INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Candida dubliniensis, which was first established as a novel yeast species in 1995, is phenotypically and genotypically closely related to the most frequently identified human fungal pathogen, Candida albicans (25). This close similarity between the two species has proved problematic in the identification of C. dubliniensis in clinical samples and in retrospective analyses of laboratory stock collections, with many isolates being misidentified as C. albicans (5, 18). However, the recent description of reliable and rapid identification tests, including the observation of differentially colored primary colonies on CHROMagar Candida medium and the use of C. dubliniensis-specific PCR primers, will greatly facilitate the identification of this species in clinical samples and establish its epidemiologic significance (6, 24). In addition, accurate species identification has been aided by the inclusion of C. dubliniensis-specific carbohydrate assimilation profiles in the databases of commercially available yeast identification kits, such as the API ID 32C and the API 20C AUX systems (20, 24). To date the majority of C. dubliniensis isolates have been identified in Western Europe and North America (1-3, 7, 10-14, 17, 18, 23, 25, 26). Most of these isolates were associated with oral carriage and oropharyngeal infection in human immunodeficiency virus (HIV)-infected individuals. In a recent study of an Irish subject group, 26% (48 of 185) of HIV-positive individuals and 32% (26 of 82) of AIDS patients with oral candidiasis yielded C. dubliniensis. In approximately 25% of these cases C. dubliniensis was the only species detected (5, 24). However, C. dubliniensis is not exclusively associated with HIV-infected individuals. In the same study C. dubliniensis was also identified in clinical specimens recovered from HIV-negative individuals, both with and without symptoms of oral candidiasis. In an analysis of oral samples taken from healthy individuals without any signs of oral disease, 3.5% (7 of 202) of subjects yielded C. dubliniensis, suggesting that this species is a minor constituent of the normal human oral flora. C. dubliniensis was also identified in 12.5% (16 of 128) of cases of oral candidiasis in HIV-negative individuals (5, 24). Although the oral cavity is the human niche from which C. dubliniensis has been recovered most frequently, there have been reports of isolates being recovered from other anatomical sites and specimens, including the vagina, the lung, feces, and sputum (16, 18, 21). Furthermore, C. dubliniensis was recently identified as the cause of three cases of systemic disease in HIV-negative Dutch patients receiving post-bone marrow transplantation immunosuppressive treatment or cytotoxic chemotherapy for the treatment of rhabdomyosarcoma (14). In this study we describe the application of routine phenotypic and rapid molecular methods to the identification of C. dubliniensis isolates from five separate HIV-negative hospitalized patients in Israel. All five isolates were shown to be susceptible to a range of antifungal agents, including fluconazole. This is the first report of the identification of this novel species in the Middle East.

	CASE REPORTS

Case 1. A 39-year-old female with a past history of thalassemia major, splenectomy, and transfusion associated hemochromatosis initially presented with acute respiratory tract infection. The physical examination on admission was notable only for the presence of a purulent postnasal drip. Given the patient's asplenic condition, she was placed on intravenous cefuroxime, 750 mg every 8 h. Oral and upper respiratory specimens were inoculated onto blood agar (5% [vol/vol] defibrinated sheep blood), chocolate agar, and MacConkey agar media and incubated at 35°C for 48 h in an atmosphere of 5% (vol/vol) CO₂. With the recovery of beta-hemolytic streptococcus (Lancefield group G), Staphylococcus aureus, C. albicans, and Candida krusei, treatment was altered to cephalexin, 500 mg every 6 h, for a total duration of 14 days.

Case 2. A 19-year-old man was admitted to the hospital due to a urinary tract infection complicating type I neurofibromatosis. Massive retroperitoneal tumors were known to obstruct the ureters and deform the bladder, with secondary renal failure and recurrent episodes of urinary tract infection. In view of previous infection with resistant bacteria (prior to hospitalization the patient was repeatedly treated for urinary tract infection, with eventual culture of antibiotic-resistant bacteria), the patient was treated with broad-spectrum antibiotics, including vancomycin, meropenem, and co-trimoxazole, and was subsequently referred for surgical revision of his urinary tract. Peri- and postoperatively, during removal of drains, the patient was frequently given broad-spectrum antibiotics as a prophylactic measure, including ampicillin, ofloxacin, and metronidazole. Urine cultures during that period grew numerous bacteria, including Strenotrophomonas maltophilia, Citrobacter koseri, S. aureus, and Enterococcus faecalis. Forty days after admission, when a suprapubic catheter was removed without adequate alternative urinary drainage, the patient became febrile. A urine culture was positive for Enterococcus faecium and Trichosporon beigelii. Administration of antibacterial agents was continued, drainage of the urinary tract was restored by intermittent catheterizations, and the patient defervesced. During this time the patient did not receive any antifungal agent. Repeated urine cultures in the following 10 days were positive for T. beigelii, and antibiotic treatment was eventually stopped. Follow-up urine culture 5 days later yielded T. beigelii and C. dubliniensis. Because the patient was asymptomatic, he was not treated. A urine culture 2 weeks later was negative for fungi.

Case 3. A 21-year-old woman with cystic fibrosis and thalassemia minor was admitted to the hospital due to worsening dyspnea, productive sputum, and fever. The patient had recurrent episodes of Pseudomonas aeruginosa respiratory tract infection, the last being 1 month prior to this admission, for which she was treated with ceftazidime and amikacin and later switched to ciprofloxacin. Bronchoscopy revealed purulent secretions, with a Gram stain of lavage disclosing gram-positive cocci and gram-negative rods. Cultures of this lavage fluid were positive for S. aureus, P. aeruginosa, and C. dubliniensis. The patient was treated with ceftazidime and amikacin and improved from the bacterial infection within 5 days. Previous bronchoalveolar lavage and follow-up sputum cultures were persistently positive only for S. aureus and P. aeruginosa.

Case 4. A 52-year-old woman was admitted for right-side pneumonia leading to acute respiratory failure. Her past medical history was significant for indicating type II polyglandular autoimmune syndrome (Schmidt's syndrome) treated with prednisone (7.5 mg daily), insulin, and thyroxine; ischemic heart disease; and perforated appendicitis 5 months prior to admission. The patient was treated with cefuroxime and ciprofloxacin, and prednisone was replaced by hydrocortisone (300 mg daily). Bronchial washings performed on admission through the endotracheal tube disclosed few granulocytes without bacteria and no significant growth. Bronchoscopy done 4 days later disclosed a minute amount of purulent secretions. A Gram stain was negative, while a culture was positive for C. dubliniensis. Thereafter, acute renal failure developed, and peritoneal dialysis was initiated. The patient died 19 days later with evidence of C. albicans peritonitis and bloodstream infection (specimens from both sources yielded C. albicans). An autopsy was not performed. HIV serology was negative.

Case 5. A 31-year-old otherwise healthy woman was treated for postpartum endometritis with ampicillin, gentamicin, and metronidazole for 10 days. One week after completion of therapy she noticed a burning sensation on the tongue accompanied by a whitish discoloration that persisted. Three weeks later a diagnosis of oral candidiasis was made. A superficial tongue specimen disclosed mixed bacterial morphotypes with few leukocytes. A culture was positive for mixed bacteria and C. dubliniensis. She was not treated, and at a follow-up examination 4 weeks later all symptoms and signs had resolved. Microscopic analysis and culture of specimens taken from the endometrium were both negative for fungi.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Yeast isolates. Yeast isolates from clinical specimens were recovered following primary culture for 48 h at 30°C on Emmon's modified SGA supplemented with 50 µg of chloramphenicol and 5 µg of gentamicin/ml. Following incubation, confluent or semiconfluent areas of yeast growth were sampled with a sterile wire loop and streaked on CHROMagar Candida medium (CHROMagar, Paris, France) to yield single colonies. Selected colonies exhibiting different colony colors were transferred to and maintained on SGA at 30°C. A single isolated colony of each species grown on SGA was transferred after 48 h of incubation at 30°C to CHROMagar medium, incubated at 30°C, and examined for colony color after 24 and 48 h. Colonies from 48-h SGA cultures were similarly used as inocula in the following standard morphological and physiological tests: (i) chlamydoconidia formation on corn meal agar supplemented with 1% (wt/vol) Tween 80, (ii) germ tube development in human serum incubated for 3 h at 37°C, (iii) sensitivity to cycloheximide as determined by growth on Mycosel agar (BBL, Cockeysville, Md.), and (iv) growth at 37, 42, and 45°C on SGA. Carbohydrate source and nitrogen source assimilation patterns were evaluated by using the API ID 32C and the API 20C AUX yeast assimilation systems (bioMérieux, Marcy l'Etoile, France), according to the manufacturer's instructions, with an inoculum derived from 48-h SGA cultures.

Serotyping. C. dubliniensis isolates were serotyped on the basis of agglutination reactions with antiserum raised against Candida antigenic factor no. 6 (Iatron Laboratories, Inc., Tokyo, Japan) as described previously (25).

Chemicals, enzymes, and oligonucleotides. Analytical-grade or molecular biology grade chemicals were purchased from Sigma-Aldrich, BDH (Poole, Dorset, United Kingdom) or Boehringer Mannheim (Lewes, East Sussex, United Kingdom). Enzymes were purchased from Boehringer Mannheim or the Promega Corporation (Madison, Wis.) and used according to the manufacturer's instructions. Custom-synthesized oligonucleotides were purchased from Genosys Biotechnologies (Pampisford, Cambridgeshire, United Kingdom).

In vitro antifungal susceptibility tests. The in vitro antifungal susceptibilities of C. dubliniensis isolates were determined by using the Etest system (AB Biodisk, Solna, Sweden) according to the manufacturer's instructions. The MIC was defined as the lowest concentration of antifungal agent at which the border of the elliptical inhibition zone intercepted the readable scale on the strip (4). All tests were quality controlled by using C. krusei ATCC 6258 and Candida parapsilosis ATCC 22019.

Immunofluorescence. An indirect immunofluorescence assay (IFA) was performed as described previously (1). Briefly, the blastospores from C. dubliniensis isolates, including the C. dubliniensis type strain, CD36 (CBS 7989) (25), and reference oral C. albicans isolate 132A (9) were grown on Sabouraud agar (Oxoid, Poole, Dorset, United Kingdom) plates for 48 h at 37°C, resuspended in phosphate-buffered saline (PBS) at a cell density of 10⁶ cells/ml, and placed on Teflon-coated immunofluorescence slides. The slides were incubated with anti-C. dubliniensis rabbit serum (1) diluted 1:5 in PBS supplemented with Evans blue (0.05% [wt/vol]) and Tween 20 (0.05% [vol/vol]) and washed, and the reacting antibodies were revealed by incubation with fluorescein-conjugated goat anti-rabbit immunoglobulin G (Sigma).

PCR identification of C. dubliniensis. PCR identification of C. dubliniensis with the C. dubliniensis-specific primer pair DUBF and DUBR (6) was carried out in a 50-µl final volume containing 10 pmol each of the forward and reverse primers, 2.5 mM MgCl₂, 10 mM Tris-HCl (pH 9.0 at 25°C), 10 mM KCl, 0.1% (vol/vol) Triton X-100, 2.5 U of Taq DNA polymerase (Promega), and 25 µl of template DNA-containing cell supernatant (prepared as described below). The primer pair of DUBF and DUBR is complementary to sequences within the ACT1-associated intron sequence of C. dubliniensis and yields an amplimer of 288 bp. Each reaction mixture also contained 10 pmol each of the universal fungal primers RNAF and RNAR (8), which amplify a fragment of approximately 610 bp from all fungal large-subunit rRNA genes, as an internal positive control. Cycling conditions consisted of 6 min at 95°C, followed by 30 cycles of 30 s at 94°C, 30 s at 58°C, and 30 s at 72°C, followed by 72°C for 10 min. Amplification products were separated by electrophoresis through 2.0% (wt/vol) agarose gels containing 0.5 µg of ethidium bromide/ml and were visualized on a UV transilluminator.

Preparation of template DNA. Candida template DNA for use in PCR experiments with the C. dubliniensis-specific primer pair DUBF and DUBR was prepared as described by Donnelly et al. (6). Briefly, a single colony from a culture grown for 48 h at 37°C on potato dextrose agar or CHROMagar Candida medium was suspended in 50 µl of sterile distilled water. Cell suspensions were boiled for 10 min, and the lysed cells were subjected to a clearing spin for 5 min at 20,000 × g. Template DNA contained in 25 µl of supernatant was used for PCR amplification.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Phenotypic characterization of putative C. dubliniensis isolates. Clinical specimens from five separate hospitalized Israeli patients yielded yeast colonies which were dark blue-green in color on CHROMagar medium. These were all found to produce germ tubes in normal human serum, to form abundant chlamydoconidia following growth on corn meal agar, and to grow in the presence of cycloheximide on Mycosel agar. Based on these findings and in accordance with previous studies (5, 22, 25) these isolates were presumptively identified as C. dubliniensis. In order to confirm this identification, all five isolates were then subjected to substrate assimilation profile analysis with the API ID 32C and 20C AUX yeast identification systems. The profiles of four of the five corresponded to excellent identification of C. dubliniensis (Table 1). The profile of one isolate corresponded to good identification of C. albicans. Since C. dubliniensis was only first described as a new species in 1995, it has only recently been added to the API 20C AUX and the API ID 32C databases. The results of a recent study show that these systems have excellent potential as a means of identifying this yeast but that database modifications are required to avoid its misidentification as C. albicans or unidentified results (20). The largest discrepancy observed was the positive trehalose assimilation results found with 15 and 30% of the 80 C. dubliniensis isolates tested with the API 20C AUX and the API ID 32C systems, respectively. In that study the authors concluded that it is reasonable to assume that incorporation of this variability in a future database would correct this problem (20). Indeed, if isolate P-7073 had not assimilated trehalose, it would have been identified as C. dubliniensis with the API 20C AUX system with the current database.

PCR-based identification of C. dubliniensis isolates. In order to confirm the definitive identification of the five Israeli C. dubliniensis isolates, template DNA from each was subjected to PCR analysis with a set of primers (DUBF and DUBR) complementary to C. dubliniensis ACT1-associated intron sequences (6). These primers amplify a DNA fragment of 288 bp from C. dubliniensis, but do not yield an amplimer from C. albicans, Candida stellatoidea, or any other Candida species. Each PCR mixture also contained the fungal universal primer pair RNAF and RNAR (8), which amplify a product of approximately 610 bp from the fungal large-subunit ribosomal RNA gene and which served as an internal positive amplification control. All five Israeli C. dubliniensis isolates and the C. dubliniensis type strain, CD36, yielded amplimers of 288 bp and approximately 610 bp (Fig. 1). In contrast, C. albicans reference strain 132A yielded an amplimer of approximately 610 bp only (Fig. 1). These findings unequivocally confirmed the results of the phenotypic tests that the five isolates were C. dubliniensis.

View larger version (62K): [in this window] [in a new window]   FIG. 1.   Agarose gel showing ethidium bromide-stained amplimers from PCRs with the C. dubliniensis-specific primers DUBF and DUBR (288-bp product) and the fungal universal primers RNAF and RNAR (610-bp product). The amplimers shown in the lanes were obtained from template DNA from yeast strains and isolates as follows: lane 2, C. dubliniensis type strain CD36; lane 3, C. albicans reference strain 132A, lanes 4 to 8, Israeli C. dubliniensis isolates P-6265, P-6785, P-7073, P-7266, and P-7276, respectively. Lane 1, molecular weight size markers (100-bp ladder); lane 9, negative control lacking template DNA.