Although most of the effort to elucidate the mechanisms of antifungal resistance has been expended in studies of C. albicans [81, 82], the results of these studies have directed similar work in C. glabrata [3, 5, 10, 23, 31, 42, 43, 57, 58, 61, 64, 65, 67, 68, 77–79, 83–85].

Each of the antifungal classes utilizes a different means to kill or inhibit the growth of fungal pathogens [82, 86, 87]. Mechanisms of antifungal resistance are either primary or secondary, and are related to intrinsic or acquired characteristics of the fungal pathogen that interfere with the antifungal mechanism of the respective drug or drug class or that lower target drug levels within the cell. Resistance can also occur when environmental factors lead to colonization or replacement of a susceptible strain with a resistant one. The antifungal effects of both polyene and azole antifungals are due to actions on the fungal cell membrane, while echinocandins act by disrupting the fungal cell wall [41, 55, 87]. Flucytosine acts as an antimetabolite to interfere with DNA and RNA synthesis [68].

Many parallels exist between the well-studied antibacterial resistance mechanisms and antifungal resistance mechanisms [88]; however, there are no data to suggest that destruction or modification of antifungal agents is an important component of antifungal resistance. Likewise, it does not appear that fungi can employ the genetic exchange mechanisms that allow rapid transmission of antimicrobial resistance in bacteria [88]. On the other hand, it is apparent that multidrug efflux pumps, target alterations, and reduced access to targets are important mechanisms of resistance to antifungal agents, just as they are important in antibacterial resistance [41, 81, 82, 86–88]. In contrast to the rapid emergence of high-level antimicrobial resistance that occurs among bacteria, antifungal resistance develops more slowly and involves the emergence of intrinsically resistant species (C. krusei, and to some extent C. glabrata, is intrinsically resistant to fluconazole) or stepwise alteration of cellular structures or functions that results in resistance to an agent to which there has been prior exposure [55, 81, 88]. Among the common species of Candida, C. glabrata appears to be unique in its ability to acquire and express drug resistance mutations in a relatively rapid time frame [3, 10, 23, 61, 67, 68, 77].

Resistance to amphotericin B, both deoxycholate and lipid formulations, remains uncommon among Candida spp. despite extensive utilization for over 50 years. Primary resistance to amphotericin B among clinical isolates of C. glabrata has not been described, although it has been noted that C. glabrata may exhibit decreased susceptibility to amphotericin B compared with C. albicans [89, 90] and may require higher doses of amphotericin B for optimal treatment [25, 91]. Notably, a recent report from France [7] found that increased use of amphotericin B in an ICU setting was followed by increased amphotericin B MICs for C. glabrata.

Our understanding of the mechanism of resistance to amphotericin B in C. glabrata stems largely from characterization of rare clinical isolates from patients failing amphotericin B therapy [10, 67, 85]. The mechanism of amphotericin B resistance appears to be from a quantitative or qualitative alteration in the sterol content of the cells [67, 85]. Accordingly, mutants of C. glabrata resistant to amphotericin B have been shown to have reduced ergosterol content and replacement of amphotericin B-binding sterols by intermediates that bind polyenes less well [67, 85]. Studies of sterol profiles of clinical isolates indicate a block in the C-24 sterol methyltransferase step with depletion of ergosterol and accumulation of Δ5,7-dienols [67, 85]. Vandeputte et al [67,85] described two different amphotericin B-resistant clinical isolates of C. glabrata, each of which contained a mutation in the ERG6 gene encoding C-24 sterol methyltransferase. In each instance the inactivation of C-24 sterol methyltransferase resulted in the interruption of ergosterol biosynthesis and thus polyene resistance. The changes in the sterol composition of the plasma membrane was thought to result in disturbances in the fungal cell wall or plasma membrane due to impairment of protein targeting. Transmission electron microscopy revealed structural modifications (thinning) of the cell wall of both mutant strains and pseudohyphal growth in one strain with a missense mutation in ERG6 [67, 85]. Both strains displayed a marked lack of fitness in vitro compared to WT strains suggesting that the ERG6 mutations constitute a selective disadvantage in the absence of selection pressure. However, given that the C. glabrata genome is haploid [69] and thus there is a higher probability of expression of a mutated allele, such mutations may be clinically relevant in the current therapeutic context, which is dominated by prophylaxis and empirical antifungal therapy [67, 85].

The azoles act by inhibiting the fungal cytochrome P450-dependent enzyme lanosterol 14α-demethylase, which is encoded by the gene ERG11. The enzyme converts lanosterol to ergosterol, and its inhibition disrupts membrane synthesis in the fungal cell [92]. Resistance can arise from a modification in the quality or quantity of the target enzyme, reduced access of the drug to the target, or some combination of these mechanisms [41,81, 86, 87]. In the first instance, point mutations in ERG11 lead to an altered target enzyme with decreased affinity for azoles. In addition to mutations in ERG11, overexpression of the gene results in the production of high concentrations of the target enzyme, creating the need for higher intracellular fluconazole (and other azoles) concentrations to inhibit the increased enzyme present in the cell [81, 82, 92]. The second major mechanism of azole resistance in Candida involves active efflux of fluconazole out of the cell through the activation of two types of multidrug efflux transporters: the major facilitators (encoded by MDR genes) and the ATP-binding cassette (ABC) superfamily (encoded by CDR genes) [81, 82].

It is now well-established that the primary mechanism of resistance to fluconazole in C. glabrata involves upregulation of the CgCDR1 and CgCDR2 genes resulting in resistance to multiple azoles [38, 42, 43, 93]. Borst et al [38] observed the rapid development of azole resistance (fluconazole, itraconazole, and voriconazole) following in vitro exposure to fluconazole among C. glabrata isolates that had never previously been exposed to azole antifungal agents. The resistance was stable and was associated with increased expression of CgCDR1 and CgCDR2, but notCgERG11. Likewise, Sanguinetti et al [43] performed a careful analysis of azole resistance profiles of clinical isolates of C. glabrata and found that the majority of the isolates were resistant to multiple azoles (fluconazole, itraconazole, ketoconazole and voriconazole) and that the resistance phenotypes were strongly associated with upregulation of the ABC efflux transporters CgCDR1, CgCDR2, and CgSNQ2. Interestingly, they too found no alteration or overproduction of CgERG11 among 20 different azole-resistant isolates. They concluded that cross-resistance is a very common feature in azole-resistant C. glabrata isolates, especially in those that are capable of expressing multiple mechanisms of resistance [43].

Although it has been suggested that alteration or overexpression of CgERG11 is not involved in azole resistance of C. glabrata [43], both Marichal et al [40] and Redding et al [93] have reported an increase in lanosterol 14α-demethylase in fluconazole-resistant C. glabrata. Marichal et al [40] found an increase in the amount of ERG11 mRNA transcript that was secondary to a duplication of the entire chromosome containing the gene, whereas Redding et al [93] found upregulation of CgERG11 along with that of CgCDR1 and CgCDR2 in a fluconazole-resistant strain of C. glabrata from a patient with oropharyngeal candidiasis (OPC) undergoing treatment with fluconazole. Taken together, the literature indicates that although efflux is a major mechanism of azole resistance in C. glabrata, the development of resistance in this species is a highly varied process involving multiple molecular mechanisms [93].

Echinocandins inhibit 1,3-ß-D-glucan synthase (GS) and thereby disrupt biosynthesis of 1,3-ß-D-glucan, a key component of the fungal cell wall. This causes the formation of a defective cell wall associated with cellular instability and lysis in yeasts such as C. glabrata [55, 61]. Mutations in the genes encoding elements of the GS complex have been associated with Candida resistance to echinocandins. In particular, reduced susceptibility or resistance of Candida to echinocandins has been linked with point mutations in two “hot-spot” (HS) regions of FKS1 (HS1 and HS2), the gene encoding for the major and presumed catalytic subunit of GS [55]. In C. glabrata, echinocandin resistance has also been associated with mutations in the FKS2 gene [19, 23, 61]. Whereas resistance to azoles in isolates of C. glabrata has been documented to be due to the overexpression of CDR efflux pumps (see previous section above), there is little or no evidence that efflux pumps are involved in resistance to the echinocandins [55, 84, 94].

Mutations in FKS proteins alter the GS enzyme kinetics resulting in significantly higher 50% inhibitory concentrations (IC50), as well as the kinetic inhibition constant (Ki), for the mutant enzymes when compared with corresponding enzymes from WT strains [61]. This pattern of decreased enzyme sensitivity to inhibition (increased IC50) extends across all three of the echinocandins, conferring resistance across the entire class. The case study of Chapeland-Leclerc et al (3) (MDR in C. glabrata) demonstrated point mutations in FKS2, associated with caspofungin resistance, as well as overexpression of CgCDR1 and CgCDR2, linked with fluconazole and voriconazole resistance.

The antifungal properties of flucytosine result from its conversion in the cytoplasm of fungal cells via the cytosine deaminase and the phosphoribosyltransferase into 5-fluorouridine triphosphate which incorporates into RNA and blocks protein synthesis or into 5-fluorodeoxyuridine monophosphate which inhibits the thymidylate synthase encoded by CDC21, a key enzyme of DNA synthesis [95]. Resistance mechanisms for flucytosine are well described among various species of Candida and include mutations in FCY2, FCY1 and FUR1 genes encoding for cytosine permease, cytosine deaminase, and uracil phosphoribosyltransferase, respectively. In studies of C. glabrata [68, 96], varying patterns of reduced susceptibility to flucytosine have been elucidated depending on the mutations present. Edlind and Katiyar (96) demonstrated that C. glabrata exhibited high-level flucytosine resistance (MIC, ≥32 μg/ml) associated with mutations in either FCY1 or FUR1 and moderately elevated MICs (MIC, 1 μg/ml) with mutations in FCY2. Vandeputte et al [68] conducted a study of laboratory mutants of C. glabrata obtained by exposure of a WT isolate to flucytosine. Mechanisms leading to flucytosine resistance were investigated in two mutant strains. One of the mutants exhibited a missense mutation in the FCY1 gene, encoding the cytosine deaminase, and a two fold reduction in the expression of the FUR1 gene, encoding the uracil phosphoribosyltransferase. The second mutant exhibited a missense mutation in FCY2, suggesting an altered function of the corresponding membrane transporter (cytosine permease), coupled with an overexpression of the CDC21 gene, which encodes thymidylate synthase, a key enzyme of DNA synthesis. Thus, although mutations in the FUR1 gene are the most common cause of resistance to flucytosine in Candida spp., other mechanisms may contribute to flucytosine resistance which may be easily acquired in C. glabrata, probably due to its haploid genome. Although the molecular events detected in the mutants described by Vandeputte et al [68] did not compromise their growth rate relative to the WT parent when assessed in independent cultures, they did represent a selective disadvantage in the absence of selection pressure since competitive growth experiments (co-culture) revealed a decreased fitness when compared with the parent isolate. Thus, without selection pressure, mutations leading to flucytosine resistance may have limited clinical consequences.

Over the past two decades, there has been a significant increase in the appearance of antifungal drug resistance in C. glabrata as a result of the increased use of various classes of antifungal agents, combined with the exceptional ability of this haploid yeast species to develop antifungal resistance [1, 3, 5, 7, 10, 13, 16, 20, 23]. In bacteria, antimicrobial resistance is often associated with fitness costs and thus results in a competitive disadvantage against otherwise drug-susceptible organisms within the host [71, 72]. Extension of this reasoning to fungi such as C. glabrata would suggest that antifungal resistance may have a fitness cost in this fungal species and thus may result in a counter-selection against resistant strains without drug pressure. Indeed decreased fitness has been described for C. glabrata strains with mutational resistance to amphotericin B [67, 85], the echinocandins [74, 76], and flucytosine [68, 96]. In some instances, however, resistance confers no change in microbial fitness or may even increase fitness [72, 73, 77–79]. Furthermore, initial decreases in fitness due to resistance may be restored by still unknown compensatory mechanisms, which can stabilize the resistant populations and render them as fit as susceptible organisms [72, 97].

Resistance to azole antifungal agents in C. glabrata is known to be mediated almost exclusively by enhanced drug efflux and overexpression of the ABC transporters [81]. Several genes encoding these transporters have been identified including CgCDR1, CgCDR2, and CgSNQ2 [38, 42, 43]. The expression of these genes is regulated by the zinc finger transcription factor CgPdr1p [98]. Single-point mutations in CgPDR1 (encodes for CgPdr1p) are known to increase the expression of CgCDR1 and CgCDR2 and thus contribute to the azole-resistant phenotype of C. glabrata [77, 83]. Recently Ferrari et al [77,78] demonstrated that these point mutations were gain of function (GOF) mutations since they conferred hyperactivity to CgPdr1p revealed by constitutive high expression of CgCDR1 and CgCDR2. These GOF mutations in CgPDR1, therefore, constitute the molecular basis of azole resistance in C. glabrata [77, 78,98]. Furthermore, Ferrari et al [77,78] found that resistant strains of C. glabrata were more virulent in mice and less susceptible to azoles in vivo as compared to WT strains, strongly suggesting a gain in fitness for the resistant isolates. The azole-resistant population effectively supplanted the WT population in the presence and absence of drug selection in vivo. The authors showed that it was the presence of GOF mutations rather than the presence of CgPDR1 that accounted for increased virulence. Subsequently, these investigators found that only two genes were upregulated by all GOFs, CgCDR1 and PUP1 (PDR1 UPregulated gene encoding a mitochondrial protein) [78]. Virulence and tissue burden studies demonstrated that these two CgPDR1-dependent and -upregulated genes contribute to the enhanced virulence of C. glabrata strains that acquired azole resistance [78].

Mitochondrial dysfunction is another of the mechanisms by which azole resistance can occur in C. glabrata [79]. Cells with mitochondrial DNA deficiency (“petite mutants”) upregulate the ABC transporter genes, CgCDR1, CgCDR2, and CgSNQ2, and thus display increased resistance to azoles. Notably, the upregulation of ABC transporters in mitochondrial mutants, while mediated by CgPDR1, is not dependent on GOF mutations in CgPDR1 [77,79]. Surprisingly, even with an in vitro growth deficiency compared to a WT strain of C. glabrata, a mitochondrial mutant strain was more virulent (as judged by mortality and fungal tissue burden) in murine models of infection [79]. The increased virulence of the petite mutant correlated with a dramatic gain of fitness in mice compared to the parental (WT) isolate. The molecular basis for this property is unclear, and the factors responsible for enhanced virulence need to be identified. Ferrari et al [79] did report that in addition to increased expression of CgPDR1 and other transcription factors in the petite mutant, some genes (CgYPS1 to −11) required for cell wall integrity, adherence to mammalian cells, virulence and survival in macrophages [69] were also upregulated. Cell wall alterations as a cause of increased virulence in mitochondrial mutants of C. glabrata is an intriguing possibility that warrants further study.

These findings highlight the need for carefully monitoring drug resistance in C. glabrata, since increased virulence may have a negative impact on the outcome of disease in infected patients. The gain in fitness associated with ABC transporter upregulation may favor the emergence of azole-resistant C. glabrata. In addition to GOF mutations in CgPDR1, several genes involved in stress response, resistance to DNA damage, and cell wall structure are upregulated in azole-resistant strains of C. glabrata, all of which may contribute individually or in combination to modulate the virulence of C. glabrata [77].

On a global scale, the proportions of HDC caused by C. glabrata vary from 22 to 26% in North American to 4 to 6% in Latin America [31,89]. Within the US, the proportion of HDC due to C. glabrata has been shown to vary from 11% to 37% across the nine US Bureau of the Census Regions [16, 99, 100] and from <10% to >30% within single institutions over the course of several years [101–103]. In a US multicenter survey spanning the years 1992 to 2007 [16], C. glabrata was shown to have increased as a cause of HDC from 18% of all bloodstream infection (BSI) isolates in the time period 1992 to 2001 to 25% in 2001 to 2007 (Table 1).

In general, the incidence of antifungal resistance among the fungal pathogens of humans is moderate, especially when compared with the incidence of antibiotic resistance among bacterial pathogens [88, 104]. Among the common species of Candida (C. albicans, C. glabrata, C. parapsilosis, and C. tropicalis), C. glabrata is unique for its decreased susceptibility to fluconazole and cross-resistance to other azoles [16, 25, 52,89]. Resistance to fluconazole among isolates of C. glabrata may vary from a low of 13.0% in the Asia-Pacific region to a high of 19.5% in North America [18]. In the USA C. glabrata has not only increased in frequency as a cause of HDC, it has also shown a concomitant increase in resistance to fluconazole from 9% in 1992 to 2001 to 14% in 2001 to 2007 [16] (Table 1).

Given the distinct differences between the mechanisms of action and resistance for the azoles versus the polyenes, echinocandins, and flucytosine [81, 86, 88], the lack of cross-resistance among the four classes is not surprising and supports the current recommendation that the echinocandins be used as first-line treatment of HDC, especially in patients with possible infection with C. glabrata [25]. The potent activity of the echinocandins against C. glabrata (and other species of Candida) has led some experts to suggest that if initial therapy is echinocandin-based, then any increase in non-albicans Candida species (e.g. C. glabrata) becomes largely irrelevant [21, 22]. This may be true given our current state of understanding; however, recent reports of acquired resistance to echinocandins, polyenes, and flucytosine in clinical isolates of fluconazole-resistant C. glabrata suggest that this species represents a major concern as a MDR pathogen [3, 5, 10, 13, 16, 18,19, 20, 23, 54, 57–60].

Data from global surveys demonstrate that the frequency of echinocandin resistance among clinical isolates of C. glabrata ranges from 1 to 3% and is higher among isolates from North America (3%) than among those from Europe (1%), Latin America (0.0%), or the Asia-Pacific region (0.0%) [105]. A recent US population-based survey conducted by the Centers for Disease Control and Prevention (CDC) found that approximately 3% of C. glabrata BSI isolates were resistant to one or more echinocandin [23]. In addition to geographic differences in echinocandin resistance among C. glabrata isolates, resistance may also vary according to the age and exposures of the infected patients. A study of 1,239 Candida BSI isolates from 79 medical centers conducted in 2008–2009 [17], found that resistance to both azoles and echinocandins was most prominent among isolates of C. glabrata, with the highest resistance rates to echinocandins (16.7%), fluconazole (16.7%), posaconazole (5.0%) and voriconazole (11.0%) among isolates from patients in the 20–39-year age group. This latter observation, coupled with case reports such as that of Chapeland-Leclerc et al [3], suggested that MDR may be an emerging threat in C. glabrata. These reports prompted us to review the experience of two large surveillance programs, the SENTRY Antimicrobial Surveillance Program for the years 2006 through 2010 and the CDC population-based surveillance conducted in 2008 to 2010 to assess the frequency of co-resistance to fluconazole and the echinocandins among clinical isolates of C. glabrata [19]. Among 1,669 BSI isolates of C. glabrata, 162 isolates (9.7%) were resistant to fluconazole (MIC, >32 μg/ml), of which 98.8% were nonsusceptible to voriconazole (MIC, >0.5 μg/ml), and 9.3%, 9.3%, and 8.0% were resistant to anidulafungin (MIC, >0.25 μg/ml), caspofungin (MIC, >0.25 μg/ml), and micafungin (MIC, >0.12 μg/ml), respectively. There were 18 fluconazole-resistant isolates that were resistant to one or more of the echinocandins (11.1% of all fluconazole-resistant isolates), all of which contained an acquired mutation in FKS1 or FKS2. By comparison, there were no echinocandin-resistant strains detected among 110 fluconazole-resistant isolates of C. glabrata from the years 2001 to 2004 (Table 2).

These data document the broad emergence of co-resistance over time to both azoles and echinocandins in clinical isolates of C. glabrata. Whereas resistance to azoles in isolates of C. glabrata has been documented to be due to overexpression of CDR efflux pumps [42, 43], there is little or no evidence that efflux pumps are involved in resistance to the echinocandins [84, 94]. The documentation of FKS mutations in isolates of C. glabrata showing in vitro resistance to both azoles and echinocandins suggests the sequential accumulation of acquired resistance mechanisms as demonstrated by Chapeland-Leclerc et al [3]. The lack of co-resistance in C. glabrata in the 2001–2004 time period is not surprising given that among the echinocandins only caspofungin was available, having been just approved in the US and Europe in 2001. Since that time, the overall use of echinocandins in the US has increased significantly, from 7.7±5.3 days of therapy (DOT) per 1,000 patient-days in 2004 to 13.1±8.6 DOT per 1,000 patient-days in 2008 (mean ± standard deviation) (P<0.001)[28]. During the same time period (2004 to 2008), the use of azoles increased as well, from 67.6±29 to 72±33 DOT per 1,000 patient-days, but this change was not significant (P=0.1520) [28].

Thus persistent and expanding antifungal drug pressure may predispose to both azole and echinocandin resistance in C. glabrata. These findings emphasize the importance of both rapid identification to the species level as well as antifungal susceptibility testing in patients with recurrent Candida isolation who are receiving or were previously exposed to antifungal treatment [21, 22, 25, 56].