Antifungal Resistance: An Emerging Reality and A Global Challenge

Fungi are important causes of disease among organisms in the plant and animal kingdoms, including humans [1]. Because the use of antifungals is widespread in both agriculture and modern medicine, the prevalence of resistant fungal infections has been on the rise. The problem of growing antifungal resistance is exacerbated by a paucity of new antifungal agents in development that have unique mechanisms of action [2]. Given the increasing number of patients with chronic immunosuppressing conditions who require antifungal therapy as part of their supportive care, there is growing concern that we may be approaching not only the postantibiotic era but also the postantifungal era [3]. The emergence of fungal infections with innate or acquired resistance to modern antifungals is the focus of this Journal of Infectious Diseases supplement. In a series of 7 articles authored by notable experts in this field [4–10], we visit different aspects of this complex area. I would like to highlight some of what I find to be intriguing questions in this dynamic field. Resistance in both Candida and Aspergillus species has become a global problem since the introduction and widespread use of triazoles in early 1990s [11]. Reports of non-Candida species resistant to azoles and of multidrug-resistant Candida species, such as Candida glabrata [12] or, lately, Candida auris [13], are increasing with alarming frequency. In addition, the emergence of Aspergillus species that are resistant to triazoles is particularly concerning because azoles are the only oral treatment option for prophylaxis and long-term treatment of invasive fungal disease [3]. Iatrogenic selection of resistance for both Candida and Aspergillus in the past was typically restricted to patients who received prolonged treatment courses in the setting of poor source control or persistent severe immunosuppression [14, 15]. More recently, azole-resistant aspergillosis has been reported on 4 continents [16] and has been hypothesized to be driven by agricultural use of azoles [3]. This creates a difficult problem because of the need for complex containment strategies that involve regulating the agricultural use of fungicidal agents. Some intriguing questions have emerged in the epidemiology of fungal resistance. Why is the prevalence of azole resistance in Aspergillus so variable between hospitals in the same region or between countries? Is this an artifact of suboptimal sampling and/or detection methods? Early experience has suggested that more-intensive and more-systematic approaches to detect azole resistance in institutions often equate with a higher prevalence of azole-resistant Aspergillus strains. Alternatively, are there unique ecological niches in the environment or patterns of spread of resistance through geoclimatic factors that are yet to be understood (eg, global warming, wind activity, patterns of migration by birds to different areas, as appears to be the case for multidrug-resistant bacteria) [17]. For example, infections by the multidrug-resistant fungus Scedosporium prolificans have been geographically restricted to countries such as Australia and Spain [18], and amphotericin B–resistant Aspergillus terreus has been described mainly in isolated centers of Austria and Texas [19]. Clearly, investment in robust epidemiological studies is urgently needed to understand the relative contributions of the environment, specific antifungal drug classes and individual drugs (including their pharmacokinetic and pharmacodynamic behavior), and, importantly, risk factors pertaining to specific hosts. Importantly, not all patients with immunosuppression are similar in terms of their risk for resistant fungal infections and the type of such infections they acquire. Some examples are telling. First, patient with chronic granulomatous disease and severe inherent defects in neutrophil function very rarely develop the resistant mold infection mucormycosis, and the amphotericin B–resistant organism Aspergillus nidulans is a typical pathogen only in these patients [20]. On the other hand, patients with other defects in neutrophil function, such as those receiving high doses of corticosteroids for treatment of chronic graft versus host disease following allogeneic stem cell transplantation, develop this infection, especially if they are hyperglycemic and have been exposed to antifungal agents that lack Mucorales activity [21]. Second, infections by Candida krusei [22] or Fusarium solani rarely spread beyond patients with leukemia and those who have undergone hematopoietic stem cell transplantation. Fungal diagnostic assays, including tests to ensure the timely diagnosis of a resistant fungal infection, are one of the most urgent needs in medical mycology. Having better diagnostic tests would increase the possibility of implementing an effective antifungal stewardship program that would allow physicians to better distinguish colonization (a setting in which resistance frequently emerges) from invasive infection. In clinical practice, there is a prevailing suspicion that a sizable portion of resistant infections could manifest as a fatal terminal pneumonia due to an unknown pathogen. The very low frequency of autopsy in modern hospitals [23] limits opportunities to detect infections caused by pathogens that acquired resistance to antifungals, new infections by an innately resistant fungus, or even mixed fungal infections. Culture-independent methods such as the detection of DNA, an antigen, or a metabolomic signature could advance the field. At this point, only polymerase chain reaction (PCR) has been shown to have promise for both early diagnosis and detection of resistance [24]. Even PCR, though, has low sensitivity for detection of resistance markers, because resistance-encoding genes are frequently found in 1 copy in the genome, compared with other genes (such as those involved in 18S ribosomal RNA sequencing) that are used for fungal identification. Matrix-assisted laser desorption/ionization–time of flight testing should in theory be a useful modality for both diagnosis of fungal infection and determination of the pathogen’s resistance patterns, but this will require the construction of libraries containing data on a sufficient number of resistant fungi [25]. However, one needs to be thoughtful on how to integrate these non–culture-based diagnostic tests into the work flow of the clinical microbiology laboratory [26] and into clinical decision making. For example, it might be difficult to differentiate colonization from invasive infection or a living from a dead fungus (especially for PCR). Nevertheless, there is little doubt that, in the future, we will see incorporation of serum- or bronchoalveolar lavage–based biomarkers in our therapeutic approaches. A key problem in this field is that information about the resistant pattern of the offending fungal pathogen is often not available to clinicians at the time when treatment decisions are made. Culture-independent methods that identify both the fungus and its resistance pattern will be needed to show how that laboratory information affects outcomes. These studies are becoming even more important as the susceptibility profile of many different fungi can no longer be inferred from identification of the pathogen to the species level. Another difficulty is that in vitro conditions for evaluating fungal resistance (eg, high inoculum, aerobic environment, high glucose medium, and planktonic pattern of growth) do not simulate fungal growth in vivo, especially for opportunistic mold infections in the lung, in which there are poor substrates (frequently in biofilms), semianaerobic conditions, and, typically, a low infecting inoculum. The frequent sequential exposure of the infecting fungus to different antifungals and the widespread use of combination therapy make interpretation of in vitro susceptibility data difficult. Not surprisingly, clinical breakpoints of resistance have been developed only for Candida species and mostly derived from episodes of mucosal candidiasis or invasive candidiasis in nonneutropenic patients. No studies have proven that using in vitro susceptibility testing has a beneficial effect on outcomes of candidiasis in very sick patients (eg, patients in the intensive care unit who have a high Acute Physiology and Chronic Health Evaluation score) or neutropenic patients [27]. The situation is similar in patients with infections caused by less frequent opportunistic yeast or mold [28] infections. In these patients, breakpoints are not based on data derived from clinical responses or outcomes in these patients but only from epidemiological cutoff values and pharmacokinetic and pharmacodynamic data from animal models [29]. Finally, the frequent lack of concordance between results of different laboratory methods [30] may be a reflection of the artificial characteristics of the laboratory methods used for detecting resistance. For example, changes in minimum inhibitory concentrations for caspofungin may be driven, in part, by methodological attributes of the broth microdilution test, such as differences in composition of growth media [31]. We may have to develop alternative methods that better simulate in vivo growth to improve the sensitivity of these tests for detection of resistance. Tremendous progress has been made in characterizing molecular and epigenetic mechanisms of resistance, especially in Candida and Aspergillus [32]. Less is known about innate or acquired mechanisms of resistance among frequently encountered opportunistic yeasts other than Candida and Cryptococcus (eg, Trichosporon) and opportunistic molds other than Aspergillus (eg, Fusarium), although some very interesting stories have emerged [33]. Most of these mutations involve target modifications or overexpression of efflux pumps [32]; however, apart from genes affecting the expression of target proteins or efflux pumps, we know less about genes encoding transcription factors involved in resistance due to gain of function mutations or epigenetic changes. Of note, several of these resistance mechanisms have been discovered through analysis of laboratory-adapted strains, and we know less about the mechanisms of resistance operating in clinical settings. Because serial isolates are not available, we often miss the opportunity to study the microevolution of resistance [34] in high-risk patients that arises from antifungal selection pressure. The exception has historically involved patients with human immunodeficiency virus infection (with or without progression to AIDS) with oral thrush, women with chronic recurring vulvovaginal candidiasis, and patients with chronic underlying lung disease with semi-invasive aspergillosis syndromes. The problem is particularly important in highly immunosuppressed patients with pulmonary mycoses such as acute or subacute invasive pulmonary aspergillosis. It is possible that many of the clinical fungal strains isolated during a breakthrough infection in patients with a past history of antifungals have multiple mechanisms of resistance or uncharacterized mechanisms of resistance [35] or tolerance that might involve organelles important for drug detoxification, such as mitochondria. Fungi have tremendous genomic plasticity that enables them to become tolerant or fully resistant to antifungals [36]. The relationship between tolerance and resistance is fascinating and incompletely understood [36, 37]. This relationship should be explored further, as the clinical scenario typically involves exposure of fungi to repetitive cycles of antifungal therapy, which creates a permissive setting for the acquisition of tolerant organisms that ultimately become resistant. In addition, we do not fully understand the relationships between antifungal resistance and pathogen fitness [38, 39] and how both susceptible and resistant fungi alter strategies to evade the host’s immune system [40]. Finally, the difference in prevalent mechanism(s) of resistance between fungi with low-level resistance and those with high-level resistance and, importantly, the clinical implications of this finding are unclear [41]. Direct genomic sequencing of resistant fungi, although appealing, remains in an early investigational stage. The complexity of the fungal genome [42], with its frequent insertions, deletions, and rearrangements, and of the contribution of gain of function mutations involving other genes and promoters to resistance create a difficult context to interpret genomic changes and differentiate them from background noise. In the absence of robust clinical experience, animal models have become an essential tool for identifying possible treatment strategies for resistant fungal pathogens [8, 43]. However, animal models rarely simulate clinical scenarios that lead to the emergence of antifungal resistance. Specifically, opportunistic fungal infections typically occur in patients at the extreme of ages, in patients with severe complex and chronic immunosuppression, and in those with multiple comorbidities. In contrast, animal models use relatively young animals without comorbidities that are exposed acutely to high-level immunosuppression and high fungal inocula to overcome their natural resistance to fungal infections. These mice are frequently inbred, and this does not reflect the genetic heterogeneity one encounters in real life in humans. In addition, although eloquent studies have been done [8] correlating pharmacokinetic and pharmacodynamic properties of antifungals in infection models caused by resistant fungi, one needs to be reminded that pharmacokinetic and pharmacodynamic behavior of an antifungal in an animal and a human might vary, partially depending on the underlying immune status of the host [8, 43]. Finally, despite their promise in studying fungal pathogenesis, it is unproven whether mini-host models such as Drosophila, Galleria, and Caenorhabditis elegans could help accelerate the study of fungal resistance. More work as needed to develop these miniature models [44] as platforms for high-throughput screening of novel antifungal compounds to treat infection due to a resistant fungus. Acquisition of a fungal infection caused by a resistant fungus is a relatively late event in the treatment scenarios of immunosuppressed patients. Specifically, major confounders, such as the activity of underlying disease, comorbidities, drug-drug interactions, immunosuppression, source control issues, and lack of choice of antifungals based on real-time information about resistance, are important biases that are difficult to eliminate in case-control studies. Current guidelines for treatment of candidiasis [45] and aspergillosis [46, 47] do not provide robust evidence-based recommendations on how to treat resistant infections. There is a major need for good data for how to treat multidrug-resistant C. glabrata infection, an emerging problem, and how to treat Aspergillus infections involving a high level of azole resistance [11, 48]. The introduction of drugs with unique mechanisms of action to our modern armamentarium will be a welcome occurrence. The market niche and commercial success of these drugs to treat fungal resistance are uncertain [2]. The discovery of new anti-infective agents in general is heavily biased toward drugs that kill the microorganism or inhibit its growth. An interesting direction in antifungal drug development would be the introduction of drugs that do not inhibit growth but impair key virulence properties, such is invasion or adherence [49]. Developing drugs that only interfere with virulence might be theoretically beneficial in terms of preservation of a healthy microbiome/mycobiome, another interesting area of research direction in this area [50]. Considerable progress has been made in the emerging field of antifungal resistance, an underestimated area of global public health concern, but significant challenges remain. We need better knowledge of the epidemiology and drivers of resistance; improved diagnostic tests, with development, ideally, of point-of-care diagnostic tests; and refinement of existing in vitro methods and antibiograms that are linked to clinically meaningful end points that are specific to hosts (eg, patients with hematologic conditions or stem cell transplants who have acute infections and patients with less severe immunosuppression who have chronic infections) and clinical units (eg, the intensive care unit and the hematology service). Improved education of prescribers and policy makers, as well as greater efforts toward antifungal stewardship [51], are also important. One can be certain that the importance of fungal resistance in modern medicine will continue to increase, creating new questions and challenges in the decades to come. Acknowledgment. I thank Russell E. Lewis for useful comments. Supplement sponsorship. This work is part of a supplement sponsored by grants from Astellas Pharma Global Development, Inc. and Merck & Co., Inc. Financial support. This work was supported by the Texas 4000 Distinguished Endowed Professorship for Cancer Research. Potential conflicts of interest. D. P. K. has received research support and honoraria from Pfizer, Astellas Pharma US, Gilead Sciences, and Merck and Company. The authors has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.