Cryptococcus neoformans Strains and Infection in Apparently Immunocompetent Patients, China

High incidence of cryptococcosis in these patients is in contrast to reports from other countries.


C ryptococcosis is caused by 2 species in the genus
Cryptococcus, C. neoformans and C. gattii (1). C. neoformans (serotypes A, D, and AD) is found worldwide and causes cryptococcosis most frequently in AIDS patients (2,3). C. gattii (serotypes B and C) is geographically restricted and is infrequently diagnosed in AIDS patients except in some areas of Africa (4). The most widely used approaches for cryptococcal strain typing include the following: PCR fi ngerprint analysis based on microsatellite DNA using M13 primers (5,6) and (GACA) 4 repeats (7), amplifi ed fragment length polymorphisms (8), sequence analysis of the rDNA intergenic spacer regions I and II (IGSI and IGSII) (9), and multilocus gene sequence typing (MLST) (10). These molecular approaches showed genetically distinguishable subgroups for each serotype. For example, phylogenetic analysis of the IGS1-5.8S rDNA-IGSII sequence showed 3 genotypes, 1a, 1b, and 1c, among the strains of serotype A collected worldwide (9). Likewise, PCR fi ngerprint patterns based on M13 microsatellite DNA identifi ed 2 major genotypes among the strains of serotype A. In this group, the VNI strain type was the most common genotype found worldwide (5). Furthermore, MLST showed 3 genetic subpopulations among strains of serotype A as well as subpopulations unique to certain geographic areas as was the case with strains from Botswana (VNB) in Africa (10). A C. neoformans genotype unique to Botswana showed epidemiologic importance of strain type. Rare genotypes of C. gattii have been reported from the recent outbreak of cryptococcosis on Vancouver Island in Canada (11). Finding C. gattii strains of a rare genotype in geographic areas outside of subtropical and tropical zones (12) further underscores the epidemiologic importance of this rare genotype.
In spite of extensive global strain typing, isolates from the People's Republic of China have rarely been included (5,8,9). Recently, data released by the Chinese Department of Health indicated that by October 2006, the total number of HIV patients and those identifi ed as AIDS patients was 183,733 and 40,667, respectively (www.chinacdc.net.cn/ n272442/n272530/n294176/n340510/15099.html).
From 1999 through 2006, a total of 224 articles on cryptococcosis have been published in Chinese journals; 2,850 case-patients were included. We cited 6 representative reports from these papers (13)(14)(15)(16)(17)(18). In these cases, only 69 strains (0.2 %) were associated with AIDS. The popu-lation structure of C. neoformans in China, however, has not been fully explored. In this study, we have used various molecular typing approaches to analyze the population structure of the cryptococcosis agent in China.

Serotype and Mating Type
The CryptoCheck kit (Iatron, Tokyo, Japan) was used for serotyping. Mating type was determined by crossing all the C. neoformans strains with B-3501 (MATα) and JEC20 (MATa) and C. gattii strains with NIH112 (MATα) and NIH198 (MATa) on V-8 agar (11,21). A few strains of C. neoformans that showed ambiguous mating results were analyzed by PCR using serotype A MATα and MATa allele specifi c primers of either the STE12 or STE20 gene (22).Yeast cultures were grown overnight in yeast extract glucose broth (0.5% yeast extract, 2% glucose) at 30°C. Genomic DNA was extracted from all strains as previously described (23).

Genotyping
Genotyping approaches included PCR and DNA sequencing. PCR fi ngerprint analysis was based on M13 primers for microsatellite DNA as well as primers containing sequence repeats of (GACA) 4 and URA5-restriction fragment length polymorphism (RFLP). Sequencing was performed for partial IGSI-5.8S-IGSII region of rDNA cluster, and MLST involving 8 genes located on 7 different chromosomes (10). PCR fi ngerprints using either the microsatellite-specifi c primer M13 (5′-GAGGGTGGCGGTTCT-3′) (5) or the (GACA) 4 sequence repeats (7) were generated as described. The URA5-RFLP profi les were generated after digestion of the PCR fragment containing the URA5 gene sequence with Sau961 and Hha1 (6). IGSI sequence analysis of the LrDNA gene was derived from PCR amplicon products (≈1.7 kb) generated by the primer pair combination LR11 (5′-TTACCACAGGGATAACTGGC-3′) and 5SR (5′-GGATCGGACGGGGCAGGGTGC-3′) (9). The amplifi cation reactions were carried out in microtubes at a fi nal volume of 50 μL. The PCR mix contained 50-100 ng of genomic DNA, 0.5 mmol/L of the forward and reverse primer pairs, 1.0 U DyNAzyme II polymerase, 1.5 mmol/L MgCl 2, and 200 mmol/L of each dNTP. The reaction was performed in an MJ Research PTC-100 thermocycler (GMI, Inc., Ramsey, MN, USA) and consisted of a denaturation step at 94°C for 1 min, followed by 40 cycles: 2 min of denaturation at 95°C, 1 min of annealing at 57°C, and 3 min of extension at 72°C. A fi nal elongation step was conducted at 72°C for 7 min. The amplicon product was purifi ed with QIAquick PCR Purifi cation Kit (QIAGEN, Valencia, CA, USA). Amplicon synthesis was confi rmed by agarose gel electrophoresis. Sequencing reactions used the forward primer IGS1F (5′-CAG ACGACTTGAATG-GGAACG-3′), located at positions 3613-3633 of the LrRNA region and the reverse primer IG2R (5′-ATG CAT AGA AAG CTG TTG G-3′), located at position 791 of the IGS1 region. Sequencing reactions were carried out with an ABI 3730 capillary sequencer (Applied Biosystems, Foster City, CA, USA) and followed the protocol described by Diaz et al. (9). Sequencing alignments of the IGS region were executed with MegAlign (DNASTAR, Inc., Madison, WI, USA) and visually corrected. Phylogenetic tree construction used PAUP* version 4.0 (24) with parsimony analysis (heuristic search, stepwise addition, random addition sequence, nearest neighbor interchange, 100 maximum tree). Reliability of the character was checked by using bootstrap analysis with 500 replications.

Construction of Dendrogram Based on M13 RFLP
To construct a dendrogram based on M13-generated DNA fi ngerprints, 12 strains representing the different provinces of China were chosen, along with 6 reference strains. The PCR product of the reference strain VNI was used twice in the panel (adjacent to the size marker and in the last lane). A total of 15 major bands that ranged in size from 3,054 bp to 506 bp were identifi ed across the lanes. Data were coded as 1 and 0; 1 represented the presence of a band and 0, its absence. This data matrix was coded into NEXUS format for input into the program PAUP* (24). PAUP* was used to generate a maximum parsimony phylogeny using a heuristic search algorithm. For the heuristic search, 10 random-addition replicates were performed. The starting tree for each replicate was obtained by stepwise addition. Seven replicates returned the same most parsimonious tree, which had a tree score of 18 steps. To quantify clade robustness, 500 nonparametric bootstrap replicates were performed. Bootstrap percentages >50% are indicated above the branches of the maximum parsimony phylogenetic tree. The tree fi gure for the maximum parsimony analysis was created using FigTree version 1.0 (http://tree. bio.ed.ac.uk/software/fi gtree).

MLST
For MLST analysis, DNA fragments of 8 unlinked genes that included CAP10, GPD, IGS1, LAC1, PLB1, SOD1, TEF1, and URE1, were amplifi ed by PCR (10) from 6 randomly chosen C. neoformans strains from China. DNA sequencing was carried out by using the dideoxy method, and sequences were compared with previously published sequences from the global collection of serotype A strains (10). To visualize the genetic relationships among different MLST genotypes, sequences were automatically aligned by using Sequencher 4.1 (Gene Codes Corp., Ann Arbor, MI, USA); the alignment was imported into MacClade 4.05 and edited manually. Because of the observed incongruence in the genealogies of several genes, combined sequence data for all 88 worldwide isolates (10) were analyzed with the neighborjoining method using uncorrected ("p") genetic distances. Statistical support for the phylogenetic groups was assessed by bootstrap analysis using 1,000 replicate data sets.

Results
The conditions of most patients (81.3%) were diagnosed at Shanghai Changzheng Hospital; the remaining patients (18.6%) were diagnosed at other hospitals. Patients without any recognizable predisposing factor for cryptococcosis such as HIV infection, malignancies, cirrhosis, organ transplantation, end-stage renal failure, autoimmune disorder, diabetes mellitus, idiopathic CD4 T-cell lymphopenia, sarcoidosis, chronic usage of corticosteroids or other immunosuppressive therapies, and any abnormal symptoms were regarded as patients "without apparent risk factors" (2). Since 2002, the period when isolates were obtained from 68% of the patients without apparent risk factors, HIV serologic testing and a battery of immunologic tests were performed on all cryptococcosis patients. Before 2002, each patient's cellular and humoral immune status was routinely determined. Any abnormality in these tests led to further HIV serology. The patients who had otherwise normal test results were also subjected to HIV serologic testing, the results of which were negative.
Of the 120 C. neoformans strains obtained from 16 provinces located in the middle to the eastern regions of mainland China (Figure 1), 84 (70 %) strains were isolated from apparently healthy patients and 27 (22.5 %) strains were isolated from patients with risk factors other than HIV infection (online Appendix Table). Only 9 (7.5 %) of the C. neoformans strains were isolated from AIDS patients. All 120 isolates were analyzed by PCR fi ngerprints using M13 primers, by URA5 RFLP patterns, and by (GACA) 4 sequence primers. Notably, all 120 strains of C. neoformans yielded an identical M13-based fi ngerprint pattern that could be distinguished from the reference types. In Figure 2, the M13 fi ngerprint patterns of 12 of the 120 strains are shown as examples along with the 6 reference strains. The major bands pattern for the serotype A isolates from China was more similar to the VNI than the other reference types (Figure 2, panel A). The URA5 RFLP (Figure 2, panel B) and (GACA) 4 patterns of 120 isolates, however, were identical to that of the VNI type (online Appendix Figure 1, available from www.cdc.gov/EID/content/14/5/755-appG1.htm).
The phylogenetic tree for maximum parsimony analysis showed the strains to be closely related to the reference strain VNI and the H99 strain ( Figure 3). The H99 strain had the same M13 fi ngerprint pattern as that of the VNI strain. Genotyping by sequence analysis of the IGSI-rDNA region indicated that the C. neoformans strains from China belonged to the C. neoformans genotype 1a (online Appendix Figure 2, available from www.cdc.gov/EID/ content/14/5/755-appG2.htm). Genotype 1a is the major genotype found among the serotype A strains of C. neoformans collected worldwide and follows a clonal pattern (9).
MLST, performed using 8 unlinked genes from 6 randomly chosen strains, showed identical sequences for the CAP10, GPD, LAC1, PLB1, SOD1, IGSI, TEF1, and URE1 genes. These results corroborate the homogeneity observed with various PCR fi ngerprint patterns. Notably, the Chinese strains of C. neoformans formed a cluster with 7 strains previously reported by Litvintseva et al. (Table) (10) and formed a distinct cluster, M5, that appeared to have diverged from the M1 genotype to which the VNI reference strain belongs (Figure 4). Two strains in the M5 that clustered with the Chinese strains, jp1086 and jp1088, were isolated in Japan and the M-13 fi ngerprint pattern of all 7 strains (Table) tested was identical to that of the Chinese strains (online Appendix Figure  3, available from www.cdc.gov/EID/content/14/5/755-appG3.htm). Our genotyping data indicates that M13-based PCR fi ngerprinting together with the MLST are powerful tools that enable discrimination of different strain types. The 3 Chinese C. neoformans strains randomly chosen to assess virulence were considerably less virulent than the H99 strain and moderately to signifi cantly more virulent than VNI and VNII reference strains ( Figure 5).
The H99 strain, the type strain of C. neoformans var. grubii (26), clustered with VNI strain based on the MLST data. The previous MLST tree, which did not place the H99 strain in the same cluster, A1 + A3 (= M1), as VNI was determined to be due to an error introduced during the sequencing process (10). The sequences from 8 genes that we sequenced matched 100% with the nucleotide sequence data posted in the H99 genomic database.
Nine isolates of C. gattii were all obtained from the eastern regions of China. Except for 1 strain from Shangdong, and 1 from Shanghai, all strains were recovered from provinces located south of Shanghai with a warm climate. Five (55%) of the 9 strains of C. gattii were from the Zhejiang province where Eucalyptus trees are commonly found. Of the 9 C. gattii strains, 2 were isolated from AIDS patients and 7 were isolated from otherwise normal patients. All of the C. gattii strains were serotype B, of MATα mating type with typical PCR patterns of the VGI type ( Figure 6). IGS sequences from 4 randomly chosen strains belonged to the genotype 4b (9).

Discussion
From 1985 to 2007, a total of 7,372 cases of cryptococcosis were reported in China. However, documentation on these patients is mostly unavailable, and fragmented documentation is available for only 1,999 cases. Among them, 323 (16.2%) cases seem to have occurred in patients with no underlying disease and who were considered otherwise healthy; 215 (10.8%) cases occurred in AIDS patients (Z. Yao et al., unpub. data).
This study provides a large scale population analysis of C. neoformans strains isolated from 129 patients with well-documented clinical cases of cryptococcosis that were treated at the Shanhai Changzheng Hospital. All patients were assessed for the common predisposing factors for cryptococcosis. Patients without any of these predisposing factors were regarded as "with no apparent risk factor." HIV serology was performed for all cases that have occurred since 2002. A total of 68% (56/82) of the patients had no underlying disease. Patient outcomes were monitored for 2 years following treatment at which time patients were determined to be cured of cryptococcosis.
Consistent with previous reports, all 120 clinical strains of C. neoformans isolated from China were serotype A and MATα (2). In contrast to the genotypic diversity of clinical C. neoformans serotype A strains found in other countries, such as Brazil, Australia, and the United States (5,27)  Chinese C. neoformans strains showed remarkable genetic homogeneity. This was evident not only in the patterns based on various PCR fi ngerprints but also by the lack of diversity in MLST results. A report from India also showed relatively low genetic diversity among 57 clinical C. neoformans isolates. However, a serotype D strain was found, indicating that the population structure of C. neoformans in India is less homogeneous than that in China (28). Remarkably, a majority of the Chinese strains caused cryptococcosis in persons without any recognizable immune defect or underlying disease. Although a considerable number of AIDS patients have been identifi ed, 8.5% of the 129 strains isolated across mainland China were recovered from AIDS patients; 71% were isolated from otherwise healthy persons. This is in stark contrast to the >80% AIDS-associated cryptococcosis cases reported in Europe and the United States and the 69% reported in Africa (29)(30)(31). Although markedly lower than the frequency in China, a relatively high number of cases of non-AIDS associated cryptococcosis due to C. neoformans were documented in Australia, New Zealand (32), and India (28). Most patients in Australia and New Zealand, however, were immunocompromised (32). The report from India showed that 41% of the cryptococcosis patients also had HIV infections; the remaining patients were determined to have no known immune defect (28). Information on the underlying diseases in these HIVfree Indian patients, however, was not recorded. We recognize the possibility that 71% of the Chinese cryptococcosis patients without apparent underlying disease may have had subtle defects in immunity that may have predisposed them to cryptococcosis. Alternatively, genetic factors may play an important role in the unusually high non-AIDS-associated cryptococcosis in China. A relationship between common functional genetic polymorphisms of the low-affi nity Fc gamma receptor genes, FCGR2A, -3A, and -3B, and the risk of cryptococcosis in HIV-uninfected patients was recently reported (33). It would be of interest to investigate the relationship between the genetic polymorphisms of the 3 genes in the immunocompetent Chinese cryptococcocis patients.
The M13 based PCR fi ngerprints of the Chinese C. neoformans strains were identical to each other and similar but distinguishable from VNI. For convenience throughout the discussion, we will hereafter refer to the M13 pattern of the Chinese strains as VNIc. The MLST-based phylogenetic tree also showed that the VNIc diverged from WM148 and formed a separate cluster with 7 previously analyzed strains (10). The 7 strains that cluster with the VNIc strains had originated from 3 different continents, which suggests that the VNIc type is not unique to China and apparently follows a widespread cosmopolitan distribution. We found that the 7 strains that cluster with Chinese strains have exactly the same M13 PCR pattern as the VNIc type. Whether most of these 8 strains were also from immunocompetent patients is not known. Since analysis of the IGSI region of Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 14, No. 5, May 2008 759  the VNIc strains showed it to be identical with WM148, the M13 fi ngerprint analysis appears to have greater discriminatory power in distinguishing the VNIc strains from the typical VNI type than IGSI analysis. The URA5 RFLP analysis or the (GACA) 4 -based PCR fi ngerprints also had lower resolution power than the M13 pattern analysis because the 2 methods could not distinguish VNIc from the WM148 strain. We do not know if VNIc is the most common type in China because the present study only analyzed clinical strains. Environmental strains should also be isolated and undergo molecular typing. Such a study would offer insight into the importance of the VNIc strain type in relation to infection in otherwise healthy persons in China. In general, the occurrence of cryptococcosis in patients without any apparent immune defect or underlying disease is relatively rare (2,34). In the United States, the incidence of cryptococcosis in immunocompetent patients has been estimated at 0.2/1,000,000 population per year in California (35); 0.9/100,000 in Atlanta, Georgia; and 0.93/100,000 in Alabama (36).
The presence of the VGI type in China is not surprising because it is the most common C. gattii strain type identifi ed in Southeast Asia and most of the C. gattii strains obtained were from the southeastern part of China. Although the total number of the C. gattii strains studied is small, the percentage of AIDS-associated C. gattii infections in China is higher than expected. Other than from certain geographic areas in Africa (4), C. gattii strains are rarely isolated from AIDS patients (2,32). In Australia where C. gattii is prevalent, for example, only 1 of 47 clinical strains of C. gattii was isolated from a patient with AIDS (32).
In our previous study, the urease-negative strain B-4578, which had been isolated from a cryptococcosis patient in China, was as virulent in mice as was the highly virulent H99 strain (23). Because the strain B-4578 was also of the VNIc molecular type, it was tempting to assume that strains isolated from immunocompetent patients in China would similarly be highly virulent in experimental animals. Indeed, the 3 strains of the VNIc type isolated from different provinces tested in mice were more virulent than the VNI reference strain. However, they were considerably less virulent than the H99 strain, which had been isolated from an immunocompromised patient (20). This suggests that a wide variation in virulence exists among the strains of the VNIc molecular type. Whether the degree of cryptococcal strain virulence manifested in mice is comparable to the human host remains unknown. Our experience with strain NIH12 has shown that virulence in mice does not necessarily correlate with that of the human host. The strain NIH12 is one of the most virulent serotype D strains tested in BALB/c mice (37); it only caused a chronic localized infection without dissemination in the human host. The patient infected by the NIH12 strain had sarcoidosis and a chronic, localized, osteomyelitis lesion later developed in the hip, which was cured by amphotericin B treatment without any dissemination or recurrence.
Because all 129 of the cryptococcal strains were isolated from Chinese, primarily immunocompetent, patients, one can ask whether any susceptibility difference to cryptococcosis is related to ethnicity. Australian studies have indicated a higher frequency of cryptococcosis in Aborigines (38) than in whites, and in Los Angeles, disease incidence was reported to be twice as frequent in Hispanics than in whites (39). However, data regarding ethnic differences in susceptibilities are scant and unconvincing. Since most of the 129 strains were isolated from non-AIDS patients in China, possible differences may exist in the reporting systems for cryptococcosis cases among AIDS patients and among non-AIDS patients. While some strains were isolated from patients in Henen, Yunan, and Xinjian provinces where HIV/AIDS was more prevalent, most strains were from patients in regions where HIV/ AIDS was not prevalent. Although the HIV status of some otherwise apparently healthy patients was unknown, they did fully recover after antifungal treatment. The outcome of the AIDS patients with cryptococcosis, however, is unknown because they were transferred to a quarantined fa- Figure 5. Virulence in mice. Mice were intranasally infected with 5 × 10 7 yeast cells from the 3 Cryptococcus neoformans Chinese strains, CHC114, CHC193, and CHC 186, and compared with H99 and the reference strains VNI and VNII for survival. cility soon after they were diagnosed to be HIV positive. To determine whether Chinese AIDS patients are more resistant to cryptococcosis than AIDS patients in other countries, differences in how AIDS patients are handled in China should be investigated.
In conclusion, this study reveals a strikingly homogeneous cryptococcal population belonging to a subtype of VNI in China. The high incidence of cryptococcosis in immunocompetent patients in China contrasts with reports from other countries. This work was supported by funds from the intramural program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health (K.J. Kwon-Chung). A.P. Litvintseva is supported by grants AI25783 and AI44975 from the National Institutes of Health.
Dr Chen is a physician and research scientist at the Department of Dermatology, Changzheng Hospital, which is affi liated with the Shanghai Second Military University. His main research interests are the epidemiology and clinical manifestations of cryptococcosis.