Prevalence of Clinical Isolates of Cryptococcus gattii Serotype C among Patients with AIDS in Sub-Saharan Africa

The Cryptococcus neoformans–species complex includes basidiomycetous yeasts that cause systemic infections in both immunocompromised and immunocompetent individuals [1]. However, most patients who acquire cryptococcosis have impaired cellular immunity and are infected with either C. neoformans var. grubii or C. neoformans var. neoformans; these varieties, which include serotypes A and D and AD hybrids, are globally responsible for 98% of all cryptococcal infections in patients with AIDS. Conversely, C. gattii which has been accorded species status [2], includes serotypes B and C, predominantly infects immunocompetent individuals, and, until recently, had not been reported in patients with AIDS

Little is known about the epidemiology and physiology of serotype C isolates. Most clinical serotype C strains were isolated from immunocompetent patients in southern California [3]; in addition, there are reports of the isolation of serotype C from immunocompetent patients in Canada [3], the United Kingdom [3], Thailand [4], Mexico [5], Brazil [6], and Colombia [7]. On the basis of molecular fingerprinting, all serotype C strains have been associated with either of 2 genetically isolated molecular types, VGIII and VGIV [5]. Most isolates possess the VGIII molecular type, and they have been isolated in the United States, India, and South America [5]; isolates of the VGIV type are extremely rare but are found in South Africa, India, Mexico, and Colombia [5, 8]

Recent reports have suggested that infection with serotype C strains may be associated with AIDS in patients in sub-Saharan Africa. In 2002, 4 cases of cryptococcal meningitis were reported from South Africa, and they were caused by C. gattii serotype C strains [9]. In addition, 8 patients with AIDS and cryptococcosis caused by C. gattii were reported from Rwanda, but the serotypes of these isolates were not indicated [10]

We examined isolates of Cryptococcus from specimens of spinal fluid from 161 Botswanan patients and 15 Malawian patients, all of whom had AIDS. As expected, most of these isolates were C. neoformans var. grubii serotype A, but 22 (13.7%) of the 161 Botswanan and 2 (13.3%) of the 15 Malawian strains were serotype C. All of the serotype C strains have the rare VGIV molecular type and possess the MATα mating-type allele. By analyzing amplified fragment-length polymorphisms (AFLP) and sequences of the intergenic spacer (IGS) region of rDNA, we discerned 10 distinct genotypes among the serotype C strains. However, there was little genetic diversity among these strains, which implies that they most likely descended from a single lineage

Subjects, materials, and methods We examined 161 C. neoformans strains that were isolated, between January 1999 and May 2002, from specimens of cerebrospinal fluid (CSF) from Botswanan patients and 15 strains that were isolated, during August and September 1997, from specimens of venous blood from Malawian patients [11]. Isolates were colony purified, confirmed to be C. neoformans by standard morphological and physiological criteria, and maintained on yeast extract–peptone-dextrose agar medium (Difco) at 30°C. Isolates were serotyped by use of commercial monoclonal antibodies (Mabs) (Iatron) and by polymerase chain reaction (PCR) amplification using serotype-specific primers and AFLP genotypes (see below). For comparison, we used 3 reference strains from the American Type Culture Collection—ATCC 34880, ATCC 34883 (=MMRL 1332), and ATCC 32608 (=MMRL 1334); 3 strains from the Medical Mycological Research Laboratory (MMRL) of Duke University Medical Center—MMRL 1343, MMRL 2650, and MMRL 2651—which were gifts from Arvind A. Padhye; and reference strains VGIII and VGIV, which were provided by Teun Boekhout and Wieland Meyer [5]. For mating assays, we used the following laboratory-test serotype C strains: MMRL 1334 MATa MMRL 1343 MATa and MMRL 1332 MATa. Additional non–serotype C strains used in the phylogenetic analysis are described in the legend to figure 1

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Genetic diversity among Cryptococcus species, on the basis of 96 amplified fragment-length–polymorphism markers, presented in a neighbor-joining dendrogram with Nei-Li genetic distances. The uppermost clade includes the C. gattii isolates reported in the present study: genotypes C1–C9 are serotype C strains isolated from Botswana, and genotype C10 was isolated from Malawi; numbers in parentheses after each genotype indicate the number of isolates that have that genotype. The letters (A–D and AD) above each branch indicate the serotypes of the isolates composing the respective clades. The numbers below each branch indicate the statistical support for each clade and are based on 1000 bootstrap recalculations of the data. The sources of the other strains are as follows: MMRL 2650 and MMRL 2651 are clinical isolates of serotype C from India; MMRL 1343 is a clinical isolate of serotype C from New York; MMRL 2544, MMRL 2560, MMRL 2572, and MMRL 2573 are clinical and environmental isolates of serotype B from Australia; ATCC 32608 and ATCC 34880 are clinical isolates of serotype C from southern California; NCA1–CNA7, NCAD1–NCAD5, and NCD1–NCD5 are clinical and environmental isolates of serotypes A, AD, and D, respectively, from North Carolina [13]; bt1, bt157, and bt65 are clinical isolates of serotype A from Botswana [12]; and VGI, VGIII, and VGIV are C. gattii–typing strains [5]. “MMRL” strains were obtained from the Medical Mycological Research Laboratory, Duke University Medical Center; “ATCC” strains were purchased from the American Type Culture Collection; and the “VG” typing strains were provided by Teun Boekhout (Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands)

Genomic DNA was extracted from each isolate, and AFLP markers were generated as described elsewhere [12]. Polymorphic bands were defined as bands of the same size that were present in some but not all isolates. To assess the reproducibility of the AFLP method, DNA was extracted and the AFLP reactions and analyses were performed on at least 3 separate occasions, for each isolate. In comparisons of replicate analyses, 98% of the AFLP bands were identical (data not shown). Only intense and reproducible bands were scored for the analyses of population structure. For each strain, the IGS1 region between the large subunit of rDNA and the 5S genes was obtained by PCR and was sequenced [12, 13]

The mating type of each strain was identified by PCR using the following mating type–specific primers, which amplified portions of either the STE3a allele or the STE3α allele: (1) STE3a—forward, 5′-ACCTTTGCGGTTTCATCAAC; reverse, AAGGTCGCATGGGTAATGAG; and (2) STE3α—forward, 5′-TAACATTGGACATCCCAGCA; reverse, 5′-GAAGACGCAGGGTACAGCTC. The conditions for these amplifications were as follows: STE3a—94°C for 5 min, followed by 30 cycles of 94°C for 30 sec, 56°C for 30 sec, and 72°C for 1 min, followed by extension at 72°C for 7 min; and STE3α—94°C for 5 min, followed by 30 cycles of 94°C for 30 sec, 62°C for 30 sec, and 72°C for 1 min, followed by extension at 72°C for 7 min. Each reaction contained PCR buffer, 2 mmol/L MgCl₂, 0.2 mmol/L each deoxyribonucleoside triphosphate, 1 μmol/L each primer, 0.065 μL Taq DNA Polymerase (Invitrogen), and ∼1 ng genomic DNA. To test for sexual reproduction and to determine mating types, the strains were crossed, in the laboratory, with MMRL 1343 serotype C MATa and with MMRL 1332 serotype C MATα, as described elsewhere [12]

Genetic relatedness among the isolates was evaluated by 2 methods: (1) nonmetric multidimensional scaling analysis using Euclidian-distance measures and Community Analysis Package 2.4 (PISCES Conservation) and (2) Nei-Li genetic distances for restriction-fragment data, followed by cluster analysis using the neighbor-joining algorithm in the Phylogenetic Analysis Using Parsimony (PAUP) program [12, 13]

Phylogenetic analyses of the IGS1-sequence data were also performed by use of PAUP. Maximum-parsimony trees were identified by use of heuristic searches based on 100 random sequence additions for each data set. To evaluate the association among loci in each sample, we used the index of association (I _A) and the new unbiased estimate of multilocus linkage disequilibrium (r _d) [12, 13]; both the I _A values and the r _d values were calculated by use of Multilocus 1.2 software, and 1000 artificially recombined data sets were used to determine the statistical (bootstrap) values for the test [12, 13]

Results and discussion Among the spinal-fluid isolates of Cryptococcus from 161 Botswanan patients with AIDS and the blood isolates from 15 Malawian patients with AIDS, we identified 24 C. gattii serotype C strains. The strains were serotyped by use of commercial Mabs (Iatron), and their serotypes were confirmed by serotype-specific PCR and AFLP analysis [12]. The Botswanan sample consisted of 139 isolates (86.3%) that were serotype A and 22 isolates (13.7%) that were serotype C; the Malawian sample consisted of 12 isolates (80%) that were serotype A, 2 isolates (13.3%) that were serotype C, and 1 isolate (6.7%) that was a serotype AD hybrid strain. None of the isolates were serotype B or serotype D. The serotype C strains are listed in table 1

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Cryptococcus gattii serotype C strains

All 24 serotype C isolates were encapsulated, produced melanin on niger-seed agar, and grew at 37°C [1]. When cultured on canavanine–glycine–bromothymol blue agar, 23 of the 24 serotype C isolates changed the color of the medium from yellow-green to bright blue

The mating type of each strain was identified by PCR using mating type–specific primers that amplified portions of either the STE3a allele or the STE3α allele, and the results were confirmed by mating assays [12]. Characteristic 600-bp amplicons were obtained by use of the primers specific to the STE3α allele but not by use of the primers specific to the STE3a allele (data not shown). In mating assays, only 1 of the strains, bt91, was fertile

We used AFLP analysis with 2 independent primer pairs to investigate the genetic diversity of the isolates [12, 13]; 10 polymorphic AFLP bands distinguished 10 AFLP genotypes, and these genotypes were used to determine genetic relationships among the isolates (figure 1). DNA sequence analysis of the IGS1 region revealed that all but 1 of the isolates, bt21, had identical IGS1 sequences (data not shown), a result that confirms that AFLP genotyping is more discriminatory than is single-locus gene sequencing [12]

Using PCR fingerprinting, Meyer et al. defined 2 genetically distinct molecular types of serotype C, which they designated “VGIII” and “VGIV” [5]. Our AFLP and IGS1-sequence analyses of serotype C strains also delineated 2 distinct molecular types of serotype C, and the serotype C isolates from Botswana and Malawi that we delineated are closely related to the VGIV molecular type (figure 1). VGIV strains are rare, and reported sources of VGIV include an immunocompetent patient in Mexico and environmental samples from Colombia [5, 7]. In addition, Ellis et al. reported the isolation of VGIV strains from South Africa and India, although they did not indicate either the number or the source of these strains [8]

Analyzing the clone-corrected data set, we detected statistically significant linkage disequilibrium among the alleles (I_A=0.73; P<.01), which suggests that there is significant clonality among the isolates. A low level of genetic diversity among serotype C strains was also observed in a multigene phylogenetic analysis of 6 strains [14]

With few exceptions, C. gattii serotype B or C strains are typically found in tropical and subtropical areas of the world and are usually associated with primary infections in immunocompetent patients [1]. One of the exceptions is the recent cluster of infections caused by serotype B that were found in immunocompetent persons in Vancouver, British Columbia; the vast majority of these cases were caused by serotype B strains with an identical genotype and the α mating type [15, 16]. Compared with cases of cryptococcosis caused by serotypes A and D, those caused by serotypes B and C exhibit more pulmonary and cerebral nodules, increased neurological morbidity, and slower response to antifungal therapy [17]. Our investigation has revealed that a significant proportion of opportunistic cryptococcal infections in Botswanan and Malawian patients with AIDS are attributable to serotype C strains, which may explain, in part, the substantial gravity of cryptococcal meningoencephalitis in sub-Saharan Africa. Because these strains were recovered from either spinal-fluid or blood specimens from patients with AIDS, we do not know the extent to which they may cause nonneurological or non-AIDS–associated cryptococcosis in Botswana and Malawi

This newly discovered association with meningeal cryptococcosis caused by C. gattii serotype C in patients in Botswana, Malawi, and South Africa who have AIDS may be a localized phenomenon. Indeed, no additional serotype C strains were discovered in a preliminary screening of 13 clinical isolates from Tanzania, 21 from Uganda, and 5 from the Democratic Republic of Congo (data not shown). However, it would be surprising if these strains were not more widespread in sub-Saharan Africa

Because of the high prevalence of HIV infection in Botswana and Malawi, a large percentage of their populations continue to be at risk for infection with indigenous Cryptococcus strains. The present report demonstrates that C. gattii serotype C strains are clearly more abundant in Africa than had previously been thought [9]. In the natural environment, C. gattii serotype B has been associated with a variety of trees, particularly with species of Eucalyptus [1], and several such species—including E. camaldulensis, E. tereticornis and E. sideroxylon—are found in Botswana. However, the ecological reservoir of serotype C has not been determined; there is only a single report that describes the isolation of C. gattii serotype C from the environment—an almond tree (Terminalia catappa) in Colombia [7]. Environmental isolates will be necessary for investigation of the source of isolates of serotype C from patients in Africa

• Received December 23, 2004.
• Accepted March 30, 2005.

• © 2005 by the Infectious Diseases Society of America