Molecular Screening of Donor Corneas for Fungi before Excision

Abstract

purpose. To develop panfungal and Candida albicans species–specific polymerase chain reaction (PCR) assays to screen donor eyes for fungal contamination before corneal excision.

methods. PCR primers were designed for either the broad-spectrum detection of fungal DNA or the specific detection of C. albicans DNA. Their sequences were based on rDNA regions highly conserved among and specific to fungi and C. albicans, respectively. PCR conditions with the two primer sets were optimized and tested for sensitivity using purified C. albicans genomic DNA and a plasmid containing the relevant region of C. albicans DNA. The specificity of the primer sets was established using higher eukaryotic, fungal, prokaryotic, and viral DNAs as PCR templates. Donor eye swab specimens were collected before corneal excision. DNA was extracted from the specimens and tested by both PCR assays.

results. The lower limit of detection for both primer sets was consistently 10³ genome equivalents, when using genomic DNA as a template and 10² copies of plasmid. The fungal PCR assay amplified DNA from all fungal species tested but did not amplify any of the selected mammalian, bacterial, or viral DNA. The C. albicans PCR detected the C. albicans DNA but was negative for all other DNA substrates, including the other fungal templates. Thirty-five percent of the donor eye samples tested were positive for fungus, and 19% were positive for C. albicans DNA.

conclusions. The PCR assays allowed the rapid screening of DNA extracted from specimens collected from corneal donors for potential fungal contamination. The assay was highly sensitive and specific for screening corneal surfaces. The results suggest that approximately one-third of donor eyes tested harbor fungi on the ocular surface.

More than 100,000 corneal transplantations are currently performed worldwide each year. ¹ A serious donor-related complication is postkeratoplasty endophthalmitis, with more than 70% of the fungal endophthalmitis cases being attributed to Candida albicans. ² The same organism is often isolated from the donor rim or the preservation medium. ² The prevalence of fungi in the normal conjunctival flora has been reported to be from 3% to 28% ^{3 4 5} and fungi have been isolated from 2.5% to 6.5% of donor corneoscleral rims. ⁶  

PCR technology has allowed the rapid detection of low copy numbers of DNA in clinical samples of small volume. The purpose of this study was to develop effective PCR-based techniques for the detection of a broad spectrum of fungal DNA and C. albicans DNA specifically from donor eye specimens. The development of these assays resulted in a highly sensitive and specific method for screening corneal donors and has the advantage of rapid detection of fungi from ocular surfaces compared with standard culture methods. Our results indicate a higher prevalence of fungi on the ocular surface of donor eyes than has been previously reported in eyes of healthy individuals. These findings may have implications for the monitoring of mycologic contamination of donor corneas before or during storage in corneal preservation medium. 

Materials and Methods

Donor Eye Swab Collection

Over a 3-month period, certified technicians at the Lions Eye Bank of Texas used sterile cotton applicators to swab cadaveric eyes within 24 hours of death (mean time of collection 12.5 hours). Specimens were collected from the lower conjunctival fornix of each eye before irrigation or corneal excision. The applicators were placed in 1 ml of sterile PBS solution and kept at 4°C until delivery to the Sid W. Richardson Ocular Microbiology Laboratory. The applicators were discarded, the samples centrifuged at 15,000g, and the pellets stored at −70°C. 

DNA Extraction

Candida albicans strain 32354 was obtained from the American Type Culture Collection (ATCC; Rockville, MD), and other control fungal and prokaryotic cultures and clinical samples were obtained from the Sid W. Richardson Ocular Microbiology Laboratory collection, Baylor College of Medicine (Houston, TX). 

Overnight cultures of C. albicans, Candida parapsilosis, Candida tropicalis, Candida guilliermondii, Candida lipolytica, and Saccharomyces cerevisiae at more than 10⁸ cells/ml were processed for DNA extraction by standard yeast genomic DNA methods, ⁷ with modifications. Briefly, 30-ml cultures were pelleted by centrifugation (10,000g), resuspended in 10 ml sorbitol-EDTA buffer, and incubated with 1.25 mg/ml Lyticase (Sigma, St. Louis, MO) for 3 hours to induce spheroplast formation. The spheroplasts were resuspended in 3 ml Tris-EDTA (TE) buffer and incubated with 10% SDS, and the proteins were precipitated with 5 M potassium acetate. The nucleic acid–containing supernatant was precipitated with 100% ethanol, resuspended in 4 ml TE buffer, and incubated in the presence of 1 mg/ml RNase A (Roche Diagnostics, Indianapolis, IN). The DNA was extracted three times with phenol-chloroform-isoamyl alcohol (25:24:1), ethanol precipitated, and resuspended in 1 ml TE. Aspergillus flavus was processed for DNA extraction by standard methods. ⁸ For the processing of the donor eye swab pellets and the C. albicans–spiked control samples, each pellet was resuspended in 1 ml PBS, and the standard yeast genomic preparation was used with the volumes of all reagents reduced proportionately. Glycogen was used as a carrier. ^{9 10 11}  

Human DNA was extracted from a human lung fibroblast cell line (MRC-5 cells, CCL-171; ATCC), and murine DNA was extracted from a mouse embryo fibroblast cell line (NIH 3T3 cells, CCL-1658; ATCC). The nucleic acids were extracted, and the DNA was isolated from confluent cell monolayers with an RNA extraction reagent (TRIzol; Gibco BRL, Gaithersburg, MD), as previously described. ^{9 10 11} Calf thymus DNA was obtained from Sigma. One to 3 μg of each DNA template, which corresponds to more than 10⁶ copies of genomic DNA, was used in the PCR reactions. 

Escherichia coli and Pseudomonas aeruginosa overnight cultures were processed for genomic DNA extraction by selective precipitation of polysaccharides with hexadecyltrimethyl ammonium bromide (CTAB). ⁷ Herpes simplex virus genomic DNA was extracted according to published methods. ¹² Purified adenovirus DNA was a generous gift from Ronald Javier, Department of Molecular Virology and Microbiology, Baylor College of Medicine. 

Cloning of C. albicans rDNA Region

Total genomic DNA was isolated from C. albicans strain 32354 and digested with EcoRI. The digested fragments were separated on a 0.7% agarose gel and transferred to a nylon membrane for Southern blot analysis. ⁷ A 200-bp double-stranded DNA probe was generated, by using C. albicans–specific primers in a PCR containingα ³²P-dATP. ¹³ An EcoRI fragment approximately 4 kb in size was identified, isolated, and used in a ligation reaction with an EcoRI-digested pUC19 plasmid. The clone pCA1 was identified by colony lift using the 200-bp probe. ⁷ The plasmid was purified with a kit (Concert Midiprep; Gibco BRL), and the concentration was adjusted to 1 × 10⁸ copies/μl. 

Polymerase Chain Reaction

PCR primers were designed for the detection of broad-spectrum fungal DNA (F-fwd and F-rev) or C. albicans–specific DNA (Ca-fwd and Ca-rev). Their sequences were based on BLAST (provided in the public domain by the National Center for Biotechnology Information and available at http://www.ncbi.nlm.nih.gov) searches of rDNA regions highly conserved among and specific to fungi and C. albicans, respectively. The forward fungal PCR primer (5′-GCA TCG ATG AAG AAC GCA GC-3′) and reverse fungal PCR primer (5′-TCC TCC GCT TAT TGA TAT GC-3′) encompass the highly conserved 5.8s rRNA, internal transcribed sequence-2 (ITS-2), and 28s rRNA regions and were designed to amplify a variety of different fungal DNA templates. Although genetic variation is expected for different fungal species, the fungal PCR was predicted to amplify a 328-bp region of C. albicans DNA. The C. albicans forward primer (5′-TCA TCG AAT CTT TGA ACG CAC A-3′) and the reverse primer (5′-GAC GTT ACC GCC GCA AGC A-3′) were designed to specifically amplify C. albicans DNA. The C. albicans–specific primers were predicted to produce a 200-bp product. PCR conditions previously described ^{10 11} were optimized for the two primer sets, which included 1 pmol of forward and reverse primers and 5 mM MgCl₂ in 50-μl reactions. The cycling was performed at 94°C, 58°C, and 72°C, 3 minutes each for 1 cycle, and 94°C, 58°C, and 72°C, 1 minute each for 30 cycles. Thirty percent of each PCR product was resolved on a 1.8% agarose Tris-borate-EDTA (TBE) gel and was visualized using ethidium bromide and ultraviolet excitation. 

In addition to size differentiation, the PCR products were confirmed by Southern blot analysis, using standard methods. ⁷ In general, the PCR products were immobilized on a nylon membrane by capillary transfer and hybridized with a probe end labeled withγ -[³²P]ATP. The probe (5′-GCA TGC CTG TTT GAG CGT CGT TTC T-3′) was designed from internal sequence of the predicted PCR product of the C. albicans–specific PCR. The hybridization was performed at 60°C for 16 hours and washed, using standard methods. ⁷ Autoradiography with the use of intensifying screens was used for visualization of the blot analysis results. 

Results

Panfungal and C. albicans PCR Sensitivity

The sensitivities of the panfungal and C. albicans PCR primers were evaluated using serially diluted C. albicans genomic DNA isolated from cultured C. albicans ATCC strain 32354, plasmid pCA1, and cultured C. albicans that was spiked into PBS and extracted with the protocol developed for the eye swab specimens. Electrophoretic analysis of PCR reactions showed that both primer sets yielded a single band at the predicted size of 328 bp for the fungal primers and 200 bp for the C. albicans–specific primers. Southern blot analysis was performed using a C. albicans–specific probe on control PCR reactions with genomic DNA amplified from C. albicans, S. cerevisiae, and A. flavus. The probe hybridized specifically to the C. albicans PCR product, confirming that the predicted sequence had been amplified in the PCR (data not shown). The lower limit of detection for both assays was consistently 10³ genome equivalents when using genomic DNA as a template (Figs. 1A 1B) , and 10² copies of pCA1 were consistently detected (Fig. 2A) . To test the sensitivity of the PCR assays in the context of the specimen-extraction protocol, known amounts of cultured organism were spiked into PBS, and the DNA was extracted. Efficacy of DNA recovery was consistent, based on subsequent PCR analysis (data not shown). The lower limit of detection of spiked C. albicans was consistently 10³ organism equivalents for both the panfungal PCR (data not shown) and the C. albicans–specific PCR (Fig. 2B) . The sensitivity of both PCR assays was within acceptable ranges for application to clinical specimens. 

Panfungal and C. albicans PCR Specificity

To determine the specificity of the panfungal and C. albicans–specific assays, the primers were evaluated, using more than 10⁶ copies of genomic DNA from several different species of higher eukaryotes, fungi, prokaryotes, and viruses—an amount greater than 3 logs above the lower limit of sensitivity of the PCR assays. The panfungal PCR specifically amplified DNA from all fungal species tested including yeasts and filamentous molds (C. albicans, S. cerevisiae, A. flavus), but did not amplify mammalian (human, bovine, murine), bacterial (E. coli, P. aeruginosa), or viral (HSV, adenovirus) DNA (Fig. 3A) . The C. albicans–specific PCR amplified only the C. albicans genomic DNA, and was negative for the other DNA templates tested, including the other fungi (Fig. 3B) . 

C. albicans PCR Species Specificity

To further define the specificity of the C. albicans PCR, DNA was extracted from several cultured clinical isolates and ATCC cultures of Candida spp., including C. albicans (n = 10), C. tropicalis (n = 3), C. glabrata (n = 3), C. guilliermondii (n = 2), C. lipolytica (n = 1), and C. parapsilosis (n = 1). The PCR products were analyzed on ethidium-stained agarose gels. Representative results are shown in Figure 4 . 

The C. albicans PCR primers amplified all 10 clinical isolates of C. albicans. To determine genotypic diversity among the C. albicans isolates tested, genomic DNA from each of the isolates was evaluated by restriction enzyme analysis. ¹⁴ Resultant DNA fragments demonstrated heterogeneity in the banding patterns between the isolates; particularly in the 2- to 6-kb range (data not shown). These results confirmed diversity among the various strains of C. albicans used, and demonstrated that the PCR was not restricted to a particular isolate. PCR analyses of the various isolates of C. tropicalis, C. glabrata, C. guilliermondii, C. lipolytica, and C. parapsilosis were all negative. 

PCR Analysis of Donor Eye Swab Specimens

The panfungal and C. albicans–specific PCR assays were used on pre-excision swab specimens to determine the prevalence of fungal flora from donor eyes that were to be excised for corneal transplantation. The samples were systematically collected during a 3-month period and analyzed in a masked fashion. Pellets from 111 swabs were processed for DNA extraction and tested in both the fungal and C. albicans PCR assays. Representative results are shown in Figure 5 . Thirty-five percent (39/111) of the samples tested were positive in the fungal PCR with 50% of those testing positive in both eyes (Table 1) . Several of the samples may have contained more than one organism, as indicated by the amplification of more than one PCR product (Fig. 5A , lanes H and N). Analysis using the C. albicans PCR showed that 19% (21/111) of the swab specimens were positive for C. albicans DNA, with 25% of the positive cadavers testing positive in both eyes. All 21 (100%) of the C. albicans–positive specimens were positive in the fungal PCR, with at least one band of the appropriate size (328 bp) for amplification of C. albicans by the fungal PCR. Multivariable logistic regression, which controlled for the effects of donor age, gender, and cause of death, was performed on the PCR data. The odds of having a positive panfungal PCR was 3.66 times greater when the time of death to time of collection was 12 hours or more (P = 0.03, 95% confidence limits, 1.44 and 9.31). Similarly, the odds of PCR showing a positive C. albicans was 4.67 times greater (P = 0.01, 95% confidence limits, 1.20 and 18.12). 

Discussion

We developed PCR-based techniques for the detection of both a broad-spectrum of fungal DNA and specifically C. albicans DNA. The panfungal primers were designed based on the nucleotide sequence from the multiple-copy, highly conserved rDNA regions of fungi. The C. albicans–specific primers were designed based on internal rDNA ITS-2 sequences specific to C. albicans. The rDNA region is ideal for both panfungal and species–specific fungal sequences and has been used successfully for targeting by PCR for fungi from serum and vitreous samples. ^{15 16} The optimized PCR assays in our study resulted in amplification of the predicted 328-bp and 200-bp products, respectively. The assays had similar sensitivities with lower limits of detection of less than 10² copies of pCA1, a plasmid containing the relevant region of C. albicans DNA. Both the panfungal and C. albicans–specific primer sets amplified target DNAs without amplification of products from nontarget templates. 

The sensitivity and specificity of the panfungal and C. albicans–specific PCR assays is advantageous, because it makes it possible to rapidly screen specimens from ocular surfaces for mycologic contamination. This allowed the screening of DNA extracted from samples collected from corneal donors for potential fungal contamination. The prevalence of positive samples (35%) determined with our panfungal primers was slightly higher than the range reported previously for ocular fungal flora by culture analysis in the United States (3%–28%). ^{3 4 5} However, the prevalence in our study is within the range of the overall carrier rate of fungi reported from various parts of the world in healthy individuals, which varies from 2.5% to 52%. The reasons for the variations in prevalence are not known, but warm moist climates, rural agricultural environments, or advanced age may be associated with an increase in positive fungal cultures. ¹⁷ A likely reason for the relatively high prevalence observed in our study is the differences between healthy individuals and cadavers. Physiological changes such as differences in tear film or a reduction in eye movements or blinking could feasibly contribute to the amount of fungi present or the ability for the organisms to grow. Other factors that may have contributed to our detection rates include the increased sensitivity of the PCR over standard culture methods and the geographic location of specimens collected. Based on logistic regression analysis, the chance of detecting C. albicans and other fungi in our study was significantly greater when the time between death and donation was prolonged. Yeasts and hyphae may proliferate on the ocular surface after death, migrate from the periocular skin, or reach the decedent’s eyes from the environment. 

Of interest, the broad-spectrum fungal PCR amplified multiple bands of varying sizes in several of the positive samples, indicating more than one organism was present in the swab specimen. This is not necessarily surprising, because many species of fungi are found in the conjunctival sac of a healthy population and because several of these, which are transient colonizers, are also known to be associated with ocular mycoses. ¹⁷ Table 2 lists several common fungal ocular pathogens and their predicted PCR product size when using our panfungal primers. Some of the non–C. albicans bands we observed were consistent with the predicted sizes of other ocular fungi. For instance, several positive samples had signals corresponding to the predicted size of Aspergillus sp., a common ocular fungus. Analysis of the PCR products by Southern blot or sequencing could be used for genera-specific identification. Regardless, the broad-spectrum PCR was a good indicator of the prevalence of fungi-positive ocular surfaces and a general indication of the diversity in the organisms present. It is also noteworthy that although the C. albicans primers were species specific, 10 different isolates of C. albicans were readily detected with the C. albicans primers, including seven different clinical isolates. This suggests that negative findings with the C. albicans–specific PCR for any of the specimens tested were not negative because of genetic diversity of a particular C. albicans strain. 

The risk of developing fungal endophthalmitis after penetrating keratoplasty is low (0.1%–0.7%), ¹⁸ but the outcome usually has devastating visual consequences. Eye banks routinely irrigate the donor eye with povidone-iodine solution before corneal excision, but the corneas are stored in preservation medium without antifungal supplements for up to 10 days before transplantation. Furthermore, most eye banks in the United States do not perform donor rim or preservation medium cultures, and the cost-effectiveness of these procedures by the corneal surgeons has been questioned. ¹⁹ Consistent with the low risk, there were no reports of fungal endophthalmitis in any of the transplanted corneas that were screened in our study. Although our findings demonstrate that fungi, especially C. albicans, are commonly found on the ocular surface of deceased corneal donors, it remains unclear how many transplanted corneas are contaminated with fungi. It is important to emphasize that the samples tested in this study were obtained before treatment of the donor eyes with povidone-iodine solution. It will be of interest in future studies to use these PCR assays for the testing of donor rims to identify the level of fungi persisting after storage in preservation medium. In conjunction with the established quality control for corneal storage in eye banks, there is potential for these PCR assays to aid in rapid screening of clinical specimens for fungal contamination and in the diagnosis of ocular infections. 

 Supported by the Sid W. Richardson Foundation, Houston, Texas; Research to Prevent Blindness, New York, New York; and National Eye Institute Core Grant 2P30EY02520-21 for Vision Research.

 Submitted for publication June 4, 2001; accepted July 11, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

The sensitivity of the panfungal (A) and C. albicans–specific (B) PCR primers was tested using 10-fold serially diluted C. albicans genomic DNA from 10⁵ genome equivalents to 10 genome equivalents. PCR products of 328 bp (panfungal primers) and 200 bp (C. albicans primers) were electrophoretically analyzed on a 1.8% agarose/0.5× TBE gel stained with ethidium bromide. Lane M: 100-bp molecular weight marker; lane H ₂ O: PCR negative control.

Figure 2.

 

The sensitivity of the C. albicans–specific primers was tested using plasmid pCA1 10-fold serially diluted from 10⁶ copies to 1 copy (A). Cultured C. albicans was serially diluted from 10⁵ organisms to 1 organism and spiked into PBS, after which the DNA was extracted (B). PCR products of 200 bp (C. albicans primers) were electrophoretically analyzed on a 1.8% agarose/0.5× TBE gel stained with ethidium bromide. Lane M: 100-bp molecular weight marker; lane H ₂ O: PCR negative control.

Figure 3.

 

The specificity of the panfungal primers (A) and C. albicans–specific primers (B) was tested using more than 10⁶ copies of genomic DNA from several different species of higher eukaryotes, fungi, prokaryotes, and viruses as a template. PCR products were electrophoretically analyzed on a 1.8% agarose-0.5× TBE gel stained with ethidium bromide. Lane M: 100-bp molecular weight marker; lane H ₂ O: PCR negative control.

Figure 4.

 

The specificity of the Candida PCR was tested at the species level. DNA was extracted from several cultured clinical isolates and ATCC cultures of Candida spp. Representative results are shown. PCR products were electrophoretically analyzed on a 1.8% agarose-0.5× TBE gel stained with ethidium bromide. Lane M: 100-bp molecular weight marker; lane + control: ATCC 32354 C. albicans genomic DNA; lane H ₂ O: PCR negative control.

Figure 5.

 

Panfungal (A) and C. albicans–specific (B) PCR analysis of donor eye swabs. Frozen pellets from donor eye swabs were processed for DNA extraction. Twenty percent of the extracted sample was used in each PCR. Several representative positive and negative results from individual extractions are shown. Lane M: 100-bp molecular weight marker; lane H ₂ O: PCR negative control.

Table 1.

 

Summary of Broad-Spectrum Fungal and Candida PCR-Positive Donor Eye Swabs

	Panfungal PCR	C. albicans PCR
Positive Eyes	39/111 (35)	21/111 (19)
Cadavers with Both Eyes Positive	12/24 (50)	6/24 (25)

 Data are total number PCR positive/total number tested, with the percent positive in parentheses.

Table 2.

 

Predicted PCR Product Size of Some Common Ocular Pathogens

Organism	Predicted PCR Product Size
Candida albicans	328
Saccharomyces cerivisiae	421
Aspergillus flavus	354
Pseudallescheria boydii	383
Cryptococcus neoformans	374
Blastomyces (Ajellomyces) dermatitidis	376
Paracoccidioides brasiliensis	391
Histoplasma capsulatum	366
Alternaria alternata	343
Curvularia spp.	∼360

 Predicted size based on PCR amplification of specific rDNA regions with the F-fwd and F-rev primers.

The authors thank Mary Scardino, RN, BSN, and Jeremy Nelson, CEBT, Chris Hanna, CEBT, Judy Garcia, CEBT, Anthony Thomas, CTBS, Rhia Pascua, Sumant Molhotra, and Bobby Tran, technicians of the Lions Eye Bank of Texas for their assistance in the collection of the donor eye swab specimens.