Beyond resistance: antifungal heteroresistance and antifungal tolerance in fungal pathogens

Abstract

Fungal infections are increasing globally, causing alarmingly high mortality and economic burden. In addition to antifungal resistance, other more subtle drug responses appear to increase the likelihood of treatment failures. These responses include heteroresistance and tolerance, terms that are more well-defined for antibacterial drugs, but are also evident in pathogenic fungi. Here we compare these antifungal responses with similarly named antibacterial responses, and we review recent advances in how we understand the routes by which antifungal heteroresistance and tolerance emerge.

Introduction

The first list of fungal ‘priority pathogens’ was released by the World Health Organization in 2022, with a warning that some species are increasingly drug-resistant and becoming more widespread [1]. Indeed, the incidence of fungal infections is rising globally. This is due, in part, to the growing number of immune-suppressed populations that are susceptible to such infections and, in part, due to the high proportion of antifungal treatment failures which result in more than 1.5 million deaths annually [2]. Treatment failures may occur because of poor drug penetrance into infected organs, weakened patient immunity that normally enhances drug efficacy, and fungal drug responses that enable growth in the presence of the drug. The latter include antifungal drug resistance, tolerance and heteroresistance. Candida albicans is the most prevalent cause of fungal infections in Western hospitals and reports of drug-resistant C. albicans drug isolates are relatively rare [3]; the increase in clinical drug resistance is driven largely by the increased prevalence of non-albicans Candida species that are intrinsically drug-resistant. Nonetheless, the high treatment failure frequency (30-50% mortality) is far more frequent than the prevalence of drug resistant isolates, and we suggest that heteroresistance and tolerance to antifungals may explain some of this discrepancy.

Fungi are eukaryotes and share conserved metabolic and regulatory pathways with their mammalian hosts, complicating the search for effective antifungal drug targets. In addition, some pathogenic species are intrinsically resistant to at least one of the three clinically available drug classes (azoles, echinocandins and polyenes). We posit that antifungal drug resistance, heteroresistance and tolerance contribute to treatment failures and infection recurrence [4–6]. This review will define antifungal drug resistance, heteroresistance and tolerance and discuss what we know about intrinsic and acquired drug responses, the molecular mechanisms that drive them, and how we might prevent the emergence of these responses in pathogens.

Main text of review

Consensus definitions of resistance, heteroresistance, and tolerance are listed in Table 1. For fungal pathogens of humans, these terms are not used consistently. The definitions for these antifungal drug responses also differ for different drugs and different pathogenic species.

Table 1.

Definitions resistance, heteroresistance, and tolerance for antibiotics and antifungals.

Antibacterials	Antifungals
Resistance--definition An organism is resistant when its MIC is above the clinical breakpoints as established by standardization organizations, e.g., CLSI [1] and EUCAST [2]. In laboratory experiments, resistance is defined as having a higher MIC than a control or ancestor strain.
MIC is the lowest concentration that completely inhibits growth of the organism.	For amphotericinB: lowest concentration with >=90% inhibition relative to drug-free control. For flucytosine, azoles and echinocandins: lowest drug concentration with >=50% of inhibition relative to drug-free control.
Measured using dilution assays or agar diffusion assays.	Measured using dilution assays or agar diffusion assays.
Heteroresistance--definition A small subpopulation of cells displays a higher MIC than the majority of susceptible cells
Instability possibly from unstable tandem gene amplification.	Instability often due to unstable aneuploidy or copy number variation.
Measured using PAP assays	Measured using PAP assays
Tolerance--definitions
The ability to survive longer times of bactericidal drug exposure without an increase in MIC.	Tolerance to echinocandins: For Candida glabrata and Candida parapsilosis,: The ability to survive fungicidal echinocandin drugs for a longer time without increased MIC.
	Measured as minimum duration to killing in time-kill assays.
	In other reports, for C. albicans and C. parapsilosis, the ability to grow better than a wild-type control (but slower than in the absence of drug) in the presence of echinocandins.
	Tolerance to azoles: For Candida species, the ability of subpopulation of a fungus to grow slowly at concentrations of fungistatic azoles that are above the MIC, yet the MIC is generally not altered.
Measured as minimum duration of killing in time-kill assays.	Measured as supra-MIC growth in liquid assays or fraction of growth in agar assays. Both parameters measure the proportion of the maximum growth.

Antifungal drug resistance

For both bacteria and fungi, drug susceptibility is measured as the minimum inhibitory concentration (MIC) of a drug; the term resistance describes the ability to grow above the MIC, with clinical breakpoints determined for some, but not all, species-drug combinations [7,8]. Resistance is often due to heritable genetic mutations that alter interactions between the drug and its cellular target (e.g., Erg11/Cyp51, the azole target or Fks, the echinocandin target) and enable relatively good growth despite the presence of the drug. Like bacteria, fungal resistance mutations are usually due to genetic mutations that alter the drug target or decrease the intracellular drug concentration via changes in efflux [9]. For the nucleoside analogue class of drugs, which are prodrugs, resistance arises rapidly during monotherapy, usually via loss-of-function mutations that inhibit prodrug uptake and intracellular processing [10]. Thus, the paradigm established for bacterial drug resistance generally holds for fungal drug resistance.

Antifungal drug heteroresistance (HetR) is the ability of a small proportion of cells in a population to grow at concentrations above the MIC, while most cells do not grow. In yeast pathogens, HetR to azoles refers to a small subpopulation of cells that acquire resistance, i.e., a higher MIC. Once acquired, these cells tend to remain resistant in a transient manner that can return (in a subset of cells) to the completely susceptible state. The degree of transience differs in different HetR strains. HetR has been reported in Cryptococcus neoformans and Cryptococcus gattii, environmental basidiomycetes that cause serious infections in immune-compromised patients. Recent studies indicate that fluconazole HetR is prevalent in both clinical and environmental isolates of C. neoformans and C. gattii [11,12].

In vitro, HetR in Cryptococcus neoformans was associated with aneuploidy, specifically an extra copy of Chr1, which carries ERG11 encoding the target of azole antifungals, and AFR1 encoding an efflux pump important for reducing intracellular azole concentrations [13]. Importantly, in a clinical study of HIV/AIDs patients in Tanzania, HetR was frequent in recurrent infections that had been treated with fluconazole monotherapy. However, patients treated with a combination of fluconazole and flucytosine did not exhibit HetR, suggesting that a combination therapy could reduce the frequency of fluconazole HetR and treatment failure [14].

In in vitro selection studies with C. neoformans, HetR can be acquired within 48h of exposure to sub-MIC fluconazole. Notably, most HetR isolates also acquired either whole chromosome 1 disomy or, more rarely, a ~62Kb duplication of Chr1 that included ERG11, and the amplification of the small region was sufficient to cause HetR [15]. Furthermore, the acquisition of FLC HetR was associated with the acquisition of cross-tolerance to amphotericin B and/or flucytosine [15].

HetR was also reported in Candida glabrata and Candida parapsilosis [11,16–19], ascomycetes that are increasingly prevalent amongst infections causing persistent and recalcitrant infections. In C. glabrata, HetR to fluconazole was associated with increased clinical persistence of a C. glabrata infection in a mouse model of systemic candidiasis [19]. More recently, echinocandin HetR (at frequencies of 0.01% to 1.7%), was seen in C. parapsilosis [6] in ~75% of bloodstream isolates from micafungin-breakthrough candidemia. The HetR subpopulations had micafungin MICs 32- to 64-fold higher than the susceptible progenitor cells in the same population. Intestinal colonization with C. parapsilosis that are micafungin HetR was correlated with the development of breakthrough bloodstream infections [6], providing strong support for the idea that HetR may be important clinically.

The frequency with which HetR arises (≥ 1/10,000 cells) is consistent with the idea that HetR can arise via changes in DNA copy number or allele frequency (due to defects in chromosome segregation or mitotic recombination).

However, aneuploidy is not the only route to heteroresistance: some HetR C. neoformans isolates had no detectable aneuploidies or copy number changes [13,14]. Similarly, in micafungin-HetR C. parapsilosis isolates, no point mutations in the FKS genes, which confer resistance to echinocandins, were detected [6], suggesting that HetR can be acquired via different routes that remain to be characterized. These could include non-genetic changes (e.g., chromatin modifications [20], metabolic shifts [21] or changes in the biophysical states of molecular interactions [22]. Another option is that HetR could emerge by downregulating the production of extracellular vesicles, which carry factors that increase the ability of biofilms to grow in drug [23]. In C. albicans, extracellular vesicles also enable drug resistance in biofilms by producing the extracellular matrix that sequesters the drug [24] and can promote cross complementation for biofilm factors between individuals and different Candida species [25] and can transmit virulence factors in C. neoformans [26]. Future studies will likely reveal more mechanistic details about these intriguing unconventional routes to HetR.

Antifungal tolerance

In bacteria, antibiotic tolerance is a response to bactericidal antibiotics and is defined as the ability of a susceptible strain to survive drug exposure for longer periods of time. Tolerant cells die more slowly in drug (have a longer minimum duration of killing (MDK) than most susceptible cells [27]).

For antifungals, the MDK is only relevant when microbes are treated with fungicidal drugs. The degree to which antifungal drugs are fungicidal or fungistatic differs with the drug administered and the species responding to the drug. For example, most azoles are fungistatic against most Candida species, yet are fungicidal against Aspergillus species; conversely, echinocandins are fungicidal against Candida species, yet fungistatic against Aspergillus species.

Antifungal drug tolerance also depends upon the drug and the species being tested. For example, tolerance to echinocandins, which are considered fungicidal in C. glabrata, improves the survival of a very small subpopulation of cells (10^-4-10^-5) over a wide range of drug concentrations [28], and thus resembles the situation with bactericidal drugs in bacteria.

C. albicans isolates exhibit different levels of intrinsic azole tolerance whose properties are different from antibiotic tolerance. First, azole tolerance is usually due to a subpopulation (usually 10-90% of cells), with the subpopulation size being strain-dependent and the size of the tolerant subpopulation correlates with tolerance level [4,29,30]. Second, azole tolerance enables slow growth, in addition to survival, of the cells, with the average growth rate being weakly dependent on drug concentrations above the MIC [4,30]. Third, azole tolerance is drug class-specific: tolerance to azoles does not correlate with tolerance to echinocandins, but cross-tolerance is often seen among the azole drugs, including fluconazole, posaconazole, and ketoconazole [30–33]. Fourth, because upon retesting, again only a subpopulation of cells grows in the drug, this cell-to-cell variability is due to phenotypic heterogeneity, rather than the acquisition of genetic changes [4]. Finally, intrinsic azole tolerance in C. albicans isolates is strain- and condition-dependent; i.e., tolerance levels change in response to temperature, medium composition, and pH [4,30,33,34]. This may be connected to the effects of metabolic flux on the ability of cells to grow under high stress conditions [21]. Clinical isolates of the “superbug” C. auris also display a similar type of tolerance to azole drugs [35].

Acquired antifungal tolerance, like resistance, can be acquired rapidly via large-scale genome changes, primarily via aneuploidy [30–32], or copy number variations [36]. Using broth medium containing wide range of supra-MIC concentrations of fluconazole, after 24 h exposure, most colonies (16.7%--66.7%) randomly tested from the cultures became tolerant to fluconazole, suggesting that tolerance is a mechanism that bypasses the drug/target interaction [30]. Even short time exposure (24 h) to a sub-MIC concentration of fluconazole could also select tolerant mutants, although the frequency is much lower (1.7%) [37]. On solid rich medium plates supplemented with supra-MIC concentrations of fluconazole, tolerant colonies emerge within three days [30]. Such acquired tolerance is usually unstable and is lost concomitant with loss of the aneuploid chromosome [30], and might be better termed ‘hetero-tolerance’ to distinguish it from the intrinsic tolerance that is stably associated with each strain.

For C. albicans and C. parapsilosis, acquired caspofungin tolerance is mediated by aneuploidy, which does not increase the MIC, yet enables cells to grow in inhibitory concentrations of the drug. As with HetR, this tolerance is not due to point mutations in FKS genes (which encode the drug target) [38–41]. Similarly, in C. auris, selection in fluconazole led to the rapid acquired tolerance via different genome changes including the acquisition of aneuploidy [42].

Recent work in multiple labs has converged on segmental or whole chromosome trisomy of ChrR as a driver of acquired tolerance to fluconazole [30,31], posaconazole [32] and miconazole [36] in C. albicans. The mechanism by which ChrR trisomy mediates acquired azole tolerance is not clear. However, one possibility is that it may indirectly affect drug uptake and/or efflux in some cells: tolerant cells usually have lower steady-state intracellular fluconazole levels [4].

Copy number variation (CNV) that causes transient gene amplification of chromosome regions, ranging in size from one to many genes, also emerges upon selection in antifungal drugs. For example, amplification of a genome segment that includes MRR1, which encodes a positive regulator of multi-drug resistance pumps, drives azole resistance; and an amplification of MRR1, together with CDR1, CDR2, and CRZ1 drives cross-tolerance to multiple azoles [36]. Intriguingly, like aneuploidy, these CNVs are selected in the presence of the drug and can be transient, with their copy number often returning to near-wild-type levels when selection is relaxed [36].

In contrast to intrinsic tolerance, which exhibits phenotypic heterogeneity, acquired tolerance frequently emerges via aneuploid chromosome acquisition following drug exposure. Importantly, specific aneuploidies (e.g., ChrR trisomy or segmental aneuploidy) emerge recurrently, resulting in transient tolerance that is largely independent of the fluconazole concentration [30,31].

Acquired flucytosine tolerance in C. parapsilosis, similar to azole tolerance, is detected as slow growth inside the inhibition zone in disk diffusion assays. The tolerance was due to the formation of chromosome 5 trisomy; however, genes known to be involved in the uptake and metabolism of 5FC are not on the aneuploid chromosomes [41].

Preventing tolerance

Although the underling mechanism(s) required for acquiring azole tolerance remain to be identified, several genes are required for maintaining azole tolerance. Azole tolerance is associated with multiple genetic components and stress pathways involving membrane integrity and cell wall function. Inhibition of Hsp90 and its client protein calcineurin, sphingolipid biosynthesis, or protein kinase C activity can abolish fluconazole tolerance. Importantly, eliminating tolerance with any of these inhibitors also causes the death of cells exposed to supra-MIC fluconazole concentrations, making fluconazole treatment fungicidal rather than fungistatic [4,29]. This suggests that tolerance is an important mechanism for viability during drug stress and thus that it could be a good target for antifungal drugs. Finding adjuvant drugs with pharmacologically useful properties that could be used to potentiate azole and/or echinocandin efficacy is a promising new direction for the identification of combination therapies that could be adopted rapidly relative to the time required to develop new antifungal drugs.

Other genes required for azole tolerance include components of the vacuolar-type proton pumping ATPase (V-ATPaseS a ubiquitous enzyme responsible for proton transport across membranes and acidification of intracellular organelles. Pharmacological inhibition of V-ATPase or deletion of the fungal-specific VMA11 gene, which encodes the c’ subunit of the V₀ subcomplex of the V-ATPase, abolish ketoconazole tolerance [33]. Similarly, deletion of RPN4 or NDU1 in C. albicans cause loss of fluconazole tolerance [43,44]. RPN4 encodes a transcription factor required for the expression of proteasome components as well as genes important for ergosterol biosynthesis. NDU1 is essential for mitochondrial respiration. Finally, both azole resistance and tolerance can be restored by small molecules such as 1,4-benzodiazepines [45]. At this point, it appears that tolerance is a process involving many interdependent processes that contribute to the ability to continue growing slowly despite the presence of inhibitory drug concentrations.

Conclusions

In fungal pathogens, antifungal drug heteroresistance and tolerance differ between species and between drug classes and we suggest that a community-wide consensus on definitions and assays for HetR and tolerance would clarify thinking and aid in progress toward understanding the mechanisms involved. We note that assays performed on agar medium allow one to distinguish between HetR and tolerant colonies and can reveal both the proportion of the subpopulation that are tolerant and the frequency of appearance of HetR. While aneuploidy is clearly one route to HetR and to elevated tolerance levels as well [13,30,31], the molecular mechanisms by which specific aneuploid chromosomes promote tolerance are not well understood and often are not due to the ‘usual suspect’ genes that confer antifungal drug resistance. Thus, we also must remain aware of the many alternative mechanisms of phenotypic heterogeneity that can allow cells to survive drug stress.

Data Availability

No data were used for the research described in the article.