately after (−20°C) for 5 min, and the temperature was stabilised for 30 min at 20°C.

Real-time TaqMan PCR

In preliminary assays, the optimal conditions (primers and probes, volumes and temperatures) were determined in which at least the equivalent of 0.1 cyst/μl of Acanthamoeba genotype T 4 were repeatedly detected. For the duplex fast real-time PCR (f-d-real-t PCR), the differences in emission wavelength spectra for each fluorophore binding to probes were >15 nm. The f-d-real-t PCR was carried out in a final volume of 25 μl containing 2X TaqMan Fast Universal PCR Mastermix (2X, Applied Biosystems, Foster City, California; AB Ref. 4352042), Acanthamoeba forward (0.5 μM) and reverse primers (0.5 μM), Acanthamoeba FAM-TAMRA probe (0.5 μM), PhHV-forward (0.5 μM) and reverse primers (0.5 μM), PhHV-VIC-TAMRA probe (0.5 μM) and 12.5 μl of the DNA eluted in distilled water. The amplification and detection were carried out with the ABI Prism 7500 sequence detector, and the cycling programme consisted of one cycle at 95°C for 20 s and 45 cycles at 95°C for 3 s and 30 s at 60°C. Measurements were followed automatically, and the cycle threshold (Ct) value for each sample was determined according to the fluorescence signal exceeding the background limit of 0.07. Each run contained negative controls and controls with quantified number of cysts. Linearity, sensitivity, reproducibility and detection limits were assessed for all the strains tested pure and diluted in distilled water before and after DNA extraction. Retesting with the same extracts was conducted with the SmartCyclerII equipment in 50 two-step cycles of 10 s at 95°, and 65 s at 60°C after mixing 12.5 μl of extracted DNA with 12.5 μl of 2X TaqMan Universal PCR Master Mix (MNL 430449 AB) containing the primers and probes each at 0.5 μM.

Confirmation of negativity (absence of Acanthamoeba) by real-time SYBR Green PCR

SYBR Green (2-[N-(3-dimethylaminopropyl)-N-propylamino]-4-[2, 3-dihydro-3-methyl-(benzo-1, 3-thiazol-2-yl)-methylidene]-1-phenyl-quinolinium) is an intercalating dye which fluoresces only when bound to double-stranded products generated by PCR. The f-d-real-t PCR is associated with SYBR Green reactions carried out in a second tube after mixing 5 μl of the DNA extract with 12.5 μl of the Fast Start Universal SYBR Green Master Mix (AB 04 913850001), 0.25 μl of the forward and reverse primers (final concentration of each 300 nmol), 1 unit of UNG (AB UNG Ref: 03539 806001) and 9.5 μl of sterile distilled water (one cycle at 50°C for 20 s, one cycle at 95°C 20 s and 45 amplification cycles at 95°C for 3 s and at 60°C for 30 s; the positive threshold was fixed at 100). In every series of testing, SYBR Green results were validated only if the DNA extracts from the reactants and the environment yielded negative results.

Specificity and interferences of bacteria and fungi

The potential interferences on the f-d-real-t PCR were studied by mixing before extraction 20 cysts of the strain T4 with 200 μl of suspensions containing 10⁶ CFU/ml of bacteria or fungi or 10⁶ PFU/ml of three different Herpesviridae isolated from corneal ulcers.

Comparison with other real-time PCR

Aliquots from the DNA extracts containing the IC obtained from clinical samples were tested simultaneously according to the primers, probes and experimental conditions for real-time PCR described by the authors.16171819

Results

The new f-d-real-t PCR detects and quantifies Acanthamoeba in one run and simultaneously assesses the DNA extraction yields and PCR inhibitors. The detection capacities of the f-d-real-t PCR (dilution vs the last positive signal) are identical using the ABI PRISM 7500 or the SmartCyclerII, and both detected at least 0.1 cyst/μl (the equivalent of one cyst or less in the original specimen). However, the signals triggered using TaqMan probes for A astronyxis were repeatedly inferior to those obtained for all the other strains (table 1) with detection limits 1 log inferior (1 cyst/μl). Nevertheless, using the same set of primers and SYBR Green, it was possible to detect at least 0.1 cyst/μl, suggesting a reduction in the annealing capacities of the of A astronyxis probe targeted sequence. On the above, for the confirmation of the absence of Acanthamoebae in a sample, the f-d-real-t PCR should be systematically associated with an SYBR Green PCR (using 5 μl of the same DNA extract) before the absence of Acanthamoeba can be stated.

Table 2 presents the results obtained after mixing Acanthamoeba (20 cysts/sample) with highly concentrated bacterial, fungal or viral suspensions. The pathogens isolated from corneal infections did not alter the detection capacities of the f-d-real-t PCR. The Ct values obtained for the controls for each sample compared with the blanks were never delayed for more than 1.5 cycles, indicating high and reproducible DNA recovery yields after the preproteolysis.22 All samples testing f-d-real-t PCR negative were SYBR Green-negative, and all negative controls from air and reactants were negative by both methods. For all the samples, the dispersion of the Cts (intra and inter assays) for the targeted sequences and for the ICs was <1.5 cycles. The SYBR Green PCR always detected 0.1 cyst/μl or less. During the preliminary SYBR Green validation procedures, highly concentrated suspensions of Candida (C) albicans, C parapsilopsis, C glabrata, C krusei and Geotricum candidum (but not the culture media) isolated from corneal lesions triggered weak signals that could have been misinterpreted. However, when the samples containing bacteria, Acanthamoeba or fungi were tested after presetting for the two equipments tested here, the threshold value at 100, only the samples containing Acanthamoeba were positive.

The results obtained with 20 samples obtained from patients with signs evoking AK tested (a) by classic routine diagnosis techniques (direct microscopic examination after Giemsa and Gram staining and culture for aerobic, anaerobic bacteria, fungi and Acanthamoeba), (b) by the real-time PCR16171819 and (c) by the new f-d-real-t PCR are presented in table 3. The direct microscopic examination detected positive 60% of the samples and the culture of Acanthamoeba from corneal scrapings positively detected seven out 10 samples (70%). The potential PCR inhibitors were eliminated (reproducible for the ICs), and the isolates genotypically characterised belong to genotypes T 4 (70%), T 3; T5 and T 11 (10% each). One hundred per cent of the positive f-d-real-t PCR results were confirmed by SYBR Green. In three different experiments, the dispersion of the Cts (intra- and interassays) for the targeted sequences and for the ICs were <1.5 cycles for all the samples The real-time PCR using primers and probes bracketing other regions of the genome16171819 detected eight out 10 samples as positive, but the new f-d-real-t PCR detected 10 out 10, and all were confirmed by SYBR Green and genotypically characterised. Moreover, the false-positive results obtained with the real-time PCR for samples containing Salmonella lexinton, Clostridium sporulans, Serratia marcescens and Propionibacterium acnes (Cts >37) were repeatedly negative with the f-d-real-t PCR and confirmed as negative by SYBR Green.

Discussion

A new test was developed for the diagnosis of Acanthamoeba. The TaqMan probe selected does not cross-react with prokaryote or eukaryote genome sequences, and as predicted by the comprehensive in silico blast analysis, the suspensions of Acanthamoeba mixed with other infectious agents do not produce false-positive or negative results. To minimise misdiagnosis for samples containing low loads of A astronyxis, the SYBR Green procedure was systematically carried out. The SYBR Green may detect traces of residual double-stranded DNA, and so we limited its use for the confirmation of the absence of Acanthamoeba.

The culture methods may produce false-negative results due to low parasitic load, small sample volumes and the use of antiseptic or antibiotics prior to sampling. Nevertheless, NAATs are independent of the growth of the micro-organisms and may detect low amounts of injured Acanthamoeba, even those suspended in antiseptics.91113142324 Several PCRs and real-time PCRs have been described for diagnosis of Acanthamoeba, but none reported their abilities to detect the species available from the ATCC.1617181924 In addition, the sensitivity of real-time PCR was generally assessed testing the genotype T4 (the most frequently detected in corneal samples), and as shown in the present study these tests do not detect all the genotypes associated with AK. Moreover, the specificity of the f-d-real-t PCR is higher compared with the other real-t PCRs171819 especially for samples that contain one of the four bacteria that triggered positive signals. These false-positive conclusions produced by the real-t PCR were confirmed by an additional in silico testing showing partial homologies between the Acanthamoeba sequences previously selected for the real-time PCR16171819 and the genomes of bacterial species (NCBI Blastn FAQs, ncbi.nlm.nih.gov.Genbank).

Conclusion

The methods previously described for diagnosis of Acanthamoeba do not fulfil the expectancies for sensitive and specificity. None of them reported the capacity for detecting low levels of the ATCC strains of Acanthamoeba associated with pathology. The f-d-real-t PCR diagnosis strategy (f-d-real-t PCR coupled to SYBR Green) allows results to be obtained within 2 h with minimal risk of false-negative diagnosis for A astronyxis. On 20 clinical samples, the sensitivity and specificity of the f-d-real-t PCR are higher than those of the real-t PCR previous described.16171819

Timely interventions inpatients with AK improve the prognosis. Nevertheless, therapeutic management of AK is based on long-lasting topical solutions administered repeatedly—at the beginning during day and night—and most of the biocidal agents are toxic for the epithelium. For the above, false-positive diagnosis may induce unnecessary iatrogenic toxic effects on the eye surface.249101124 In summary, the new real f-d-real-t PCR targeting sequences shared by all the Acanthamoeba available from the ATCC appears as a specific, sensitive, reproducible, quantitative and easy-to-perform tool. The diagnosis strategy can be adapted to the different real-time PCR platforms and produces results in less than 2 h with a significant low cost of reactants. To our knowledge, the f-d-real-t PCR is the first NAAT detecting and quantifying simultaneously all of the ATCC Acanthamoeba strains (including the genotype T5 and T11) that simultaneously assesses the DNA extraction yields and the presence of PCR inhibitors. Larger trials are necessary to confirm the usefulness of the f-d-real-t PCR for disease management and environmental studies.