Mitigation of Human-Pathogenic Fungi That Exhibit Resistance to Medical Agents: Can Clinical Antifungal Stewardship Help?

Claire M Hull; Nicola J Purdy; Suzy C Moody


Future Microbiol. 2014;9(3):307-325. 

In This Article

Weighing up the Threat of Fungi Exhibiting Resistance to Medical Antifungal Agents

In addition to the lack of horizontal gene transfer in fungi, it has previously been argued that any mutations that do confer resistance to medical agents are likely to be selected and retained at low frequency in human-pathogenic fungi.[14–16] Several explanations for this have been offered and provide the basis of the following appraisal.

Phenotypic Consequences of Point Mutations May Be Masked

C. albicans : A Case in Point? Since numerous human-pathogenic fungi are diploid, it has been reasoned (often with reference to C. albicans) that point mutations in one chromosome are likely to be compensated, in part, by the unaffected allele.[15] This explanation is conceptually sound and explains why the phenotypic and clinical outcome (antifungal resistance) of certain mutations may not always be observed. Nonetheless, it is somewhat paradoxical that over the last decade numerous studies of drug-refractory, pathogenic C. albicans isolated from patients in the clinical setting have demonstrated precisely how certain mutations can, and do, confer resistance to antifungal drugs. For example, in the ergosterol biosynthetic ERG genes, which encode enzymes essential for fungal sterol production, mutant isolates of C. albicans exhibiting cross-resistance to azole and polyene agents have been reported. These include strains harbor mutations in ERG3, which encodes sterol Δ5,6-desaturase,[26]ERG5 encoding C-22 sterol desaturase[27] and ERG11, which encodes the azole drug target (sterol 14α-demethylase).[28]

Other Clinically Important Yeasts. The impact of point mutations in the ERG genes in other clinically isolated species of Candida that exhibit poor sensitivity to azoles and polyenes have also been described and include C. glabrata (ERG11[29] and ERG2, which encodes C-8 sterol isomerase,[30] and ERG6 encoding C-24 sterol methyltransferase[31]) and, recently, C. tropicalis (ERG11 and ERG3).[32] Point mutations in genes that regVulate fungal cell wall manufacture and integrity including FUR1 (encoding uracil phosphoribosyltransferase) and FKS1/FKS2 (which encode putative subunits of 1,3-β glucan synthase) have also been linked to flucytosine (5-FC) and echinocandin resistance in C. glabrata.[33–35] This is certainly significant because C. glabrata has a haploid genome (thus the phenotypic effects of genetic alterations are readily seen) and is increasingly associated with the number of drug-refractory infections observed in the clinical setting.[3–5,17] The importance of haploidy for the propagation of resistance traits in A. fumigatus is also noteworthy and is considered hereafter (see section 'Human-pathogenic fungi do not reproduce sexually – at least in hosts'). Point mutations in ERG2 that result in reduced susceptibility to polyenes have been documented in a strain of Cryptococcus neoformans isolated from an patient with AIDS[36] and perturbations in sterol 14α-demethylase (ERG11) have been associated with fluconazole resistance in a recurrent clinical C. neoformans infection.[37]C. neoformans is intrinsically resistant to echinocandin antifungals, and fluconazole treatment is not completely curative,[38] thus, there is a clear need for specific antifungal stewardship guidelines and awareness among clinicians regarding the potential for acquired, multidrug resistance in this species.

Loss of Fitness & Genetic Reversion in the Absence of Selection Pressure

Fitness & Longevity. Mutations conferring antifungal resistance can result in a loss-of-fitness, such that mutants are observed to be weaker or less competitive than the parental strain in an antifungal-free environment. It has consequently been proposed that mutations conferring resistance do not favor the longevity of resistant strains.[15] The possibility of genetic reversion to wild-type, which can occur when selection pressure (e.g., from a specific antifungal drug class) is removed, is also an issue. Indeed, fitness costs have been documented in azole-resistant petite mutants (possessing respiratory defects) of C. glabrata[39] and associated with point mutations in several ERG genes, including ERG6[31] and ERG1 (encoding squalene epoxidase).[40] Of these, mutations in ERG3, which confer azole polyene cross-resistance, are perhaps the best documented in clinical isolates. It has been argued that mutations in ERG3 lead to reduced virulence in C. albicans[41] and that because erg3 mutants often exhibit hyphal growth impairment, their importance as agents of disease is negligible. However, aside from numerous accounts of clinically-isolated strains,[26] recent studies, in which C. albicans erg3 mutants have been observed to exhibit hyphal growth, provide grounds for further debate.[26,42] In fact, it might be argued that the mechanism through which mutations in ERG3 enable pathogenic C. albicans to circumvent the inhibitory effect of azole antifungals[43] presents a classic example of how conditional mutations may facilitate survival in a clinically-relevant context. Compensatory mutations in genes that regulate alternative cellular processes are also important. For example, it has been demonstrated how cumulative mutations that affect sterol biosynthesis in fungi – often resulting in synthetic lethality - can be suppressed by alterations in sphingolipid profiles.[44,45] Changes in sphingolipid content are understood to mediate echinocandin resistance[34] and lipidomic profiling has also demonstrated how dynamic changes in overall membrane structure can influence the antifungal susceptibility of C. glabrata.[46] Suppressor mutations that occur in antifungal-resistant clinical isolates underscore the complexity of molecular factors that regulate the fitness of human-pathogenic strains. These require fuller characterization because diagnostic tools for drug-resistant fungal infections (e.g., DNA screens for species-specific gene mutations) are lacking and are needed to enhance the efficacy and capacity of clinical antifungal stewardship in the future.

Survival & Virulence in the Human Host. A pathogenic ergosterol-deficient clinical isolate of C. glabrata harboring a missense mutation in ERG11 (historically understood to be an essential gene) and exhibiting facultative sterol uptake, slow flocculent growth and cross-resistance to azoles and amphotericin B (in vitro) has recently been documented.[29] From an 'in vivo' survival angle, it is entirely possible that the flocculent growth of this isolate compromised the efficiency of phagocytic clearance by the host immune system, as such, it was able to persist as a chronic systemic infection. Conversely, gain-of-function mutations in C. glabrataPDR1 (encoding a transcriptional regulator of multidrug transporters) have been proposed to mediate antifungal resistance and also enhanced virulence[47] by promoting colonization of the host through increased adherence to epithelial cell layers.[48] It is important to note that definitive explanations for the association between changes in fitness, virulence and antifungal resistance, which are often observed in clinical isolates harboring mutations, are rarely straightforward. For example, isolates of A. fumigatus exhibiting azole-resistance (induced during azole therapy) and increased expression of the cyp51A gene, have recently been reported in a fatal case of aspergillosis.[49] It remains unclear whether increased cyp51A expression accounted for the azole-resistant phenotypes or whether these isolates were exclusively responsible for invasive disease. However, despite being less virulent in a murine model than isogenic azole-susceptible and wild-type comparators, only azole-resistant A. fumigatus mutants were recovered from patient respiratory samples.[49]

Factors Outside the Clinical & Laboratory Setting. It is significant that mutations in A. fumigatus Cyp51A, found in environmental and clinical isolates, that confer multiazole resistance and are not associated with any fitness costs are an emergent concern.[19] This has been discussed in more depth elsewhere (see sections 'Human-pathogenic molds present in the environment' and 'Mapping the epidemiology of resistance: considerations & challenges'); however, the recovery of azole-resistant A. fumigatus from antifungal drug-naive patients (where prophylactic selection pressure from medical agents is absent) is very important because it highlights the potential for selection pressures exerted outside the clinic (and/or other unforeseen ones within[50]) to promote the development and conservation of antifungal resistance traits. Finally, it is noteworthy that despite measurements of fitness cost and instances of genetic reversion that are very often based on laboratory observations in vitro, the fact that drug-refractory human-pathogenic fungi harboring mutation(s) are frequently isolated from the clinic underscores the fact that these strains can and do persist as chronic infections in the human host. The biology, ecology and molecular determinants of resistance in these isolates demand further attention.

Human-Pathogenic Fungi Do Not Reproduce Sexually – at Least in Hosts

Cryptic Sexuality & Complex Lifecycles: Yeast. Until very recently, evidence for sexual reproduction in human-pathogenic fungi – especially within the human host – has remained limited. It has consequently been presumed that mutations conferring resistance in clinically relevant fungal isolates will not be expanded into other lineages through genetic exchange and recombination events.[15] However, gradual advances in our understanding of fungal biology, particularly new insights into the complexity of fungal lifecycles, suggest that traditional concepts concerning reproduction in fungi warrant closer evaluation. Approximately 8 years ago, it was recognized, for the first time, that the encapsulated yeast C. neoformans (ordinarily haploid nuclei) can undergo a process of diploidization to produce blastospores, which, following meiosis and recombination, yield haploid spores that can be dispersed.[51] This process, termed monokaryotic fruiting, is believed to promote recombinational repair in the oxidative, DNA-damaging environment of the host macrophage and it has been suggested that this may contribute to the virulence of C. neoformans in humans.[51] More recently, a cryptic mating cycle – termed same-sex mating – has been documented in C. albicans, a species once thought to be strictly asexual.[52] Parallels between same-sex mating in C. albicans and monokaryotic fruiting in C. neoformans have already been drawn[52] and have heightened speculation that same-sex mating could promote the survival of fungal infections within mammalian hosts. The extent to which genetic exchange during sexual and cryptic life processes occurs and could contribute to the survival and adaptation (including the acquisition of resistant traits) of other pathogenic fungi is unknown, and it is currently impossible to gauge the implications of these factors on the success of antifungal stewardship. The actual extent and clinical relevance of cryptic sexuality in human-pathogenic yeasts represents an important avenue for further research.

Cryptic Sexuality & Complex Lifecycles: Molds. For numerous environmentally ubiquitous molds ( Table 2 & Table 3 ), spore-formation is essential for both survival and dispersal and the impact of spore production on the environmental epidemiology of azole-resistant A. fumigatus[5,19,21] is currently attracting global interest (see sections 'Human-pathogenic molds present in the environment' and 'Mapping the epidemiology of resistance: considerations & challenges'). With this in mind, the very recent and first documentation of sexual cycles in both Aspergillus terreus[53] and Aspergillus lentulus[54] demands closer attention as Neosartorya fumigata (teleomorph of A. fumigatus) is also known to possess genomic elements that are required for sexual reproduction.[55] The specific environmental and biological factors that influence the onset and frequency of sexual reproduction in A. terreus and A. lentulus remain under investigation. However, it is tempting to speculate that under favorable conditions, the combination of sexual reproduction and subsequent spore formation could constitute a mechanism for the transfer of resistance traits in certain environmentally ubiquitous, human-pathogenic molds. It is noteworthy that, in terms of overall numbers, A. terreus currently accounts for fewer human-pathogenic infections than A. fumigatus, but it is an emergent concern that exhibits intrinsic resistance to amphotericin B.[20] Since amphotericin B is one of the few systemic antifungals that are effective against aspergillosis, resistance in any species/species complex should certainly be regarded as a warning against any temptation to discount the clinical relevance of A. terreus (and its potential to cause disease) either now or in the future. Importantly, A. terreus and A. lentulus occupy a worldwide distribution in soil and cause invasive aspergillosis in humans; here, molecular diagnostics that can differentiate between them and other closely related species are needed to inform antifungal stewardship and clinical treatment guidelines. Genetic screens for intraspecies resistance traits are also required; however, the design of these is a major challenge in itself. In the first instance, the mechanisms (including specific gene mutations) that govern antifungal resistance in the Aspergilli at large are not well documented. Moreover, aside from the recent documentation of sexual cycles in A. terreus and A. lentulus,[53,54] additional evidence for intra-kingdom horizontal gene transfer in fungi is steadily growing[56] and it is now suspected that the exchange of genes between fungi might happen much more frequently than conventionally assumed.[56] Undoubtedly, intra-kingdom horizontal gene transfer has the potential to impact the epidemiology of drug-resistance traits and could further complicate the design of diagnostics and antifungal stewardship protocols both now and in the future. Further research is now required to enhance our understanding of the ubiquity, ecology and biology of cryptic sexuality and sexual cycles in human-pathogenic molds.

Human-Pathogenic Fungi Lack Drug Inactivating Enzymes

Enzymes & Resistance. To date, enzymatic inactivation or xenobiotic metabolism of medical azoles has not been documented in human-pathogenic yeasts. This has provided the basis for further argument against the likelihood that antifungal resistant human-pathogenic fungi will emerge on a scale similar to drug-resistant bacteria.[14–16] However, for human-pathogenic fungi, the importance of alternative enzyme-dependant processes should not be overlooked.[17,18] Of these, drug efflux (and mutations that lead to increased export) mediated by ABC transporters and enzymes from the major facilitator superfamily are key to drug resistance in numerous fungi including the emergent human-pathogenic yeast C. glabrata.[47] Membrane-anchored enzymes, which are believed to mediate azole import via facilitated diffusion,[57] are also important. Here, it is possible that the alterations in cellular lipid composition and/or membrane order that are frequently documented in antifungal resistant yeast mutants could affect the distribution and function of export and import enzymes – providing an additional level of protection against the activity of antifungals.

Non-antifungal Drugs That Exhibit Fungicidal Activity. An important consideration concerning the 'lack' of antifungal drug inactivating enzymes in human-pathogenic yeast is the interpretation of what actually constitutes an antifungal. In addition to the official classes of medical agents ( Table 1 ), numerous non-antifungal clinical drugs (e.g., anesthetics, antipsychotics, statins, immunosuppressants, anticancer agents and steroid hormones) also exhibit antifungal activity.[50] The exposure of C. albicans to steroid hormones is understood to induce the upregulation of MDR genes, which can heighten resistance to alternative antifungal agents, and several well-known human-pathogenic molds (e.g., Aspergillus and Rhizopus spp.) possess enzymes that enable them to hydroxylate and so detoxify steroids. It is noteworthy that among the filamentous fungi, Cunninghamella spp, including the human pathogen Cunninghamella bertholletiae are able to metabolize a wide variety of xenobiotics (e.g., antidepressants, muscle relaxants, antipsychotics, antihistamines and anticholinergics) in the regio- and stereo-selective manners that are similar to those in mammalian enzyme systems.[58] The biotransformations performed by Aspergillus, Rhizopus and Cunninghamella spp. are already exploited commercially; however, they could have implications for the design and efficacy of future combination therapies that seek to repurpose existing clinical drugs.[59,60] At present, we simply do not know enough to disregard or understate the importance of the 'non-antifungal' drug degrading enzymes that are known to exist or have not yet been discovered in human-pathogenic fungi.