Use of Peptide Nucleic Acid-Fluorescence In Situ Hybridization for Definitive, Rapid Identification of Five Common Candida Species

Abstract

We investigated a 2.5-h peptide nucleic acid-fluorescence in situ hybridization (PNA-FISH) assay with five Candida species-specific probes to identify Candida colonies and compared it to standard 2-h to 5-day phenotypic identification methods. Suspensions were made and slides were prepared and read for fluorescence per the manufacturer's instructions. Sensitivity was 99% (109/110), and specificity was 99% (129/130). PNA-FISH can rapidly identify those Candida species isolated most frequently.

Candida spp. are important nosocomial bloodstream pathogens and account for disproportionate morbidity and mortality (2, 3). Although Candida albicans remains the most common isolate, nearly half of all invasive infections are now caused by other species (7, 8). Since different species have distinct antimicrobial susceptibility profiles, the prolonged time needed to identify organisms to a species level (up to 5 days) precludes the early use of less expensive, targeted therapy. Peptide nucleic acid-fluorescence in situ hybridization (PNA-FISH) is a diagnostic tool that uses fluorescence-labeled probes that hybridize to species-specific rRNA sequences. The PNA-FISH C. albicans assay (AdvanDx, Inc., Woburn, MA), which can identify C. albicans in ≤3 h directly from positive blood culture bottles (4-6), is now commercially available and has been demonstrated to be cost-effective (1).

We hypothesized that the use of the C. albicans probe and the specific PNA probes for C. glabrata, C. krusei, C. parapsilosis, and C. tropicalis would reduce the time to identify the colonies on agar plates and slants to a species level. Therefore, we evaluated PNA-FISH probes for rapid identification of the five species of Candida most commonly isolated.

We tested 238 recent clinical isolates, including 62 C. albicans, 63 C. glabrata, 42 C. tropicalis, 37 C. krusei, 21 C. parapsilosis, 7 C. dubliniensis, 5 Cryptococcus neoformans, and 2 Saccharomyces cerevisiae. The isolates were simultaneously identified using standard phenotypic tests such as the germ tube test, morphology on corn meal agar with caffeic acid, 7-carbohydrate fermentation, urease activity, phenoloxidase activity, and API 20-C strips (when required). The isolates were maintained on Sabouraud's dextrose agar at 4°C.

Fresh suspensions of yeast (newly subcultured with Sabouraud's dextrose agar and incubated for 18 to 24 h at 35 to 37°C) were prepared by inoculating the isolate into sterile deionized water to achieve a 0.5 McFarland standard. PNA-FISH probes for five Candida species (C. albicans, C. glabrata, C. krusei, C. parapsilosis, and C. tropicalis) were stored at 4°C. Slides were prepared according to the PNA-FISH manufacturer's recommendations. One drop of fixative and 10 μl of suspension were added to each slide and gently emulsified. Slides were then air dried and fixed with methanol. Next, 1 drop of PNA-FISH probe was added to each slide, and the slides were incubated at 55 ± 1°C for 90 min. Subsequently, slides were immersed in a preheated (55 ± 1°C), buffered wash solution for 30 ± 5 min to remove unbound probe, air dried, and mounted with a coverslip. A technologist who was not involved with the study coded the slides, including the positive and negative control slides. The slides were then examined by a trained reader (M.E.R. or A.B.M.).

Slides were viewed under a magnification of ×60 or ×100 (low-viscosity oil) with a dual-band filter (fluorescein isothiocyanate/Texas Red) fluorescence microscope to determine the presence or absence of morphologically consistent fluorescent organisms. In the analysis, the performance of the PNA-FISH assay was compared with that of standard phenotypic identification tests.

Finally, we estimated the potential impact of using PNA-FISH to hasten identification of Candida obtained from sterile sites at our institution. We searched the 2005 to 2006 laboratory database to identify how many patients with infection by Candida isolated from a sterile site could have had a specific diagnosis earlier had species-specific PNA-FISH probes for Candida been available.

A total of 238 slides, in addition to positive and negative control slides, were prepared and examined within 2 h of preparation. The five species-specific probes were found to be 99% sensitive (95% confidence interval [CI], 95 to 100%) and 99% specific (95% CI, 96 to 100%) when tested against the matching species and a panel of five to eight other yeasts (Table 1). There was one false negative and one false positive (a C. albicans and a C. parapsilosis isolate tested with a C. albicans probe); both were identified correctly on repeat testing.

TABLE 1.

Identification of Candida by PNA-FISH versus standard phenotypic identification

PNA-FISH probe	No. of isolates found positive/no. isolates testeda								Test characteristics
	C. albicans	C. dubliniensis	C. glabrata	C. tropicalis	C. krusei	C. parapsilosis	S. cerevisiae	C. neoformans	% of sensitivity (95% CI)	% of specificity (95% CI)
C. albicans	32/33 (97%)	0/3	0/9	0/6	0/6	1/2	0/1	0/1	99 (84-100)	100 (82-100)
C. glabrata	0/10	0/1	34/34 (100%)	0/6	0/6	0/3	ND	0/1	100 (90-100)	100 (87-100)
C. tropicalis	0/8	0/2	0/11	21/21 (100%)	0/7	0/2	ND	0/1	100 (75-100)	100 (87-100)
C. krusei	0/6	0/1	0/7	0/6	13/13 (100%)	0/5	ND	0/1	100 (66-100)	100 (81-100)
C. parapsilosis	0/5	ND	0/2	0/3	0/5	9/9 (100%)	0/1	0/1	100 (84-100)	100 (88-100)

^aND, not done.

During 2005 to 2006, Candida species were isolated from a total of 192 sterile-site specimens from the same number of patients. The use of the five Candida species-specific probes could have identified 190 of 192 (99%) yeasts recovered from sterile sites, since probes are still not available for two species (C. dubliniensis and C. lusitaniae). Excluding the 118 isolates of C. albicans for which other rapid tests (germ tube or rapid enzymatic assay) are available for identification, the use of species-specific PNA-FISH probes still would have hastened definitive identification of 72/74 (97%) isolates from sterile sites.

We found that PNA-FISH can provide rapid and accurate identification of Candida at the species level from colonies, thereby enabling more directed early treatment for patients. At our institution, the number of patients from whom one of these five common Candida species is recovered from a sterile body site is ∼100 per year. PNA-FISH could allow identification of common non-Candida albicans species on average 24 h sooner than is possible with current phenotypic tests. Most clinical laboratories have access to a fluorescence microscope; therefore, the assay requires only a specialized heating chamber. Since consistent interpretation of fluorescence is necessary, parallel reading by two technologists for an initial period might be helpful; however, once readers attain competency, the assay has outstanding sensitivity and specificity. Importantly, the detection of positive fluorescence by PNA-FISH provides a final identification; therefore, tedious confirmatory testing is not necessary. Confirmation of improved clinical outcomes with early targeted treatment enabled by definitive identification of Candida at the species level from sterile sites using PNA-FISH, as well as the assay's cost effectiveness, will require prospective study.