Genotyping of Candida albicans Oral Strains from Healthy Individuals by Polymorphic Microsatellite Locus Analysis

Abstract

Analysis of a polymorphic microsatellite locus was applied to 85 Candida albicans strains from healthy individuals. Comparison with strains from nonhealthy individuals previously analyzed in our laboratory showed an overall similarity, suggesting that all commensal strains have the ability to develop as pathogens.

The dimorphic fungus Candida albicans is a ubiquitous eucaryotic organism that develops as a saprophyte of the mucosa in humans. It is detected in one or more body locations in 70% of healthy women. In immunocompromised or intensive-care patients, the organism may overcome host defenses, resulting in increased mucosal colonization and eventually invasion into the bloodstream through epithelial and endothelial layers. Candidemia accounts for 8 to 15% of nosocomial bloodstream infections, and C. albicans is the causative agent in 50 to 70% of disseminated Candida infections (8, 9). Typing methods have confirmed the genetic similarity between C. albicans strains obtained from blood cultures and colonizing strains obtained from the same patients, confirming endogenous acquisition as the main source of dissemination (17, 20). Multilocus typing studies also demonstrated a strong linkage disequilibrium between independent markers, indicating that the population structure of C. albicans is primarily clonal (1, 11, 19, 23). According to this model, C. albicans comprises distinct lineages that propagate independently. If such clonal lineages have variations in biological traits, such as pathogenicity or host adaptation, identification of the corresponding genotypes should contribute to a better understanding of the natural history of candidiasis. With this aim, bloodstream and nonbloodstream C. albicans strains obtained from patients treated in a tertiary-care hospital were previously compared (7) by determining their allelic frequencies at a polymorphic microsatellite characterized in the promoter region of the elongation factor 3 (CEF3) gene (6). Although three major genotypes were identified in the study, they were overrepresented in both bloodstream and nonbloodstream strains, and no pathogenic genotype, i.e., no genotype with a propensity for bloodstream invasion, was identified (7). This finding was subsequently confirmed by a multilocus comparison of bloodstream strains and oropharyngeal isolates from human immunodeficiency virus-infected patients (12). However, both studies compared series of strains obtained from nonhealthy individuals. Therefore, they do not rule out the possibility that among commensal C. albicans strains, i.e., strains that normally develop in healthy individuals, some are more prone than others to undergo increased mucosal proliferation in nonhealthy individuals. To address this question, we collected a series of commensal C. albicans strains from healthy individuals. In the present paper, we report the population structure of these isolates at the CEF3 microsatellite and compare allelic frequencies with those previously reported at the same locus in bloodstream and nonbloodstream strains from nonhealthy individuals living in the same area.

Between June 2000 and May 2001, 700 oropharyngeal swabs were collected from healthy individuals presenting as outpatients in a dental practice or at occupational medicine departments in Dijon and Chatillon sur Seine, two cities in Burgundy, in northeast France. To ensure that no risk factors could lead to the selection of particular genotypes and that no saprophytic colonization by nosocomial strains had occurred, the following characteristics were considered exclusion criteria: age over 65; pregnancy; oral contraceptive use; antibiotic or antifungal treatment; hospitalization within 3 months prior to sampling; work in a hospital; immunodeficiency state, including iatrogenic immunosuppression; chronic disease, such as diabetes or heart, respiratory, or kidney failure; and long-term medical treatment. Denture wearers and individuals with past or present oropharyngeal candidiasis or oral disease were also excluded. Cultures of the samples were performed on Sabouraud dextrose agar, and strains were determined to be C. albicans by filamentation at 37°C for 4 h in human serum followed by determination of the sugar assimilation profile with an API 32 C kit (Biomerieux, Marcy L'Etoile, France). DNA isolation, PCR amplification, and fragment size analysis of the CEF3 microsatellite by automated fluorescent capillary electrophoresis were performed as previously described (7).

All together, 85 C. albicans strains were identified and processed for genotyping as described above. The frequencies of the CEF3 alleles in this collection are shown in Table 1 (group A). Thirteen different alleles, including a previously undescribed 156-bp allele, were found among the 85 C. albicans strains and organized in 19 distinct allelic combinations. Of these, five combinations (126-135, 131-131, 136-145, 130-136, and 133-144) together accounted for 67% of the isolates, whereas 14 genotypes were represented by one to four isolates only. Such organization of the yeast population in a few major and multiple minor genotypes has already been reported at this locus (5, 6, 7).

TABLE 1.

Comparison of allelic frequencies at the CEF3 locus in C. albicans strains from healthy individuals (group A [this study]) and in bloodstream isolates (group B) or nonbloodstream isolates (group C) from nonhealthy individuals

CEF3 allele association	Frequency in groupa
	Ab	Bc	Cd	Total for B and Ce
126-135	25 (29.4)	13 (27)	14 (29.2)	27 (28.1)
131-131	9 (10.58)	8 (16.7)	7 (14.6)	15 (15.6)
130-136	7 (8.23)	8 (16.7)	8 (16.7)	16 (16.7)
136-145	10 (11.76)	1 (2)	5 (10.4)	6 (6.25)
133-144	6 (7.05)	1 (2)	3 (6.25)	4 (4.2)
131-139	4 (4.7)	0	0	0
136-144	4 (4.7)	3 (6.25)	3 (6.25)	6 (6.25)
126-126	3 (3.52)	2 (4.2)	3 (6.25)	5 (5.2)
133-136	3 (3.52)	1 (2)	0	1 (1)
130-130	2 (2.35)	1 (2)	2 (4.2)	3 (3)
135-135	2 (2.35)	4 (8.3)	0	4 (4.2)
136-136	2 (2.35)	0	1 (2)	1 (1)
136-141	2 (2.35)	0	0	0
129-156	1 (1.17)	0	0	0
130-139	1 (1.17)	0	0	0
130-144	1 (1.17)	2 (4.2)	1 (2)	3 (3)
135-139	1 (1.17)	1 (2)	0	1 (1)
137-145	1 (1.17)	0	0	0
145-145	1 (1.17)	0	0	0
130-135	0	1 (2)	0	1 (1)
130-131	0	1 (2)	0	1 (1)
136-137	0	1 (2)	0	1 (1)
137-139	0	0	1 (2)	1 (1)

^aNumber of strains with the corresponding allele association (percentage of strains with the corresponding allele association within the group). Data for groups B and C are from reference 7.

^bHealthy; 85 strains.

^cBloodstream; 48 strains.

^dNonhealthy; 48 strains.

^eNonhealthy; 96 strains.

Allelic frequencies in the collection of isolates from healthy individuals (group A) were compared with those previously observed in our laboratory in bloodstream and nonbloodstream isolates from nonhealthy individuals (7), two series of strains identified as groups B and C in the present paper (Table 1). The overall genetic diversity at the CEF3 locus was slightly higher in group A than in the pooled groups B and C (Table 1) (19 genotypes in 85 group A isolates, i.e., an average 4.47 strains per genotype, versus 17 genotypes in 96 group B plus C isolates, i.e., an average 5.65 strains per genotype), a finding consistent with a previous study by Xu et al. comparing clinical isolates of C. albicans to isolates from healthy students (23). Comparison of allelic frequencies in group A and groups B and C (Table 1) showed that six minor allelic combinations (131-139, 136-141, 129-156, 130-139, 137-145, and 145-145) were found exclusively in group A. Similarly, four minor allelic combinations identified in group B or C (130-135, 130-131, 136-137, and 137-139) were not detected in group A. On the other hand, 13 genotypes were common to group A and group B and C isolates. The diversity of the profiles obtained from both healthy and nonhealthy individuals and the fact that most strains belonged to allelic combinations common to group A and groups B and C suggest that the ability to undergo increased mucosal proliferation in nonhealthy individuals is a widespread biological property, common to most C. albicans genotypes. The fact that few strains belonged to genotypes apparently specific to healthy or nonhealthy individuals could simply result from the limited size of the C. albicans samples tested.

Among the three main allelic combinations previously identified in groups B and C, i.e., 126-135, 131-131, and 130-136, genotype 126-135 was identified in 29.4% of group A isolates, a percentage almost identical to those reported in groups B (27%) and C (29.2%). The 126-135 allelic combination therefore identifies a yeast population that is clearly predominant as a commensal and in infections. Combinations 131-131 and 130-136, which accounted, respectively, for 10.58 and 8.23% of commensal strains, were underrepresented compared to groups B and C (Table 1). However, these two allelic combinations were still among the four most prominent genotypes identified in group A. Interestingly, variations of the same order were encountered among clinical isolates analyzed in two previous studies of the CEF3 microsatellite (19.35 to 30.1% for genotype 126-135, 15.07 to 19.35% for genotype 131-131, and 9.6 to 16.13% for genotype 130-136) (5, 6). In this context, the discrepancies between group A and groups B and C reported here could simply reflect sampling fluctuations due to the limited size of the collections rather than true variations in pathogenicity or host adaptation.

The aim of the present study was to determine whether all C. albicans commensal populations have the same ability to undergo mucosal proliferation and bloodstream invasion. We collected a set of C. albicans isolates from the oral cavities of healthy individuals. Strict exclusion criteria were applied to avoid risk factors for candidiasis and to minimize the risk of contamination of the oral flora by hospital strains. The allelic frequencies in this group of commensal strains were thus compared to those observed in bloodstream and nonbloodstream strains collected in patients treated in a tertiary-care hospital (7). The individuals sampled to establish these three collections were not related, and one isolate per patient was used to avoid artificial homogenization of the genotypes. All strains originated from individuals living in the same geographical area, thereby ruling out regional specificity as a cause of discrepancy (18). Genotyping of the three series of strains was performed in the same laboratory, using identical equipment and control isolates. Because the mode of reproduction of C. albicans is essentially clonal (1, 4, 10, 11, 15, 16, 19, 23), a unique marker was used. This approach was based on the assumption that in a clonal organism, allele combinations are linked across the genome. Therefore, variations in biological properties may be traced by only one variable marker (19). The rationale for a microsatellite-based approach was that microsatellite regions are codominantly inherited and allow the distinction of heterozygotes, which is critical in the case of the diploid yeast C. albicans (22). Moreover, PCR typing of microsatellites is discriminatory, robust, and highly reproducible. The CEF3 gene that we chose to target is a single copy in the C. albicans haploid genome (13), and variability at this locus cannot result from polymorphic copies interspersed in the genome. In addition, this marker is stable and has been characterized in terms of discriminatory power (6). However, several limits of the present investigation must be pointed out. On one hand, despite strong evidence for clonality, there is recent evidence that some recombination may also occur in C. albicans. If such is the case, analysis of loci associated with pathogenicity would provide more accurate data (10, 19). On the other hand, the present set of commensal strains and the hospital strains previously tested were not perfectly matched temporally. We must emphasize, however, that the stability of the CEF3 microsatellite has been demonstrated by subcultures corresponding to >300 generations in four reference strains representing four allelic combinations (6). Finally, size homoplasy, i.e., the fact that an allelic class could include alleles identical by descent (truly homologous) as well as alleles that achieved the same length via convergent evolutionary events (14), is a limit intrinsic to the microsatellite-based approach (19). Despite these limits, the diversity and overall similarity at the CEF3 microsatellite locus in C. albicans strains collected in commensal and pathogenic situations is a striking result. Our data indeed suggest that all commensal strains have the ability to develop as pathogens in nonhealthy individuals. This is consistent with the few studies we are aware of that have addressed the relationship between genotypes and commensalism or pathogenicity in C. albicans (3, 12, 21, 23). In this context, our contribution is to present data based on three large series of independent strains that were clearly characterized in terms of the host-parasite relationship (commensal, colonizing, or bloodstream invasive). Taken together, this body of information suggests that no firm correlation exists between genotypic groups, at least as defined by the available markers, and the position of the strain on the commensalism-pathogenicity spectrum. Interestingly, Bernhardt et al. recently showed that isolates from patients with severe clinical forms of candidiasis adhered to host cells and invaded reconstituted tissues in vitro (i.e., under conditions independent of host factors) to a greater extent than did commensal isolates (2). Distinct C. albicans populations thus exhibit distinct phenotypes of virulence that correlate with clinical profiles. The fact that no such correlation apparently exists with genotypic populations highlights the complexity of the gene regulation mechanisms involved in the pathogenicity of C. albicans.