|Home | About | Journals | Submit | Contact Us | Français|
Cryptococcus gattii emerged in North America in 1999 as a human and veterinary pathogen on Vancouver Island, British Columbia. The emergent subtype, VGIIa, and the closely related subtype VGIIb can now be found in the United States in Washington, Oregon, and California. We performed multilocus sequence typing and antifungal susceptibility testing on 43 isolates of C. gattii from human patients in Oregon, Washington, California, and Idaho. In contrast to Vancouver Island, VGIIa was the most frequent but not the predominant subtype in the northwest United States. Antifungal susceptibility testing showed statistically significant differences in MICs between the subtypes. This is the first study to apply antifungal susceptibility testing to C. gattii isolates from the Pacific Northwest and the first to make direct comparisons between subtypes.
Cryptococcus species are basidiomycetous fungi, some of which are capable of causing invasive infection in both humans and animals. The major pathogens within the genus Cryptococcus are Cryptococcus neoformans and Cryptococcus gattii (2). C. neoformans and C. gattii differ significantly with regard to their geographical distributions and their ecological niches. While C. neoformans has a global distribution and is primarily associated with soil contaminated by bat or bird guano (19), C. gattii was until recently thought to be largely confined to tropical and subtropical regions, such as Australia and New Zealand, where it has been found primarily in association with Eucalyptus trees (11). C. gattii has also been isolated from clinical specimens from numerous other parts of the world, including Mexico, parts of Latin America, Europe, southern California, Hawaii, India, and Africa (15, 19, 25, 30).
In human immunodeficiency virus (HIV)-positive patients, the two species share similar clinical aspects; however, the diseases may differ significantly in otherwise healthy patients. C. neoformans primarily causes meningoencephalitis, while C. gattii has a propensity to cause cryptococcomas, focal lesions in the central nervous system (CNS), and significant neurological sequelae (5, 25).
In 1999, C. gattii began infecting animals and humans living in or traveling from Vancouver Island (VI), British Columbia, a temperate climate very different from that which was considered the traditional ecological niche of C. gattii (10, 14, 18). The majority of cases in this outbreak were caused by C. gattii of the clonal subtype VGIIa, with a minority of isolates caused by the clonal subtype VGIIb (12, 17, 18), but other subtypes have also been isolated during this ongoing outbreak (17).
In 2005, investigators documented the presence of C. gattii subtype VGIIa isolates in environmental samples taken from parts of the Pacific northwestern United States (16). Since that time, both human and animal cryptococcoses have been recognized in Washington and Oregon (4, 21, 29). Many of these cases have been caused by subtype VGIIa, but VGIIb and VGIIc isolates also play a role in this emergence (4). Because the VI outbreak major strain (VGIIa) is a unique subtype seen in no other part of the world (18), there has been substantial interest in subtyping all strains of C. gattii isolated from the Pacific Northwest to monitor whether other unique strains were emerging or whether the region where VGIIa could be found was expanding.
Through the collaboration of a public health working group, 38 recent and 5 archived isolates of C. gattii from the Pacific Northwest states of California, Washington, Oregon, and Idaho have been collected, genotyped, and tested for susceptibility to six antifungal drugs.
A total of 43 isolates of Cryptococcus gattii were analyzed at the Centers for Disease Control and Prevention. Public health institutions from the states of California, Oregon, Washington, and Idaho submitted 38 human and animal isolates from recent cases, including 28 isolates from the state of Oregon (17 human, 11 veterinary), four isolates from California (one human, three veterinary), five veterinary isolates from Washington, and one human isolate from Idaho. The CDC also had five archived C. gattii isolates, including two human isolates from Oregon from 2005, two human isolates from California from 1992 (3), and one veterinary isolate from California from 2001 (23). All isolates were confirmed as C. gattii using canavanine-glycine-bromthymol blue (CGB) medium (20). A single isolate that grew poorly and did not develop the characteristic cobalt blue color on CGB medium was confirmed as C. gattii by molecular typing, as described below.
All isolates were subtyped by multilocus sequence typing (MLST). The URA5, IGS1, and CAP59 gene fragments were amplified as described previously (22). Briefly, isolates were grown on yeast extract-peptone-dextrose (YPD) plus 0.5% NaCl, and DNA was prepared using an UltraClean DNA isolation kit as described by the manufacturer (MO BIO Laboratories, Carlsbad, CA). Gene fragments were amplified using the primers of Meyer et al. (22) and the same primers were used to sequence each fragment in both directions. Sequences were compared to published sequences in GenBank to assign an allele number and subtype for the IGS1 and CAP59 genes according to the supplementary allele table of Fraser et al. (12). Alleles for unpublished IGS1 and CAP59 sequences and alleles for the URA5 gene were assigned allele numbers by Wieland Meyer, University of Sydney, the current curator of the C. gattii MLST allele data set (W. Meyer, personal communication). The VGIIc alleles for CAP59 and IGS1 in this study were identical to the alleles first reported for VGIIc isolates by Byrnes and coworkers (4).
Antifungal susceptibility testing was performed on all isolates by broth microdilution with fluconazole, voriconazole, itraconazole, posaconazole, and flucytosine as outlined in the Clinical and Laboratory Standards Institute (CLSI) document M27-A3 (7) using RPMI broth and trays custom manufactured by TREK Diagnostics (Cleveland, OH). An Etest was used for amphotericin B susceptibility testing. All results were read visually after 72 h of incubation.
The alleles in this study can be found under the following GenBank numbers: URA5_2 (AY973125), URA5_7 (AY973124), URA5_12 (AY973114), URA5_19 (AY973153), URA5_23 (GU250873), IGS1_1(DQ096311), IGS1_3 (DQ096313), IGS1_4 (DQ096314), IGS1_10 (DQ096319), IGS1_15 (DQ096324), IGS1_18 (DQ096327), CAP59_1 (DQ096432), CAP59_2 (DQ096433), CAP59_6 (EU937819), CAP59_16 (GU250874), CAP59_18 (GU250875), CAP59_29 (GU250876), and CAP59_30 (GU250877).
Following the establishment of the C. gattii Public Health Working Group in 2008, public health institutions in Alaska, California, Idaho, Montana, Oregon, and Washington began passive surveillance for human and veterinary cases of C. gattii in their states. From this surveillance the CDC received 38 isolates from the states of Oregon, California, Washington, and Idaho. Isolates came from human cases as well as a number of different animals, including companion animals such as dogs and cats, as well as livestock, including llamas, alpacas, horses, and elk (Table (Table11).
Because the Vancouver Island major outbreak strain VGIIa was a unique subtype seen in no other part of the world (18), there has been substantial interest in subtyping all strains of C. gattii isolated from the Pacific Northwest to monitor the potential for emergence of other unique strains and/or expansion of the region where VGIIa could be found. To this end, we typed all of the submitted isolates by MLST (22). Our results showed that 16 of the isolates were the major VI outbreak strain VGIIa, 7 of the isolates were the minor VI outbreak strain VGIIb, 12 of the isolates were the unique Oregon strain VGIIc (4), and 8 of the isolates were either subtype VGI or VGIII (Table (Table11).
The MICs for each isolate against four azole drugs (fluconazole, voriconazole, itraconazole, and posaconazole), flucytosine, and amphotericin B are shown in Table Table2.2. All isolates had relatively low MICs with flucytosine and amphotericin B. There was a wide range of MICs for all four of the azole drugs. The range for voriconazole was 0.008 to 1 μg/ml, for fluconazole, 0.5 to 32 μg/ml, and for itraconazole, 0.06 to 2 μg/ml (Table (Table3).3). The range for posaconazole was somewhat smaller, 0.12 to 1 μg/ml.
The geometric mean MICs for all isolates of each subtype are given in Table Table4.4. The geometric mean MICs vary widely by subtype, with VGIIc isolates generally having the highest values and VGI/VGIII isolates generally having the lowest values. There is a statistically significant difference in MICs between VGI/VGIII isolates and VGIIc isolates for four of the six antifungal drugs tested (Table (Table5).5). There was a statistically significant difference in MICs between VGIIa and VGIIc isolates for fluconazole, voriconazole, and posaconazole. Itraconazole was the only drug tested for which there were no statistically significant differences in MICs among subtypes.
We provide genotypic evidence that the strain dynamics of C. gattii isolates involved in the Pacific Northwest C. gattii emergence may be different from those seen on Vancouver Island, with VGIIb and VGIIc subtypes playing roles of equal significance to VGIIa types. We also show that human and animal infections caused by genotypes VGIIa, VGIIb, and VGIIc are occurring along the northern Pacific coast, now appear in California, and may have expanded eastward to Idaho. Because the described surveillance system is passive and not active, this study does not encompass the full scope of C. gattii disease in the Pacific Northwest United States, but it does provide a snapshot of human and animal cases that are occurring at this time.
There is no subtype bias in any one state in this study for either VGIIa or VGIIb, with each being found in Washington, Oregon, and California. This is the first report of the contemporary occurrence of genotypes VGIIa and VGIIb in California. As noted earlier by Byrnes and coworkers (4), we also found that the VGIIc subtype is unique to Oregon with the exception of the single isolate from Idaho. The lone case-patient from Idaho had a travel history to Oregon and Washington and might have acquired the infection in either state. However, it must be noted that the CAP59 allele from the Idaho isolate differs from all of the other VGIIc CAP59 alleles.
There were no significant differences in the distribution of subtypes between human and veterinary cases. Many of the veterinary cases had no history of travel, so it can be assumed that those cases arose from direct local environmental exposure. Because neither the strain types nor the MICs differed significantly between human and animal isolates, the animal cases may serve as a good sentinel for human cases in any particular area. Although we have no environmental isolates for comparison, earlier studies from both British Columbia and Washington state indicated that the genotypes of environmental isolates roughly correlated with the genotypes of patient isolates by the percentage of each genotype found (16, 21). It is interesting to note that no VGIII isolates were found during the environmental testing in British Columbia and Washington State (16, 21), and yet this type was not infrequent in either human or veterinary isolates in our study. Clearly, there is a need for more environmental testing in the Pacific Northwest United States.
This is the first analysis of antifungal susceptibility of C. gattii isolates from the ongoing emergence in the Pacific Northwest and the first time that MICs were compared among subtypes. While much research has been conducted on the antifungal susceptibilities of C. neoformans, very few studies have addressed the susceptibilities of C. gattii and none have compared the susceptibilities of specific subtypes. Only a few studies have used CLSI broth microdilution to report the susceptibilities of C. gattii (5, 6, 9, 24, 25, 27, 28), and none of these studies involved isolates from the Pacific Northwest emergence. As we found in our study, none of the earlier publications described elevated MICs to flucytosine or amphotericin B. Chen and coworkers (6) reported a higher geometric mean MIC of their 21 Taiwan isolates to amphotericin B than was found for the isolates in this study, but the mean was still within the range considered susceptible. Although we did not find any isolates with fluconazole MICs of >32 μg/ml, Trilles and coworkers reported fluconazole MICs of ≥64 μg/ml in their study of 57 C. gattii isolates from Brazil (28), although the proportion of their isolates with very high MICs was not given. Chen and coworkers also reported that 20% of the 18 C. gattii isolates from Australia that they tested had fluconazole MICs of ≥64 μg/ml (5). Our finding that 23.3% of the Pacific Northwest isolates have fluconazole MICs of 16 to 32 μg/ml is roughly double the 12.7% of isolates in this category reported from Spain (24) and reflects the findings of De Bedout and coworkers, who tested 15 C. gattii isolates from Colombia (9) and reported that 33% of their isolates had fluconazole MICs of 16 to 32 μg/ml. Voriconazole MICs in this study were relatively low, with the overall geometric mean MIC of 0.1 μg/ml and only a single isolate with an MIC of ≥1 μg/ml. Even so, our values are still higher than the geometric mean of 0.03 μg/ml reported for the isolates from Spain (24). Twenty-three percent of our isolates had itraconazole MICs that would be considered resistant for Candida species, but the clinical significance of these values for C. gattii is unknown.
The most interesting aspect of our study is that the MICs to fluconazole could be correlated with subtype. All of the VGI and VGIII isolates had comparatively low fluconazole MICs, while the majority of isolates with MICs of 16 to 32 μg/ml were of the subtype VGIIc. Clearly, there are differences between the subtypes, and these differences need to be further explored.
Unfortunately, what is not known is the relationship between elevated MICs and clinical outcome. While there have been no studies to date which correlate C. gattii MICs to clinical outcome, two recent studies of C. neoformans compared fluconazole MICs to clinical outcome (1, 8). The authors of these manuscripts reached opposite conclusions, with one study stating that among AIDS patients high fluconazole MICs may be predictive for patients who will not respond to fluconazole therapy (1), and the other study stating that among HIV-positive and -negative patients treated with fluconazole alone, there was no correlation between MICs and clinical success or failure (8). This is an area where more research is needed.
While fluconazole MICs of 16 and 32 μg/ml seem high compared to typical MICs for C. albicans and fluconazole, in the largest global study of C. neoformans susceptibility to date (1,811 human isolates) 44% of the isolates had an MIC to fluconazole of ≥8 μg/ml (26). Given this study and the studies cited above, it is clear that Cryptococcus species may have elevated MICs to fluconazole compared to values and breakpoints for Candida species. While we are not seeing MICs in the 64- to 256-μg/ml range, predictive of clinical failure for Candida species, VGIIb and VGIIc isolates from the Pacific Northwest have elevated MICs to fluconazole and comprise almost half of the collected isolates from the Pacific Northwest United States emergence. Unfortunately, very little is known about the clinical course and outcome for patients infected with these strains.
Treatment of C. gattii patients in the Pacific Northwest can be challenging, and these patients may have a difficult clinical course (31). In light of the difficulty in treating these patients and the 8.7% case fatality rate for these types of patients seen in British Columbia (13), it is prudent that physicians in the new areas of C. gattii emergence are aware that their patients may have an infection with C. gattii rather than with C. neoformans. It is also critical that we develop a better understanding of the correlation, if any, between MICs and clinical outcome. Public health surveillance for C. gattii infection, including tracking further spread of subtypes, may help us solve some of these problems.
We acknowledge Sky Blue, Christine G. Hahn, and Kris K. Carter in Idaho, Dawn Daly and Robert Vega at the Oregon State Public Health Laboratory, Hailu Kinde of the California Animal Health and Food Safety Laboratory, Cyndi Free at the Washington State Department of Health, and Dan Bradway at Washington Animal Disease Diagnostic Lab. We also thank Wieland Meyer for help with the allele assignment.
The findings and conclusions of this article are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Published ahead of print on 9 December 2009.