Correlation between Azole Susceptibilities, Genotypes, and ERG11 Mutations in Candida albicans Isolates Associated with Vulvovaginal Candidiasis in China

Abstract

The relationship between susceptibilities to fluconazole and itraconazole and microsatellite CAI genotypes were examined from a total of 154 Candida albicans isolates (97 isolates causing vulvovaginitis in Chinese women and 6 vaginal isolates and 51 oral cavity isolates from asymptomatic carriers). The two dominant genotypes, CAI 30-45 (45 isolates) and CAI 32-46 (33 isolates), associated with vulvovaginitis showed significantly different azole susceptibility patterns with strong statistical support. CAI 32-46 isolates were usually less susceptible to both fluconazole and itraconazole than CAI 30-45 isolates and than the oral isolates with other diversified CAI genotypes. Remarkably different mutation patterns in the azole target gene ERG11 were correspondingly observed among C. albicans isolates representing different genotypes and sources. Isolates with the same or similar CAI genotypes usually possessed identical or phylogenetically closely related ERG11 sequences. Loss of heterozygosity in ERG11 was observed in all the CAI 32-46 isolates but not in the CAI 30-45 isolates and most of the oral isolates sequenced. Compared with the ERG11 sequence of strain SC5314 (X13296), two homozygous missense mutations (G487T and T916C) leading to two amino acid changes (A114S and Y257H) in Erg11p were found in CAI 32-46 isolates. The correlation between azole susceptibility and C. albicans genotype may be of potential therapeutic significance.

Vulvovaginal candidiasis (VVC) is a common vaginal infection, affecting up to 75% of women of child-bearing age at least once in their lifetime (7, 21, 22). The most frequent cause of VVC is Candida albicans, which is responsible for 70 to 90% of vulvovaginitis cases. Non-C. albicans species of Candida, predominantly Candida glabrata, are responsible for the remainder of cases (21). On the basis of the severity of symptoms, frequency, and causative agents, VVC is usually classified as either uncomplicated (mild and sporadic) or complicated (recurrent, severe, or caused by non-C. albicans species) (7, 21). Ten to 20% of women suffer complicated VVC in their lifetime (21). When properly diagnosed, uncomplicated VVC may be treated easily and reliably. However, complicated VVC often causes long-term physical and mental discomfort, significant economic burden from treatments, and considerable negative effect on sexual relations (21-23).

At present, prolonged suppressive therapy using fluconazole is recommended as the standard management for chronic, recurrent Candida vulvovaginitis (23). Therefore, there is a great concern about the emergence and spread of azole resistance of C. albicans isolates associated with VVC. Indeed, susceptibility testing of VVC-causing isolates has been performed in different countries and regions of the world (1, 2, 4, 5, 6, 13-15, 17, 18, 20, 24). Although relatively high frequencies of fluconazole- and/or itraconazole-resistant C. albicans isolates causing VVC have been observed in a few reports (13, 20, 24), most studies failed to identify any clear correlation between azole susceptibility and VVC association among C. albicans isolates (1, 2, 4, 5, 6, 14, 15, 17, 18).

Recently, we compared the genotype distribution patterns among independent C. albicans isolates associated with VVC in Chinese women and those from various extragenital sites by using the polymorphic microsatellite locus CAI (8, 11). The results showed that the CAI genotypes of C. albicans isolates from extragenital sites were highly diversified. In contrast, isolates associated with VVC from unrelated patients were more homogeneous and belonged to only a few genotypes, with two genotypes, CAI 30-45 and CAI 32-46, being the most common. These two dominant genotypes were rarely found among isolates from extragenital sites (11). In addition, the distribution of the dominant genotypes correlated positively with the severity of VVC (8, 11). These results suggested that C. albicans isolates with genotypes CAI 30-45 and CAI 32-46 might be more virulent and/or more resistant to the commonly used azole drugs than those with other genotypes as causative agents of vaginal infection.

Antifungal susceptibility testing using the Etest method revealed that the C. albicans isolates causing VVC in Chinese women were generally susceptible to fluconazole, amphotericin B, ketoconazole, and flucytosine; however, 19.1% of the isolates could be interpreted as being resistant to itraconazole in vitro. Interestingly, most of the itraconazole-resistant isolates belonged to a specific genotype (13). Contrary to the report described above, recent susceptibility testing and microsatellite typing of vulvovaginitis-causing Candida isolates from Europe did not find an association between azole resistance and any particular genotype cluster among C. albicans isolates (1). In the present study, fluconazole and itraconazole susceptibilities of the C. albicans isolates with the dominant genotypes CAI 30-45 and CAI 32-46 from VVC patients were compared with those of isolates possessing other minor genotypes and of isolates from the oral cavity by using the standard broth microdilution method. Furthermore, ERG11 (encoding lanosterol-14-α-demethylase, the target of azoles) gene sequences of C. albicans isolates representing different genotypes and sources were determined. The correlation between azole susceptibilities, genotypes, and ERG11 mutations was examined.

MATERIALS AND METHODS

C. albicans isolates.

A total of 154 independent C. albicans isolates, including 97 vaginal isolates from VVC patients in Beijing, Guandong, and Wuhan, located in northern, southern, and central China, respectively, and 6 vaginal isolates and 51 oral cavity isolates from healthy volunteers in various regions of China, were employed for comparative study. The CAI genotypes of the isolates were determined as described by Sampaio et al. (19).

Antifungal susceptibility testing.

In vitro fluconazole and itraconazole susceptibility testing was performed using the broth microdilution method following the guidelines outlined in NCCLS (now Clinical and Laboratory Standards Institute [CLSI]) document M27-A2 (16). The trays were incubated at 35°C for 24 h. The MICs were determined as the lowest antifungal concentration with a prominent decrease in turbidity (50% inhibition in growth) relative to growth in the antifungal-free control well. C. albicans ATCC 90028 and C. parapsilosis ATCC 22019 were employed as quality controls. The MIC values were interpreted according to the NCCLS/CLSI interpretive breakpoints (16) as follows: the breakpoints for susceptibility were ≤8 μg/ml for fluconazole and ≤0.125 μg/ml for itraconazole, the breakpoints for dose-dependent susceptibility were 16 to 32 μg/ml for fluconazole and 0.25 to 0.5 μg/ml for itraconazole, and the breakpoints for resistance were ≥64 μg/ml for fluconazole and ≥1 μg/ml for itraconazole.

ERG11 sequence analysis.

The ERG11 gene was amplified and sequenced essentially using the methods described by Lee et al. (10). Phylogenetic relationships among the ERG11 sequences determined were analyzed. For the polymorphic sites at which heterozygosity occurred, each base was rewritten twice for a homozygous (A, C, G, or T) datum or once each for the two component bases for a heterozygous (K, M, R, S, W, or Y) datum. The modified sequence alignment was analyzed by the unweighted-pair group method with arithmetic means (UPGMA) by using MEGA version 4 (25). The numbers of nucleotide differences were used as the substitution model, and bootstrap analyses were performed on 1,000 random resamplings.

Nucleotide sequence accession numbers.

The ERG11 sequences of C. albicans isolates determined in this study have been deposited in GenBank (accession numbers HM194154 to HM194227).

RESULTS

CAI genotypes.

From the 51 oral isolates, 27 CAI genotypes were identified. The most common genotypes were CAI 17-21 and CAI 18-27, which were shared by six (11.8%) and five (9.8%) isolates, respectively. The 103 vaginal isolates represented 21 CAI genotypes (see the legend of Fig. 1). For the consideration of clearer comparison and discussion, the vaginal isolates were further classified into five groups (V1 to V5) based on their CAI genotypes. The majority of the isolates belonged to either group V2 (genotype 30-45, 45 isolates) or group V4 (genotype 32-46, 33 isolates). Isolates in group V1 (11 genotypes, 12 isolates) possessed CAI genotypes similar to those of the oral isolates, while isolates in groups V3 (5 genotypes, 9 isolates) and V5 (2 genotypes, 4 isolates) possessed CAI genotypes similar to those of isolates in group V2 or V4. A clear difference in CAI genotype distributions between the oral and vaginal isolates was that the CAI alleles of the majority of the oral isolates were usually shorter than 30 trinucleotide repeats while those of the majority of the vaginal isolates were usually longer than 30 trinucleotide repeats.

FIG. 1.

Fluconazole and itraconazole susceptibilities of C. albicans isolates with different origins and CAI genotypes. The oral (O) isolates were separated from the vaginal (V1 to V5) isolates and are listed individually from left to right according to the lengths of their CAI alleles (the shorter allele has the priority if the locus is heterozygous in an isolate). They are listed as follows: O, 51 isolates with CAI genotypes 11-11, 11-20, 11-21 (2 isolates), 16-16 (2 isolates), 16-17 (2 isolates), 16-20, 16-21 (3 isolates), 16-26 (2 isolates), 17-17, 17-21 (6 isolates), 17-27, 18-18, 18-26 (2 isolates), 18-27 (5 isolates), 18-29, 20-23, 21-21 (3 isolates), 25-25 (2 isolates), 25-31, 25-34 (3 isolates), 26-26 (3 isolates), 26-27, 26-34 (2 isolates), 27-27, 32-35, 33-33, and 36-36; V1, 12 isolates with CAI genotypes 11-21, 17-21, 18-26, 21-21 (2 isolates), 23-32, 23-35, 24-24, 25-33, 28-41, 29-45, and 30-39; V2, 45 isolates with the same CAI genotype, 30-45; V3, 9 isolates with CAI genotypes 30-46, 30-47 (3 isolates), 31-32 (2 isolates), 31-45 (2 isolates), and 32-39; V4, 33 isolates with the same CAI genotype, 32-46; and V5, 4 isolates with CAI genotypes 32-47 (3 isolates) and 33-46. a, MIC = 16.0 μg/ml; b, MIC = 1.0 μg/ml.

Azole susceptibility.

All the oral C. albicans isolates were susceptible to fluconazole (MIC = 0.125 to 1 μg/ml). Among the vaginal isolates, two with genotypes CAI 30-39 (group V1) and CAI 32-46 (group V4), respectively, were dose-dependently susceptible to fluconazole (both MICs were 16 μg/ml). The remaining vaginal isolates were susceptible to fluconazole (MIC ≤ 4 μg/ml). Interestingly, the fluconazole MIC values of isolates from groups V3 to V5 were usually higher than those of isolates from groups V1 and V2 (Fig. 1).

A similar trend was found in itraconazole susceptibility testing. All the oral isolates tested were susceptible to itraconazole (MIC = 0.0625 to 0.125 μg/ml). Among the vaginal isolates, the two isolates that were dose-dependently susceptible to fluconazole could be interpreted as being resistant to itraconazole (MIC = 1 μg/ml) according to the NCCLS/CLSI interpretive breakpoints (16). Generally higher itraconazole MIC values for isolates in groups V3 to V5 were also observed (Fig. 1). The proportions of isolates with itraconazole MICs of ≥0.25 μg/ml (dose-dependently susceptible to resistant) in groups V1 to V5 were 16.7% (2/12), 2.2% (1/45), 44.4% (4/9), 84.8% (28/33), and 100% (4/4), respectively (Fig. 1).

The geometric mean (GM) values for fluconazole and itraconazole MICs of the oral group and the five vaginal groups of C. albicans isolates are summarized in Table 1. The GM value for fluconazole MICs of group V4 was significantly higher than that of group V2 (P < 0.001) and that of the oral group (P < 0.001). Consistently, the GM value for itraconazole MICs of group V4 was also significantly higher than that of group V2 (P < 0.001) and that of the oral group (P < 0.001). Group V2 and the oral group were similar in their GM values for fluconazole (P = 0.90) and itraconazole (P = 0.63) MICs. Because the numbers of isolates in each of groups V1, V3, and V5 are limited and they represent only minor genotypes among C. albicans isolates associated with VVC in China (11), the statistical significances of their azole susceptibility differences from each other and from the other groups are not presented.

TABLE 1.

Azole susceptibilities of oral and vaginal C. albicans isolates determined by the CLSI M27-A2 broth microdilution method

Groupa	No. of isolates	Geometric mean MIC50 (μg/ml)
		Fluconazole	Itraconazole
O	51	0.356	0.074
V1	12	0.375	0.111
V2	45	0.218	0.058
V3	9	0.735	0.107
V4	33	2.043	0.255
V5	4	3.364	0.297

^aSee the legend of Fig. 1 for the origin and CAI genotypes of each group.

ERG11 mutations.

A total of 74 isolates representing different CAI genotypes and origins were selected for ERG11 sequence comparison. Multiple isolates for each of the two dominant genotypes, CAI 30-45 and CAI 32-46, from groups V2 and V4 were randomly selected. The whole open reading frame of the ERG11 gene was determined for each isolate. C. albicans isolates with the same or similar CAI genotypes usually possessed the same or similar ERG11 gene sequences and thus were clustered together in the phylogenetic dendrogram with strong bootstrap supports (Fig. 2). In correspondence with the high degree of CAI genotype variation, the oral isolates exhibited more diversified ERG11 sequences than the vaginal isolates (Fig. 2). When the ERG11 sequence of strain SC5314 (GenBank accession no. X13296) was employed as the standard for comparison, a total of 30 mutations (with an average of 10.2 mutations per isolate) were found from the 34 oral isolates analyzed. However, most of the mutations were silent, and only one or two missense heterozygous or homozygous mutations were detected from 13 (38.2%) of the oral isolates compared.

FIG. 2.

Correlations between ERG11 sequence mutations, sources, CAI genotypes, and Erg11p mutations of C. albicans isolates. The dendrogram was drawn from UPGMA analysis of ERG11 sequences. The branch lengths are in scale to the numbers of nucleotide differences, and the bootstrap percentages over 50% from 1,000 bootstrap replicates are shown. The homozygous and heterozygous missense mutations in ERG11 are differentially highlighted. A few other silent mutations sporadically occurred in ERG11 sequences of isolates SZ169 (C488T and G1557G/A), XZ271 (T606G and G630A), 4242 (A935G), XZ219 (T1404T/C and G1497G/A), and XZ35 (A1594G) are not listed. O, oral cavity; V, vagina. Periods indicate conserved bases or residues.

The vaginal isolates with different CAI genotypes exhibited quite different ERG11 sequence mutation patterns. Isolates with CAI 32-46 or similar genotypes from groups V3, V4, and V5 showed identical ERG11 sequences which differed from the standard sequence by only two nucleotides (G487T and T916C), which corresponded to homozygous missense mutations, leading to two amino acid changes (A114S and Y257H) in the enzyme Erg11p. No heterozygous site was found in ERG11 sequences of these isolates. Except for isolate SZ169, which exhibited two silent mutations, all the isolates with CAI 30-45 and similar genotypes showed 9 or 10 heterozygous mutations in their ERG11 sequences (Fig. 2). For the two component bases at each of these 9 or 10 heterozygous sites where mutations occurred, one base remained unchanged compared with the standard sequence. Among the changed bases, two were missense (Fig. 2).

The differences in ERG11 mutation numbers of CAI 32-46 isolates from those of CAI 30-45 isolates and those of oral isolates are all statistically significant (P < 0.001). Though the ERG11 mutation numbers of the CAI 30-45 and oral isolate groups were similar (P = 0.15), the exclusively heterozygous mutation pattern found in CAI 30-45 isolates was specific. Consequently, the vaginal isolates were clearly separated from the oral isolates in the dendrogram (Fig. 2).

DISCUSSION

The results of this study suggest that different genotypes of C. albicans may have significantly different azole susceptibilities. The data obtained in this study indicated that the C. albicans isolates with genotype CAI 32-46 (group V4) were usually less susceptible or more tolerant to fluconazole and itraconazole than isolates with genotype CAI 30-45 (group V2) and isolates from oral cavities of asymptomatic carriers. The C. albicans isolates in groups V3 and V5 which possess similar CAI allele lengths to isolates in group V4 usually exhibited similar azole susceptibility to this group. The correlation between similar azole susceptibilities and similar CAI allele lengths was also found for isolates from groups V1 and V2. Since the numbers of isolates in groups V1, V3, and V5 tested were relatively small, the degree of correlation between similar CAI allele lengths and similar azole susceptibilities needs to be tested further using more isolates.

The azole susceptibilities of the C. albicans isolates correlated well with ERG11 mutations. The isolates with the CAI 32-46 genotype and higher azole MICs exhibited exclusively two homozygous missense base substitutions in ERG11, leading to two amino acid changes in its encoded protein, Erg11p (Fig. 2). It is interesting that the same two exclusive ERG11 missense mutations were also observed in 14 out of the 15 fluconazole-resistant C. albicans isolates from different clinical sources in China (27). The two point mutations together with another mutation, T395A (F83Y in Erg11p), occurred in the ERG11 gene of a fluconazole-resistant C. albicans strain (MIC = 256 μg/ml), induced by growth in a gradually increased concentration of fluconazole in vitro, relative to that of the starting sensitive strain (9). Another special feature of the ERG11 sequences of isolates with the CAI 32-46 genotype is the loss of heterozygosity (LOH), compared with the vaginal isolates possessing CAI 30-45 genotypes and with the oral isolates. LOH of ERG11 has been correlated with reduced susceptibility to fluconazole in C. albicans (5, 26).

The correlation between microsatellite CAI polymorphisms and ERG11 mutation patterns in C. albicans as shown in Fig. 2 is also worthy of note. Since the two loci are located on different chromosomes of C. albicans (CAI locus is on chromosome 4, while ERG11 is on chromosome 5), their correlated variation might imply more significant genetic differentiation. Statistical analysis showed the significant differences between ERG11 nucleotide mutation numbers for different groups of isolates with different CAI genotypes. Molecular phylogenetic analysis based on ERG11 sequences revealed, from another aspect, that isolates with the same or similar CAI allelic length profiles were usually clustered together with strong bootstrap supports, while isolates with different CAI profiles were separated into different clusters (Fig. 2). The dendrograms drawn from other algorithms of phylogenetic analysis including neighbor joining and maximum parsimony (data not shown) showed the same topology as that presented in Fig. 2. The clustering of C. albicans isolates based on ERG11 sequence analysis is in agreement with that shown in a previous study of ours based on sequence analysis of three housekeeping genes (11). The correlated variations of different gene sequences with CAI genotypes confirm the merit of C. albicans strain typing based on CAI polymorphism.

The observations of this study differ from those of Antonopoulou et al. (1), who did not find any association between azole-resistant isolates and microsatellite genotype clusters among C. albicans isolates causing vulvovaginitis in women from two European regions. One reason might be that the strategy of genotyping used by Antonopoulou et al. (1) was different from that used in our study. Different genotyping methods may result in different groupings of isolates. For example, inconsistent with our observations based on CAI polymorphism (8, 11), genotyping of vaginal C. albicans isolates by using randomly amplified polymorphic DNA (RAPD) analysis failed to show any relationship between genotypes and VVC patient groups (3, 12). Another possible reason is that the genotype distribution patterns of VVC causing C. albicans isolates in Chinese and European women may be different. Sampaio et al. (19) determined the CAI genotypes of 23 isolates from Portuguese patients with recurrent VVC. Ten genotypes were identified from these isolates, but none of the independent isolates shared the same or similar genotypes. The dominant genotypes found in isolates from Chinese women with VVC did not exist in the Portuguese isolates. The difference in genotype distribution patterns of VVC-causing isolates from women in different regions of the world or belonging to different ethnic groups need to be clarified by comparing a larger number of cases and isolates.

The difference in susceptibilities toward the two common azole antifungal drugs between C. albicans isolates of genotypes CAI 30-45 and 32-46 is of potential diagnostic and therapeutic significance. These two genotypes are the most frequent causative agents of VVC in China and are often positively associated with severe and recurrent infection (8, 11), suggesting that the two genotypes possess enhanced virulence in vaginal infections and are more likely to induce complicated VVC. The significant differences in susceptibilities to fluconazole and itraconazole between the two dominant genotypes suggest the potential for strategic consideration of target treatment for patients with vulvovaginitis. Specifically, VVC caused by C. albicans isolates of genotype CAI 30-45 might be more effectively treated with azoles than the disease caused by genotype CAI 32-46 isolates. Conversely, patients with VVC caused by isolates of genotype CAI 32-46 might need a higher-dose azole treatment or a non-azole-drug treatment.