| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Faculty of Medicine,1 Centre for Infectious Diseases and Microbiology, Westmead Millennium Institute,2 School of Molecular and Microbial Biosciences,4 Post Graduate Foundation in Veterinary Science,5 Veterinary Pathology Diagnostic Services, Faculty of Veterinary Science, University of Sydney,6 Centre for Infectious Diseases and Microbiology Laboratory Services, Institute of Clinical Pathology and Medical Research, Sydney West Area Health Service, Sydney, New South Wales, Australia3
Received 8 September 2006/ Returned for modification 7 November 2006/ Accepted 15 November 2006
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
Recent efforts to improve the sensitivity and specificity of diagnostic tests have focused on culture-independent methods, in particular nucleic acid-based methods, such as PCR assays. These can be applied to fresh and formalin-fixed, paraffin-embedded (PE) sections. Numerous studies have highlighted the advantages of using PCR technology to detect viable and nonviable fungal pathogens in a variety of clinical specimens. The majority of assays target multicopy genes, in particular the ribosomal DNA (rDNA) genes (18S, 28S, and 5.8S) and the intervening internal transcribed spacer (ITS) regions (ITS1 and ITS2), in order to maximize sensitivity and specificity. To date, most assays have been designed to detect Candida or Aspergillus species only (14, 16, 17, 22, 24, 35, 38, 39, 43). Given that more than 200 fungal species have been reported to cause disease in humans and companion animals (11), the clinical utility of a species-specific or even a genus-specific assay is limited. Panfungal PCR assays, on the other hand, have the potential to detect all fungal species, but many rely on additional, time-consuming techniques, such as species-specific probes and hybridization, to identify the pathogen (10, 12, 15, 25, 29, 34, 37). Furthermore, probe design is restricted to known pathogens and does not allow the identification of new and emerging agents. Sequence-based identification of PCR products is a sensitive alternative, provided that accurate sequences have been submitted to public databases, e.g., GenBank (5, 32, 42).
In this study, we developed and evaluated a panfungal PCR assay to detect and identify fungal pathogens directly from fresh and PE tissue specimens obtained from patients with culture-proven and/or histologically proven IFIs. We chose to target the ITS1 region, located between the 18S and 5.8S rRNA genes, because (i) it is multicopy (
100 copies in the fungal genome), (ii) universal fungal primers are available, and (iii) it contains highly variable regions for species identification (6, 15, 17, 18, 20, 23, 24, 26), and we used DNA sequence analysis for species identification. The results of the molecular identification were correlated with those obtained by histological examination and culture.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
|
3 h in proteinase K and lysis buffer at 55°C, and the DNA was extracted using the MagNAPure LC instrument with the MagNAPure LC DNA isolation kit II (Tissue) (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. The DNA was stored at 20°C prior to use. To monitor contamination, each specimen was shadowed by a negative control containing molecular-biology-grade water (Eppendorf AG). PCR amplification and DNA sequencing. PCRs were performed in a 25-µl volume consisting of 1x PCR buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl, 1.5 mM MgCl2, and 0.001% gelatin) (Applied Biosystems, Foster City, CA), 2 mM MgCl2 solution (Roche Diagnostics), 5% glycerol (Sigma Chemical Co., St. Louis, MO), 0.25 mM deoxynucleoside triphosphates (Roche Diagnostics), 0.8 µM primers ITS 1 (5'-TCC GTA GGT GAA CCT GCG G) and ITS 2 (5'-GCT GCG TTC TTC ATC GAT GC) (40) (Sigma-Genosys, Australia), 1.25 U of AmpliTaq Gold DNA polymerase (Applied Biosystems), and 10 µl of DNA. Amplification was performed on a Mastercycler gradient thermocycler (Eppendorf AG). The thermal cycling conditions were 95°C for 10 min, followed by 60 cycles of 94°C for 15 s, 55°C for 30 s, and 72°C for 30 s, and a final extension at 72°C for 5 min. A negative control of molecular-biology-grade water and an inhibition control composed of an equal mixture of specimen DNA and Candida parapsilosis (strain no. ATCC 22019) positive-control DNA were included for each specimen. The inhibition control was used to exclude the presence of inhibitory substances. PCR products were separated by 1.5% agarose gel electrophoresis, stained with SYBR Safe DNA gel stain (Molecular Probes, Eugene, OR), and visualized by UV light transillumination. PCR products were purified using the GFX PCR DNA and gel band purification kit (Amersham Biosciences, Castle Hill, Australia) and were sequenced using the ITS 1 primer and the BigDye Terminator (version 3.1) cycle sequencing kit in the ABI PRISM 3100 genetic analyzer (Applied Biosystems). Sequences were edited using Chromas (version 2.23) software (Technelysium Pty. Ltd.) and entered into a BLASTN sequence analysis search (2) provided by BioManager, ANGIS (http://www.angis.org.au), for species identification.
Contamination control. Prior to any experimental procedure, all work surfaces and equipment, including cabinets, pipettes, and racks, were wiped down with NucleoClean decontamination solution (Chemicon International, Temecula, CA). DNA extraction, PCR setup, PCR amplification, and agarose gel electrophoresis were performed in separate, independently equipped laboratories, set out in a unidirectional workflow to prevent carryover contamination. PCR master mixes were prepared using a PCR cabinet cleaned with 70% alcohol and UV irradiated for 20 min, positive-displacement pipettes, and aerosol-resistant pipette tips.
| RESULTS |
|---|
|
|
|---|
Specimens from patients with culture-proven IFIs. Thirty-one fresh and 16 PE tissue specimens from 38 patients with culture-proven IFIs were tested. Patients were infected with a diverse range of yeasts and molds, including the uncommon pathogens Metarhizium anisopliae and Microsphaeropsis arundinis (Table 1). A PCR product was obtained from each specimen; however, DNA sequencing was unsuccessful for three (patients 36 to 38). For two of these, DNA was extracted from PE sections in which only scant fungal hyphae were seen upon histological examination. For the remaining specimen (patient 37), multiple bands were present on the gel, suggesting a mixed fungal infection; however, a single isolate recovered by culture was identified as Glomerella lagenaria by ITS sequence analysis.
Sequencing results correlated with culture identification for 39 of 44 (88.6%) specimens (Table 1). For the remaining five specimens, analysis yielded sequences with 98 to 100% identity to sequences of fungi that are phylogenetically closely related to the species identified by culture-based methods (given in parentheses): patient 18, Neosartorya pseudofischeri (Aspergillus fumigatus); patient 27, Exophiala spinifera (Exophiala jeanselmei); patient 35, Rhizomucor pusillus (Absidia corymbifera) (Table 1). Forty-one of 44 (93.2%) sequences showed
98% identity to sequences deposited in the GenBank database. The other three demonstrated 96 to 97% identity. Sequence analysis of the ITS1 region was unable to differentiate between members of the Cryptococcus neoformans complex (Cryptococcus gattii, Cryptococcus neoformans var. neoformans, and Cryptococcus neoformans var. grubii); Fusarium species, including Fusarium verticillioides, F. subglutinans, and F. proliferatum; Neosartorya spp. and Aspergillus lentulus; Rhizopus oryzae and Amylomyces rouxii; and Exserohilum rostratum and Exserohilum mcginnisii.
Specimens from patients with histologically proven IFIs.
Six fresh and 22 PE tissue specimens from 24 patients with histologically proven IFIs were analyzed by PCR. Nine of these specimens (six patients) were obtained at postmortem examination. Fungal DNA was amplified from 18 (64.3%) samples, and DNA sequencing identified a variety of fungal pathogens, including Candida spp., Trichosporon spp., Aspergillus spp., and zygomycetes (Table 2). With the exception of one specimen (patient 51), sequence analysis showed
98% identity to sequences in the GenBank database. The molecular identification was consistent with the histological diagnosis in 17 of 18 (94.4%) cases. The exception was patient 40, where C. parapsilosis DNA was detected in a nasopharyngeal biopsy specimen but histopathology (showing narrow-necked budding, encapsulated yeasts) was consistent with a C. neoformans complex infection. All of the PCR-negative specimens (n = 10) were PE sections; six of these (patients 59 to 62) had morphological characteristics indicative of a zygomycete, while scant fungal elements were seen in three of the remaining four (patients 56 to 58). Cryptococcus albidus and Cryptococcus albidosimilis (patient 41) were not distinguished by sequence analysis of the ITS1 region.
Specimen from a patient with a suspected IFI. PCR and DNA sequence analysis of a skin biopsy specimen from patient 63 identified the causative pathogen as Trichophyton verrucosum, with 100% identity to T. verrucosum sequences in the GenBank database.
| DISCUSSION |
|---|
|
|
|---|
The validity and clinical applicability of the assay were confirmed by testing specimens from patients with culture-proven IFIs (Table 1). Although the PCR amplified fungal DNA for all 47 culture-positive specimens, DNA sequencing was unsuccessful in 3 cases. In one case, this was probably due to the presence of mixed fungal species, since multiple bands were present on the gel. In the other two, insufficient DNA may have been extracted from the paraffin sections, or nonspecific PCR products may have been generated.
Comparative sequence analysis confirmed the conventional culture-based identification for 39 of the 44 (88.6%) remaining specimens. Discordant results were observed for five specimens from three patients (Table 1). The cultures of three retrobulbar specimens from one patient yielded A. fumigatus by conventional identification, but subsequent DNA sequence analysis of the complete ITS region identified the culture as N. pseudofischeri. Although genetically distinct, the asexual state of N. pseudofischeri (and the closely related Neosartorya fischeri) is morphologically similar to A. fumigatus. Until recently, Neosartorya spp. have rarely been reported to have caused disease, although the difficulty in morphologically distinguishing A. fumigatus from Neosartorya spp. may have led to underestimation of the frequency of infection caused by the latter (3). Accurate identification is clinically important, since it has been reported that Neosartorya spp. are less susceptible in vitro to antifungal agents than A. fumigatus (3).
For the second patient, molecular analysis of the fungal pathogen from a skin biopsy specimen (PE) identified it as Exophiala spinifera, while the isolate was identified morphologically from culture of fresh tissue as Exophiala jeanselmei. Comparison of ITS1 sequences of multiple strains of E. spinifera and E. jeanselmei deposited in GenBank revealed that the two species can be distinguished easily. The isolate from patient 27 showed 98% sequence identity to six E. spinifera strains from the Centraalbureau voor Schimmelcultures. The variable morphological characteristics of Exophiala spp. make definitive identification difficult; as a result, molecular analyses are increasingly used for species confirmation (30).
PCR amplification and DNA sequencing twice identified the fungal pathogen from a nasal turbinate biopsy specimen (PE) from the third patient as Rhizomucor pusillus rather than Absidia corymbifera. Although these two organisms share antigenic similarities, they are morphologically distinct, and comparison of their ITS1 sequences reveals substantial differences (33). Given that the organism was isolated by another laboratory in 1992, it was not possible to check the details of the original identification, but it is likely to be a case of misidentification.
The panfungal PCR assay performed well on specimens where fungal elements were visualized but no pathogen was grown. A molecular identification could be assigned to 18 of 28 (64.3%) specimens, including all 5 fresh specimens, and the identification was consistent with the histological findings for all but 1. In this case, C. parapsilosis DNA was detected from a nasopharyngeal biopsy specimen, but histologically the pathogen was identified as C. neoformans complex. It is likely that C. parapsilosis was present as a commensal, rather than as the infecting agent, in the nasopharynx of this patient and was amplified by the PCR assay. This case and those reported by others (42) demonstrate that the application of a sensitive, broad-range nucleic acid test to specimens from nonsterile sites should always be interpreted in the appropriate clinical context. The PCR was negative for 10 specimens from eight patients with histologically proven IFIs (Table 2, patients 55 to 62). Since all of these specimens were PE, the quality of DNA is likely to have been compromised by the routine processes required for histological examination, in particular the duration of contact with formalin fixative prior to paraffin embedding (41, 42). Additionally, insufficient amounts of DNA may have been present for detection by PCR, since at least three of the specimens had scant fungal elements upon histological examination (Table 2).
DNA sequencing proved useful in two additional clinical contexts. First, tissue obtained at autopsy, where IFI is not clinically suspected antemortem, is frequently not submitted for culture. In our study, the fungal pathogen was identified by PCR and DNA sequencing in specimens obtained at postmortem examination from five of six patients (Table 2). Second, for a patient with follicular skin lesions for whom fungal infection was suspected, the diagnosis of T. verrucosum infection achieved in the absence of either culture or histology results was central to selecting appropriate antifungal therapy.
Studies have demonstrated that the most promising targets for molecular identification of fungi are the ITS1 and/or the ITS2 region, followed by the D1-D2 region of the large-subunit DNA gene (28S rDNA) (7, 17, 23, 31). For this reason, we targeted the ITS1 region. Nevertheless, our results indicate that there is insufficient sequence variation to differentiate between species of certain genera by using the ITS1 region alone (e.g., C. neoformans complex, C. albidus and C. albidosimilis, Neosartorya spp. and A. lentulus, and Fusarium spp.). These results were not unexpected, since it has been reported that ITS sequence variations among some species of fungi, including Aspergillus spp., are minimal (13, 17, 19, 36), and the differentiation of Fusarium species is complicated by the presence of >1 ITS sequence variant in a single strain (27). For those species that cannot be discriminated using the ITS1 region alone, future work may involve sequence analysis of additional genes, such as ITS2, the D1-D2 region of 28S rDNA, or intergenic spacer regions (19).
Despite the usefulness of the panfungal PCR assay for the identification of fungal pathogens in histologically positive but culture-negative tissue specimens, the limitations of the assay must also be considered. Environmental contamination of specimens and/or the PCR master mixture by ubiquitous fungal spores is a possible cause of "false-positive" results in a sensitive, broad-range PCR assay. This was not likely to have occurred in the present study, because strict precautions were taken throughout the whole procedure, including processing each specimen with its own negative control during DNA extraction and PCR amplification and performing all procedures using a unidirectional workflow pattern. The turnaround time for results is realistically a minimum of 48 h for fresh specimens and 4 to 5 days for PE sections. Two working days are required to remove the wax from PE tissue specimens, and one working day is required to extract the DNA and perform PCR amplification and product detection. Additionally, we rely on an external DNA sequencing facility, and its turnaround time is 24 to 48 h. Importantly, accurate sequence-based identification of fungal pathogens is dependent on the quality and accuracy of sequences in existing databases. GenBank sequence submissions are not peer reviewed, and it has been estimated that 10 to 20% of fungal sequences in GenBank are misidentified organisms (4; R. Summerbell, presented at the 16th Congress of the International Society for Human and Animal Mycology, 2006). Given that in our study, the molecular identification was consistent with the histological diagnosis for 17 of 18 specimens for which histology only was positive, the inaccuracy of some GenBank entries should not have affected our results. However, the development of an accurate sequence database for fungal species would be beneficial for this assay.
In conclusion, this study illustrates the potential benefits of using the panfungal PCR assay in combination with conventional laboratory tests for sensitive and specific identification of fungal pathogens in both fresh and PE tissue specimens. For rapid diagnosis of IFIs from fresh tissue, we recommend that the panfungal PCR be performed as soon as fungal elements are seen on microscopy. For PE specimens, we envisage the assay to be most valuable in cases where fungal hyphae are visualized and where either the culture result is negative or culture is not performed. Evaluation of the panfungal PCR assay on specimens from other sterile body sites, e.g., blood, is indicated.
| ACKNOWLEDGMENTS |
|---|
This work was supported in part by a Centre of Clinical Excellence grant (264625) from the National Health and Medical Research Council of Australia. A.L. is supported by an Australian Universities Post-graduate student award.
| FOOTNOTES |
|---|
Published ahead of print on 22 November 2006. ![]()
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Antimicrob. Agents Chemother. | Clin. Microbiol. Rev. |
|---|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|