Confirmation of Confocal Microscopy Diagnosis of Acanthamoeba Keratitis Using Polymerase Chain Reaction Analysis

Background  Acanthamoeba keratitis has commonly been identified with in vivo confocal microscopy and confirmed with histologic examination of an epithelial biopsy specimen.

Objective  To determine if Acanthamoeba keratitis can be verified using polymerase chain reaction (PCR) of epithelial biopsy specimens.

Methods  Epithelial specimens from patients with suspected Acanthamoeba keratitis by confocal microscopy were tested for Acanthamoeba with PCR of Acanthamoeba ribosomal DNA.

Results  Twenty-four of 31 patients with evidence of Acanthamoeba keratitis were positive for Acanthamoeba on PCR analysis using 3 sets of primers. In 22 cases, the sequence obtained closely matched Acanthamoeba castellanii.

Conclusions  This study demonstrates that PCR analysis of epithelial biopsy specimens can provide definitive verification of the confocal microscopic and histologic identification of Acanthamoeba organisms associated with keratitis. Acanthamoeba keratitis is probably quite common, especially in contact lens wearers, although more than half of the patients in this study did not wear contact lenses.

ACANTHAMOEBA HAS been increasingly recognized as an ocular surface pathogen since Acanthamoeba keratitis was first diagnosed in 1974.¹ Originally associated primarily with contact lens usage, more recent data indicate that many cases are associated with other predisposing conditions.^2-5 It is frequently difficult to identify Acanthamoeba on the ocular surface since the presentation can mimic herpetic keratitis or bacterial keratitis.^4,6-8 This diagnostic confusion has often led to delays in making the diagnosis and, consequently, increased severity of Acanthamoeba keratitis when detected.^5,9-12Acanthamoeba may also be present concurrently as an opportunistic organism, particularly in patients with herpetic keratitis.^6,13

Most cases of Acanthamoeba keratitis have been diagnosed by epithelial biopsy for histologic analysis and culture.^5,8,12,14-16 Epithelial biopsy has largely replaced stromal biopsy as the method of choice to obtain specimens for culture or histologic examination. It is less destructive and excellent tissue preservation can be obtained with recent advances in collection techniques.¹⁷

Confocal microscopy has also been reported to be a useful method to diagnose noninvasive Acanthamoeba keratitis and to permit earlier diagnosis.^5,18-20 Although confocal microscopy lacks sufficient resolution to be the only method of diagnosis, it can be used to screen patients suspected of having Acanthamoeba keratitis. At our institution we used initial evidence of Acanthamoeba from confocal microscopy and confirmation by histologic examination of epithelial biopsy specimens to diagnose more than 200 cases of Acanthamoeba keratitis since we began using the confocal microscope in 1993.

Following widespread flooding in Iowa in July 1993, we noted an increase in the number of cases of Acanthamoeba keratitis, particularly from counties in Iowa that experienced contamination of their water supply.²¹ For 10 years prior to 1993 we averaged 2 cases per year, in 1993 we diagnosed 8 cases, and in 1994 we diagnosed 60 cases. This level of diagnosis has not declined to the preflood level and we continue to diagnose approximately 50 cases of Acanthamoeba keratitis per year at our institution. In 1993 and 1994, most of our cases of Acanthamoeba keratitis were quite severe, characterized by symptoms of pain and loss of vision. In the past several years, many patients were seen with only mild or moderate disease that was limited to the epithelium.

It has not been possible to determine whether this epidemic was caused by typical Acanthamoeba pathogens or whether many of these less severe cases were caused by other similar opportunistic organisms or less pathogenic strains of Acanthamoeba. Culture results have been inadequate; only 1 positive culture was obtained for the first 39 cases diagnosed. Other investigators have also reported low culture rates.^4,7,8,15,16 We have, therefore, sought other methods to detect and confirm the presence of Acanthamoeba on the ocular surface. Polymerase chain reaction (PCR) has become the standard method of analyzing small amounts of DNA and can easily be used to detect small numbers of bacteria and other pathogens.²² Polymerase chain reaction has recently been reported as a sensitive method of detecting Acanthamoeba cultured from the ocular surface.^23-27 It can also be used to detect Acanthamoeba taken from the cornea without an intermediate culture.²⁸ We tested a series of patients with Acanthamoeba keratitis diagnosed by confocal microscopy and epithelial biopsy to confirm the presence of Acanthamoeba organisms using PCR.

A series of 117 patients were seen in the Corneal Service of the University of Iowa, Iowa City, between March 1, 1998, and September 15, 1998, for evaluation of possible Acanthamoeba keratitis. Following their clinical eye examination, each patient was examined with confocal microscopy to assess whether Acanthamoeba organisms were visible in the epithelium and anterior stroma. In 31 patients, organisms suspected to be Acanthamoeba were detected and the patients underwent epithelial biopsy. In 2 patients the results of confocal microscopic examination were negative for Acanthamoeba and the patient still underwent epithelial biopsy. Twenty-eight of the 117 patients were evaluated as positive by confocal microscopic examination and did not undergo biopsy. These patients usually had a confirmed diagnosis of Acanthamoeba keratitis and were being followed up. The remaining 56 patients were evaluated as negative for Acanthamoeba by confocal microscopy and did not undergo epithelial biopsy.

The epithelium of each patient was removed around the area of the lesion. The epithelial specimen was divided into 2 parts: one for histologic evaluation and the other for analysis by PCR. The biopsy specimen for histologic examination was placed immediately in an alcohol-based fixative (Saccomanno; Star Lab, Lewisville, Tex) solution. Specimens were processed by cytospin centrifugation at 400 rpm for 10 minutes. Twelve slides were made of each specimen and alternating slides were stained with hematoxylin-eosin. The slides were read in masked fashion by our ophthalmic pathologist (R.F.) for the presence of Acanthamoeba organisms. Part of the biopsy specimen for PCR was transferred to a vial containing 100 µL of sterile water and immediately frozen.

Pure cultures of Acanthamoeba were enumerated using a Coulter counter and diluted in water. Samples (100 µL) of pure cultures and/or clinical samples were prepared. Ten microliters of 1-mol/L sodium hydroxide was added to give a final concentration of 0.1 mol/L. Samples were boiled for 3 to 5 minutes at 100°C, then extracted with 100 µL of phenol followed by 100 µL of phenol chloroform (1:1). DNA was precipitated by adding 10 L of 3-mol/L sodium acetate (pH 5.2), 250 µL of 100% ethanol, and 2 µL of 20-µg/µL glycogen as a carrier.

Samples were incubated overnight at −20°C and spun at 14,000 rpm to pellet the DNA. Pellets were washed with 500 µL of 70% ethanol followed by 500 µL of 100% ethanol. Pellets were dried and resuspended in 50 µL of nuclease-free water.

Acanthamoeba organisms were detected by PCR using specific primer pairs as described by Vodkin et al²³ and Lehmann et al.²⁸ Because of problems with these primers, we also developed and tested the samples with a new set of primers referred to herein as the Nelson primers:

Forward 5‘ GTT TGA ]GGC AAT AAC AGG T 3‘ Reverse 5‘ GAA TTC CTC GTT GAA GAT 3‘

Five-microliter aliquots of the solution containing the DNA were used to test each run. Polymerase chain reactions were performed using 100-µL reactions that consisted of 1× buffer containing 10-mol/L Tris-hydrochloride (pH 8.3) and 50-mol/L potassium chloride (Gene Amp PCR Buffer II; Perkin-Elmer, Norwalk, Conn). The reaction mixture also included magnesium chloride, 0.5 to 7.5 mmol/L; deoxynucleotriphosphates, 200 µmol/L; primers, 0.5 µmol/L; and Taq gold polymerase enzyme, 2.5 U.

The Taq enzyme was activated by a preincubation for 10 minutes at 95°C and DNA was amplified by 2-step cycling first for 1 minute at 94°C and then for 1 minute at 65°C for 50 to 55 cycles.

The PCR products were combined with 10× loading buffer and run on 2% agarose gels containing ethidium bromide for 1 hour at 300 V. Products were visualized with a UV transilluminator (Fotodyne Inc, Hartland, Wis) and photographed.

To sequence the positive PCR products, DNA was extracted from the agarose gel and sequenced using the dideoxynucleotide sequencing method from an automated sequencer (ABI TM 373; Applied Biosystems Inc, Foster City, Calif).

The results on 33 patients are included in this report. The confocal microscope was able to make a positive diagnosis of Acanthamoeba keratitis in 30 patients. Six of these were from patients previously diagnosed as having Acanthamoeba keratitis and 25 were new diagnoses over the 7-month period of the study. In 8 cases, the confocal microscopic image was not completely typical and the confocal microscopic diagnosis was questionable. Two patients were negative for Acanthamoeba by all 3 methods—confocal microscopic examination, epithelial histologic examination, and PCR analysis. One additional patient who was negative for Acanthamoeba by confocal microscopy and epithelial histologic examination was positive by PCR. The confocal microscopic appearance of the Acanthamoeba was typically a round or ovoid shape, 15 to 25 µm in diameter, and highly reflective. In some cases there was evidence of internal structure with vacuoles (Figure 1).

Histologic evidence of Acanthamoeba in the epithelial biopsy specimen was found in 13 (48%) of 27 cases with adequate specimens. Two of these were not substantiated by subsequent PCR. There was insufficient material to make a diagnosis in 5 cases. Fourteen cases were read as an adequate specimen but negative for Acanthamoeba. Macrophages were identified in one of these. In 1 case, no specimen was obtained for histologic examination.

Serial dilutions were made to the purified Acanthamoeba DNA, reducing the concentration by 1 log unit each time. Ten-microliter aliquots of each solution containing the DNA were used to test each run. The PCR product was detectable to a concentration of 5 × 10⁻⁴ ng/mL. Whole Acanthamoeba organisms in solution were also diluted using a Coulter counter to a concentration of less that 1 organism per 10-µL aliquot. The organisms were then prepared and the DNA extracted using the methods described above. We found detectable PCR product at a 90% detection rate for solutions containing approximately 1 to 5 organisms per 10 µL, indicating close to single organism detection sensitivity (Figure 2).

A summary of the results of the PCR patient data is presented in Table 1. Of the 31 patients with evidence of Acanthamoeba by confocal microscopy, 24 were positive by PCR in at least 1 of the 3 primer pairs (77%). One patient's results were positive with PCR only, the confocal microscopic and histologic results having been negative.

Twenty-six cases were tested with all 3 primer pairs. In this group, the Nelson primers were positive in 22 (85%), the Lehmann primers were positive in 14 (54%), and the Vodkin primers were positive in 14 (54%). All 3 primers were positive in 9 of 26 cases. In 37 cases PCR product was removed from the gel, sequenced, and compared with product from a standard A castellanii. Exact or very close sequence verification was obtain in 13 of 14 Vodkin PCR products, 1 of 1 Lehmann PCR product, and 19 of 21 Nelson PCR products. A total of 22 patients had Acanthamoeba confirmed by the sequence closely matching A castellanii. In an additional 2 of the Nelson PCR products the sequence was similar but not as close to A castellanii as the others (Figure 3).

The most common predisposing condition was the use of contact lenses (Table 1). Ten patients had a history of wearing soft contact lenses and 1 wore rigid gas-permeable contact lenses (35%). Seven of the 10 patients wore their contacts as extended wear (overnight). Five patients had an initial diagnosis of herpetic keratitis. Four patients had a history of trauma, 4 had previously undergone corneal transplantation, and 4 were diagnosed as having basement membrane dystrophy. The most common presenting sign was corneal ulcer (11 patients). Ten patients had epithelial microcysts.

Amplification of Acanthamoeba ribosomal DNA with PCR represents a new method for detection of the organism on the ocular surface that has wide clinical applicability. Because a single Acanthamoeba organism may contain more than 100 copies of ribosomal DNA, the detection of a very small number of organisms is possible.^29,30 Because this method is so sensitive, considerable care must be taken in the collection of the specimen and in the assay to avoid replicating extraneous sources of the DNA. This very high sensitivity is coupled with a variable specificity that is dependent on the primer selection.

We report these organisms as Acanthamoeba without any species identification. Classification systems for Acanthamoeba based on morphologic characteristics and enzyme analysis of cultured organisms have been the only standard available. New genetic information indicates that previous species classifications may not be correct compared with genetic differences.^24,26,31,32 Gene-based classification systems are made more difficult by the relatively high variability that has been found between Acanthamoeba of the same species (9.8%) and between species of Acanthamoeba (20%).^33,34 With such variability, primer selection becomes critical.

Most pathologic specimens of Acanthamoeba taken from patients with Acanthamoeba keratitis belong to group 2, based on the classification proposed by Pussard.³⁵Acanthamoeba castellanii is the most common pathogen in this group. There are, however, at least 7 species in addition to A castellanii that are known to cause keratitis.³⁶ Less closely related organisms such as Vahlkampfia, Hartmannella, and Naegleria also are known to cause keratitis.^24,37-42 The primers we used in this study produce positive results only for Acanthamoeba organisms. Other primers must be used for the detection of Vahlkampfia, Hartmannella, and Naegleria.^23,43

Several researchers have reported using specimens cultured from patients to test for the presence of Acanthamoeba with PCR. Because of the large number of organisms available from cultures, this method is a relatively easy way to confirm these organisms with PCR. Obtaining sufficient material for successful PCR results from epithelial biopsy specimens is more difficult.²⁸ The small number of organisms available from biopsy specimens has been suggested as a primary reason Acanthamoeba may be relatively difficult to culture from the eye. DNA of Acanthamoeba can also be successfully amplified with PCR from the tears of patients with Acanthamoeba keratitis.²⁸ Although this has a lower success rate than epithelial biopsy, it may be repeated frequently and is completely noninvasive thereby allowing the screening of a large number of patients.

Once Acanthamoeba have been detected with PCR, it is possible to use other methods to discriminate between species and classify each organism detected. Restriction fragment length polymorphisms has been successfully used for this, as has PCR and sequencing of longer DNA segments that can be matched for species identification.^34,44,45 Since we found some variability even within the small number of base pairs sequenced with the primers used in this study, such methods may be useful in determining if the organism(s) isolated from the ocular surface is a known pathogen or an opportunistic organism not contributing to the disease process.

Confocal microscopy is also a very sensitive and powerful technique for investigation of the ocular surface. It can detect single organisms and because it is a noninvasive and relatively rapid technique, the examination can be done repeatedly. The current level of resolution available with in vivo confocal microscopy limits its ability to make the definitive diagnosis. Although one can distinguish between inflammatory cells, fungus, and Acanthamoeba, we have found it is not possible to discriminate between macrophages and Acanthamoeba organisms.

During the past 10 years we have relied on histologic examination of epithelial biopsy specimens to make the definitive diagnosis of Acanthamoeba keratitis at our institution. This study also strongly supports the utility of this approach. Only 2 of 13 cases positively identified by histologic examination were negative by PCR analysis. Since the genetic variability of Acanthamoeba organisms is so great, even within species, it is very likely that these 2 false-positive results may have correctly identified Acanthamoeba that did not have the required DNA sequence to obtain PCR product with the primers we used.

Although this study demonstrates that PCR is probably more sensitive than histologic techniques, the extreme specificity of PCR renders some disadvantages that histologic techniques do not share. Examination of biopsy specimens with light microscopy reveals important information not obtainable with PCR, such as the cell type, severity of the inflammatory response, and the appearance of the Acanthamoeba organisms.

Previous reports in the literature stress the relationship between contact lens use and the development of Acanthamoeba keratitis. In our series of more than 200 cases between 1993 and 1998, approximately 40% have been associated with contact lens use. Most of the remaining patients have some predisposing condition, such as herpetic keratitis, basement membrane dystrophy, diabetic epitheliopathy, or ocular trauma, that allowed the organism to evade the eye's immune defenses. In the series reported herein, 35% of cases were associated with contact lens use. The frequency of Acanthamoeba keratitis as a complication of contact lens wear has yet to be determined. Our experience supports previous reports that the frequency of Acanthamoeba keratitis in contact lens wearers may be 1 per 10,000 per year or higher.^36,46Acanthamoeba may be responsible for a considerable percentage of cases commonly diagnosed as contact lens overwear. Our experience suggests that this is likely. The causal relationship between Helicobacter and gastritis may prove to be analogous to the releationship between Acanthamoeba and many forms of epitheliopathy.

Although 2 of our patients had severe Acanthamoeba keratitis and bacterial corneal ulcers, many patients in this report have relatively milder disease than that described in other large series. None of these patients has required corneal transplantation to date. We speculate that the relatively stable and mild acanthamoebic epitheliopathy we found with confocal microscopy and verified with PCR may represent a different strain of Acanthamoeba or even a different species than that which causes severe disease. Likewise, this epitheliopathy may represent a diminished immune response or inadequate defense system to nonpathogenic Acanthamoeba.

Accepted for publication September 4, 1999.

This research was supported in part by grant RO1EY11667-02 from the National Eye Institute, Bethesda, Md, Research to Prevent Blindness Inc, New York, NY, and Iowa Lions Club.

Reprints: William Mathers, MD, Casey Eye Institute, Oregon Health Sciences University, 3375 SW Terwilliger Blvd, Portland, OR 97201.