The principle of the cycling probe technology. F indicates fluorescence; Q, quencher.
Itahashi M, Higaki S, Fukuda M, Shimomura Y. Detection and Quantification of Pathogenic Bacteria and Fungi Using Real-Time Polymerase Chain Reaction by Cycling Probe in Patients With Corneal Ulcer. Arch Ophthalmol. 2010;128(5):535-540. doi:10.1001/archophthalmol.2010.66
To detect and quantitate the causative pathogens in patients with corneal ulcer using real-time polymerase chain reaction (PCR) by cycling probe.
Clinical and laboratory study of 40 eyes of 40 patients diagnosed with corneal ulcer. Two methods were used for pathogen detection: bacterial culture and real-time PCR with the patient's corneal scrapings. Probes and primers of real-time PCR were designed to be pathogen specific for simultaneous detection of Staphylococcus aureus, Staphylococcus pneumoniae, Pseudomonas aeruginosa, methicillin-resistant S aureus, Candida species, and Fusarium species. Results by both methods were evaluated and compared.
Of 40 eyes, 20 eyes had the same pathogens detected by both methods and those were S aureus (3 eyes; mean [SE], 3.8 [1.3] × 101 copies/sample), S pneumoniae (5 eyes; mean [SE], 5.6 [5.1] × 103 copies/sample), P aeruginosa (8 eyes; 5.1 [4.0] × 103 copies/sample), methicillin-resistant S aureus (1 eye; 1.0 × 102 copies/sample), and Candida species (3 eyes; mean [SE], 8.8 [4.9] × 103 copies/sample). Six eyes showed negative results by both methods. Results of both methods disagreed in 14 eyes; specifically, 11 had positive PCR results only, 2 had positive culture results only, and 1 eye had positive results for different pathogens.
The real-time PCR assay can simultaneously detect and quantitate bacterial and fungal pathogens in patients with corneal ulcer. Real-time PCR can be a fast diagnostic tool and may be useful as an adjunct to identify potential pathogens.
Corneal ulcer, including bacterial keratitis, fungal keratitis, and Acanthamoeba keratitis, can cause corneal opacity, deteriorated visual acuity, or even lead to some lifelong complications. Bacterial culture and smear examination using corneal scrapings is the conventional method to detect causative pathogens of corneal ulcer. However, bacterial culture is time-consuming and results of smear examination depend on the laboratory technician's skill. Therefore, a fast and accurate diagnostic method is highly desirable.
In recent years, polymerase chain reaction (PCR) has been widely used for clinical bacterium-1- 6 and virus-specific7- 10 detection and various technologies11- 18 have been developed for the PCR assay. Multiplex PCR11 with multiple primers and real-time PCR with linear probe,12 structured probe,13 or cycling probe14- 18 (Cycleave PCR; Takara Bio Inc, Shiga, Japan) are such examples.
Cycleave PCR uses a chimeric DNA-RNA-DNA probe with a strand length of 10 to 14 bases (Figure).14- 18 When the probe hybridizes to its complementary target DNA, RNase H cleaves the probe at the RNA linkage and this allows emission of strong fluorescence. By measuring the intensity of the emitted fluorescence, the amount of the amplified product can be measured (Figure). As compared with linear probe or structured probe with longer-length probes, Cycleave PCR is highly specific with its cycling probe.14- 18
In addition to real-time PCR, DNA sequencer is another method to detect causative pathogens.19,20 However, this method is time-consuming and complicated and thus less efficient than real-time PCR.
Despite its increasing popularity, PCR has many limitations and it is still quite a challenge to accurately identify the causative pathogens of corneal ulcers by PCR.5,6 Though real-time PCR12- 18 is not generally available to most ophthalmologists, this technique can be useful in identifying the causative pathogens. The utility of DNA copy number in real-time PCR enables clinicians to investigate how DNA copy number has varied along the clinical course. It also helps to distinguish the ocular flora from the causative pathogens. Thus, appropriate management of corneal ulcers can be achieved in a prompt fashion. We have previously succeeded in using real-time PCR to detect herpes simplex virus in patients with herpetic keratitis.8- 10 However, to our knowledge, real-time PCR has not been used for the diagnosis of bacterial and fungal keratitis. We therefore used 2 methods, microbial culture and Cycleave PCR, to detect and quantitate the causative pathogens in bacterial and fungal ulcers. We assessed and compared the performance of both methods and aimed to determine the clinical potential of Cycleave PCR as a diagnostic tool for corneal ulcer.
We examined 40 eyes of 40 patients (mean age, 51.6 years; range, 16-89 years) who were diagnosed with corneal ulcer between November 2006 and January 2009. All the patients except 1 (case 10) were pretreated with antibiotics before coming to our clinic, Kinki University Hospital. This study adhered to the tenets of the Declaration of Helsinki. All the patients agreed to participate in this study and informed consent was obtained.
Subjects were under local anesthesia by oxybuprocaine eye drops, 0.4%, and an eye speculum was put on before sampling. Two samples were collected from each patient for culture, smear examination, and real-time PCR by corneal scraping. All sterile precautions were taken to avoid contamination during sample colletion.6 A stainless-steel blade (disposable scalpel No. 15; Feather Safety Razor Co LTD, Gifu, Japan) was used to scrape the lesion. One of the samples was transferred onto the culture media and slide glass with a cotton swab. The smear on the slide glass was examined by gram staining at the Department of Bacteriologic Examination, Kinki University Hospital.
The samples were cultured on blood agar medium (Nissui Pharmaceutical Co Ltd, Tokyo, Japan), chocolate agar medium (Biomerieux Japan, Tokyo), or bromthymol blue glucose agar medium (Biomerieux Japan). If fungal species including Candida species and Fusarium species were suspected, Candida medium EX (Nissui Pharmaceutical Co, Ltd) and Sabouraud agar were used, respectively.
The other corneal scraping for real-time PCR was placed into a sterile microcentrifuge tube that included 500 μL of saline solution. DNA was extracted from a corneal scraping specimen using EXTARGEN2 (Tosoh, Tokyo) according to the manufacturer's protocol. Briefly, we first added 100 μL of reagent and 2 μL of detergent for DNA coprecipitation in the tube, and the mixture was vortexed for 5 seconds. Subsequently, 500 μL of protein-denaturing detergent containing 60% (volume to volume ratio) isopropanol was added and the solution was vortexed again for 10 seconds, followed by a 3-minute centrifugation at a speed of 6000 × g. After the supernatant was removed, the precipitate was added to 200 μL of 40% (volume to volume ratio) isopropanol and was vortexed and centrifuged as described earlier. It was then added to 500 μL of 70% ethanol and was vortexed and centrifuged again. Finally, after adding DNase- and RNase-free water to the harvested DNA pellet, the DNA sample was prepared. With this kit, the DNA extraction was completed in approximately 20 minutes.
The primers and probes used for Cycleave PCR in this study were designed to simultaneously detect the following 6 pathogens: Staphylococcus aureus, Streptococcus pneumoniae, Pseudomonas aeruginosa, methicillin-resistant S aureus, Candida species, and Fusarium species (Table 1). The selected target genes for the specific pathogens were Cap5G (GeneBank U81973) and Cap8G (GeneBank U73374) for S aureus,21LytA (Takara Bio Inc) for S pneumoniae,gyrB (GeneBank EF064840.1) for P aeruginosa,22mecA (GeneBank X52593) for methicillin-resistant S aureus,23,24ITS2 (GeneBank AB032174) for Candida species,25 and EF1-alpha (GeneBank DQ247583) for Fusarium species.26 All the probes were labeled with a 6-carboxyfluorecein reporter dye at the 5′ end and with an Eclipse Quencher dye (Epoch Biosciences, Bothell, Washington) at the 3′ end. We bought the primers and probe for S pneumoniae from Takara Bio Inc. However, the company has not officially published the sequence of these primers and probe.
The real-time PCR assay used 2 × Cycleave PCR reaction mixture (Takara Bio Inc). The reaction mixture contained 20 μL of PCR Master Mix (Takara Bio Inc), 20μM primer, 5μM probe, and 5 μL of DNA sample. The final volume of the reaction mixture was adjusted to 25 μL. The PCR assay was carried out on an ABI PRISM 7000 Sequence Detector (Applied Biosystems, Foster City, California). The cycling conditions were 10 seconds at 95°C followed by 40 cycles of 10 seconds at 95°C, 10 seconds at 55°C, and 31 seconds at 72°C. The PCR assay was finished in about 2 hours.
The sensitivities for the 6 target pathogens were analyzed using the stock cultures from our laboratory. A standard curve of threshold cycle with values obtained using positive controls was constructed with an established linear range from 1.0 × 101 to 1.0 × 106 copies per 25 μL of reaction tube (Table 2). Quantification of serially diluted DNA sample was performed by ABI PRISM Sequence Detector software version 1.0 (Applied Biosystems).
Culture results after 48 hours were compared with the detected pathogens and DNA copy number obtained by real-time PCR.
Of the 40 eyes, Cycleave PCR and culture showed positive results in 32 eyes (80.0%) and 23 eyes (57.5%), respectively (Tables 3, 4, and 5). Results of both methods agreed in 26 eyes (20 eyes with matched pathogens and 6 eyes with negative results) (Tables 3 and 4) and disagreed in 14 eyes (11 eyes, positive PCR results only; 2 eyes, positive culture results only; and 1 eye, positive results for different pathogens) (Table 5). In the 2 eyes that had positive culture results only (cases 22 and 23) (Table 5), the detected pathogens (Staphylococcus warneri and Corynebacterium species) could not be detected by the real-time PCR assay because they were not included in the 6 target pathogens designed for the pathogen-specific probes and primers in this PCR study.
Among the matched pathogens in those 20 eyes, P aeruginosa was the most frequent pathogen (8 eyes; mean [SE], 5.1 [4.0]× 103 copies/sample) and was followed by S pneumoniae (5 eyes; mean [SE], 5.6 [5.1] × 103 copies/sample), Candida species (3 eyes; mean [SE], 8.8 [4.9] × 103 copies/sample), S aureus (3 eyes; mean [SE], 3.8 [1.3] × 101 copies/sample), and methicillin-resistant S aureus (1 eye; 1.0 × 102 copies/sample) (Table 4).
Of the 40 eyes, results of smear examination were positive in 8 eyes (20.0%) (Tables 4 and 5). This rate was low as compared with other reports,5,27 because we dipped the swab into the culture media and applied it on the slide glass for smear examination.
By comparing the sensitivity and specificity of real-time PCR with those of bacterial and fungal culture, the relative sensitivities and specificities5 of the real-time PCR assay for the 6 target pathogens were determined. A sensitivity of 100% was achieved for 5 pathogens of 6 and the relative specificities ranged from 89.7% to 100% (Table 6).
The current results obtained by culture method and real-time PCR coincided in 26 (65.0%) of the 40 eyes. Of those, 20 eyes had matched pathogens (Table 4) and the other 6 eyes showed negative results (Table 3) by both methods.
Accurate identification of causative organisms is essential to better management of corneal ulcer. A national surveillance27 of infectious keratitis in Japan reported that 90% of cases of infectious keratitis are caused by bacteria and 10% are by fungi or Acanthamoeba. Approximately 50% of the bacterial pathogens are gram-positive cocci, such as the frequently occurring S aureus, Staphylococcus epidermidis, and Streptococcus, whereas P aeruginosa is the most prevalent pathogen among gram-negative bacilli, including Moraxella species, Acinetobacter species, and several other bacteria.27 Furthermore, the US Centers for Disease Control and Prevention has issued a warning about increasing cases of Fusarium keratitis.28 Multidrug-resistant pathogens and fungi have also been increasing because of improper use of antibiotics and steroids.28 Based on the surveillance result, we designed the real-time PCR assay in this study for the 6 pathogens that are commonly detected in clinical ophthalmology and succeeded in simultaneous detection and quantification within 2 hours.
In our study, P aeruginosa was the most frequently detected pathogen and S aureus was the second. These results were a little different from those of other studies.5,27 We think this is because almost all our cases were referrals from other clinics and many cases had been already treated by some antibiotics when they came to our hospital.
While the pathogen in 1 of the 3 cases with discrepant findings (cases 21-23) (Table 5) was thought to be Acanthamoeba (case 21) (Table 5), the other 2 cases (cases 22-23) were cured by multidrug therapy and the causative pathogen could not be verified. Eleven cases had positive PCR results only (2 fungal and 9 bacterial pathogens) (Table 5). Although PCR has a high risk of false positivity,5,6 we actually treated the patients with positive PCR results only according to their real-time PCR results and the treatment outcomes were all satisfactory. This may demonstrate a better detection sensitivity in the PCR assay. Vengayil and colleagues6 compared the results of PCR and the conventional microbiological techniques (gram smear, culture, and potassium hydroxide wet mount). They concluded that PCR is more efficient and sensitive than the conventional microbiological techniques in diagnosing fungal keratitis. Our current results of the 2 cases with positive PCR fungal pathogen results only agreed with their findings. In cases of patients pretreated with antibiotic therapy such as quinolone, the conventional culture is more likely to give false-negative results whereas real-time PCR may still show positivity. However, the high sensitivity of real-time PCR could lead to false-positive results. These can be caused by laboratory contamination from reagents and intrasample contamination.6 In conclusion, not only culture results but also clinical symptoms, PCR findings, and conventional smear results should all be carefully considered to accurately determine the causative pathogen. Real-time PCR can be a fast diagnostic tool and may be useful as an adjunct to identify potential pathogens.6
Real-time PCR obtained high sensitivities (except Fusarium species) and specificities (89.7%-100%) for the 6 target pathogens against the gold standard culture technique (Table 6). A previous PCR study using a structured probe reported a sensitivity of 96.2% and a specificity of 93.2% for S pneumoniae,13 and our results for S pneumoniae (100% and 89.7%) (Table 6) were comparable with the previous results. Kim and colleagues5 reported that PCR is more sensitive to fungal than to bacterial pathogens. Because we had a small number of fungal cases in this study, a future study with a large number of bacterial and fungal cases can better validate the feasibility of the real-time PCR assay.
In conclusion, though the numbers included in this study were limited, particularly with fungal ulcers, we have demonstrated that real-time PCR can accurately and simultaneously detect bacterial and fungal pathogens in a speedy fashion. We targeted 6 pathogens this time. If the testing durations for the selected target pathogens are about the same, we suspect that theoretically it might be possible to simultaneously detect many bacterial and fungal pathogens in a single run. In the future, applications of the real-time PCR assay to more pathogens are of interest. We have already accomplished real-time PCR system for Acanthamoeba and Aspergillus species. With real-time PCR, it may be possible to develop a diagnostic kit for pathogen-specific detection in the busy ophthalmic clinical practice.
Correspondence: Shiro Higaki, MD, PhD, Department of Ophthalmology, Kinki University School of Medicine, 377-2 Ohno-Higashi, Osaka-Sayama 589-8511, Japan (firstname.lastname@example.org).
Submitted for Publication: June 18, 2009; final revision received August 25, 2009; accepted September 4, 2009.
Financial Disclosure: None reported.
Additional Contributions: Mayumi Mizuno provided technical assistance.