Cryptococcal therapies and drug targets: the old, the new and the promising

Summary

Half a century after the introduction of Amphotericin B the management of cryptococcosis remains unsatisfactory. The disease, caused primarily by the two fungal species Cryptococcus neoformans and Cryptococcus gattii, remains responsible for considerable morbidity and mortality despite standard medical care. Current therapeutic options are limited to Amphotericin B, azoles and 5-flucytosine. However, this organism has numerous well-characterized virulence mechanisms that are amenable to pharmacological interference and are thus potential therapeutic targets. Here, we discuss existing approved antifungal drugs, resistance mechanisms to these drugs and non-standard antifungal drugs that have potential in treatment of cryptococcosis, including immunomodulatory strategies that synergize with antifungal drugs, such as cytokine administration or monoclonal antibodies. Finally, we summarize attempts to target well-described virulence factors of Cryptococcus, the capsule or fungal melanin. This review emphasizes the pressing need for new therapeutic alternatives for cryptococcosis.

Cryptococcosis is caused by two species of the genus Cryptococcus, Cryptococcus neoformans and Cryptococcus gattii (only in rare circumstances are other cryptococcal species found to cause disease). The correlation of cryptococcosis with impaired host immunity was noted as early as 1955, when an increased frequency of disease was observed in patients with haematopoietic malignancies or undergoing corticosteroid treatment (Baker and Haugen, 1955). Beginning in the 1980s there was an exponential increase in the incidence of cryptococcosis when the AIDS epidemic resulted in millions of individuals affected by severe immunosuppression. This fungal disease is characterized by pneumonia and life threatening meningoencephalitis (Park et al., 2009). Mortality from cryptococcal infection remains as high as 30% and for many individuals with severe immunosuppression the infection is essentially incurable, requiring lifelong antifungal therapy.

Different patients require different treatments

The majority of individuals who develop cryptococcosis can be grouped in three general categories: those with advanced HIV infection, those with organ transplants entailing immunosuppressive therapy, and lastly non-HIV, non-transplant patients without an obvious immune disorder. Although clinical management of all groups is similar, the therapeutic approach is case-tailored in the sense that antifungal drug therapy is generally the same for all patients while management is individualized with regards to the nature of the immune impairment. As an example, in HIV+ individuals the best course of action is to control viral load and CD4+ T cell counts while in transplant patients modifications of immunosuppressive cocktails to reduce immunosuppression could improve prognosis. These strategies entail the caveat that improvements in immune function can lead to Immune Reconstitution Syndrome (IRIS), a paradoxical worsening of symptoms attributed to tissue damage because of an exuberant immune response directed at the fungi still in the tissue. IRIS requires delicate therapeutic management entailing control of excessive inflammation with corticosteroids. In patients with life-threatening IRIS neutralization of excessive TNF-α with adalimumab can be helpful (Sitapati et al., 2010; Scemla et al., 2015), despite the fact that TNF-α blockade increases risk of cryptococcal disease (Horcajada et al., 2007). These examples illustrate the paradigm that a sweet spot of immune regulation is necessary to control cryptococcosis (Fig 1).

Fig. 1.

Schematic of antifungal drug targets in Cryptococcus spp.

Immunocompetent individuals are not generally considered at risk for cryptococcosis but cryptococcal disease may occur if an individual is exposed to a particularly virulent strain or a high inoculum. Another possibility is that the patient possesses a cryptic immune insufficiency condition. Therefore an apparently immunocompetent patient who develops cryptococcal disease should be thoroughly investigated for predisposing factors, such as malignancies and/or conditions that result in immune deficits, such as autoimmune neutralization of GM-CSF (Bratton et al., 2012; Saijo et al., 2014), genetic polymorphisms of FcReceptor (Baker and Haugen, 1955; Meletiadis et al., 2007; Rohatgi et al., 2013) and/or idiopathic lymphopenia (Netea et al., 2004; Park et al., 2009).

The efficacy of available antifungal drugs against Cryptococcus neoformans

More than a half-century after its introduction into clinical use Amphotericin B (AmpB) remains the gold standard for the therapy of cryptococcosis. AmpB or fluconazole with or without 5-flucytosine (5-FC) are the mainstays of therapy for cryptococcosis (Sitapati et al., 2010; Perfect et al., 2010; Scemla et al., 2015). AmpB binds to ergosterol causing pores and leakiness in the fungal cell membrane. Azoles inhibit lanosterol 14 α-demethylase, a critical step in ergosterol biosynthesis. Severe toxicity of AmpB was a significant problem pushing the development of novel liposomal formulations which succeed in improving AmpB safety. The initial therapy often uses AmpB in combination with 5-FC, a pyrimidine nucleoside analog of nucleic acids. Other antifungal drugs, such as nystatin or allylamines, are either not sufficiently absorbed or are too toxic, and are therefore of limited use for cryptococcal meningitis or pneumonia.

Drug-resistance mechanisms of Cryptococcus

C. neoformans strains have been described that manifest resistance to every family of drugs (reviewed in (Perfect and Cox, 1999; Horcajada et al., 2007)). For AmpB naturally resistant strains are rare although some can acquire resistance during the course of clinical treatment, while for 5-FC appearance of resistance is so common that it precludes its use as monotherapy. For the azoles most if not all isolates include a subpopulation intrinsically resistant to azoles (Bratton et al., 2012; Sionov et al., 2013; Saijo et al., 2014). While heteroresistance may contribute to the fungistatic nature of azoles, because a percentage of the population will survive azole exposure, this class of drugs remains useful therapeutically achieving a high treatment success rate.

Echinocandins (e.g. caspofungin) target cell wall β-glucan synthases and are very effective against certain fungal infections. However they are relatively ineffective against Cryptococcus spp (Feldmesser et al., 2000; Maligie and Selitrennikoff, 2005) and are therefore not used clinically in cryptococcosis. The intrinsic resistance of Cryptococcus spp is not because of a different biosynthetic pathway because both α-1,3-glucan synthase (Reese et al., 2007) and α-1,6-glucan synthase (Baker and Haugen, 1955; Gilbert et al., 2010) mutants show a deficient cell wall architecture accompanied by decreased virulence in mouse models. Furthermore, echinocandins bind to the cryptococcal cell wall in vitro (Franzot and Casadevall, 1997; Park et al., 2009). One possible explanation for the lack of efficacy in vivo is because of fungal melanization in the host brain because cell wall-associated melanin may prevent the drug reaching its enzymatic target on the yeast. This hypothesis is supported by studies showing that fungal melanin adsorbs echinocandins and melanization reduces cell wall permeability (van Duin et al., 2002; Nosanchuk and Casadevall, 2006; Sitapati et al., 2010; Scemla et al., 2015). In addition, melanization in other pathogens has been described to greatly increases rigidity of cell wall (Money et al., 1998; Horcajada et al., 2007), which could contribute to decreased susceptibility to echinocandins in vivo.

Non-standard antifungal agents that can be used to target Cryptococcus

Antiparasitical and antiviral agents

Several antimicrobial agents have activity against Cryptococcus. Protease inhibitors used to treat HIV have activity in vitro against Cryptococcus (Blasi et al., 2004; Bratton et al., 2012; Saijo et al., 2014) and increase killing of the yeast by innate immune cells (Blasi et al., 2004; Monari et al., 2005; Meletiadis et al., 2007; Rohatgi et al., 2013). The antiprotozoal drugs chloroquine and quinacrine prevent lysosomal acidification and interfere with a myriad of cellular processes. Daily chloroquine administration to mice infected with C. neoformans reduced their mortality (Levitz et al., 1997; Netea et al., 2004). Chloroquine antifungal effects range from inhibition of Cryptococcus growth (Harrison et al., 2000; Perfect et al., 2010) to facilitating an increase in yeast killing by macrophages (Perfect and Cox, 1999; Khan et al., 2005). Additionally chloroquine effects on the host contribute to its beneficial effects (Narayanan et al., 2011; Sionov et al., 2013). A case report in a patient with IRIS found chloroquine was perfectly tailored to the patient’ symptoms: reduction of hypercalcemia, dampening of excessive inflammation and the aforementioned antifungal activity resulted in a successful control of IRIS in this patient.

Drugs targeting neurotransmitters

Some anti-psychotic drugs have activity against Cryptococcus growth. The antidepressant sertraline reduced fungal burden in mice to levels similar to fluconazole (but not as efficiently as AmpB) (Feldmesser et al., 2000; Maligie and Selitrennikoff, 2005; Zhai et al., 2012; Treviño-Rangel et al., 2015). However results from clinical trials phase I/II were encouraging enough to proceed to phase III as adjunctive to combination therapy of AmpB and fluconazole (Reese et al., 2007; Rhein et al., 2016). The mechanism of how sertraline inhibits Cryptococcus growth is unknown, but there is evidence from Candida that it may function as a translation inhibitor.

Calcineurin pathway inhibitors

Calcineurin signalling is a highly conserved pathway, shown to affect virtually all phenotypes associated with virulence in C. neoformans (Steinbach et al., 2007) and C. gattii (Chen et al., 2013), including the ability to grow at high temperature. Compounds inhibiting calcineurin are immunosuppressive and are commonly used in transplant recipients for immune tolerance of the graft. The importance of the calcineurin pathway in fungal virulence led to a prospective clinical study to determine the balance between antifungal and immunosuppressive effects of calcineurin inhibitors. This study showed an earlier onset of cryptococcosis in patients receiving either tacrolimus or cyclosporine A but decreased mortality in this cohort (Singh et al., 2007). Consequently current calcineurin inhibitors are not considered appropriate for treatment of cryptococcosis, although they might be an appropriate alternative for IRIS-associated cryptococcosis. Nevertheless this pathway remains an attractive target for the development of fungal-specific calcineurin inhibitors that presumably would exhibit greater direct antifungal effects than immunosuppressive effects.

HMG-CoA reductase inhibitors

Fungal ergosterol and human cholesterol are structurally distinct but their 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMG-CoA) reductases share 76% homology. HMG-CoA inhibitors, a class known as statins, are widely used in humans to reduce cholesterol levels. The similarity between human and fungal HMG-CoA has raised interest in the potential of statins as antifungal drugs (Bergman and Björkhem-Bergman, 2013). However a meta analysis of five retrospective studies found no positive effects for the use of statins during fungal infection (Bergman and Björkhem-Bergman, 2013).

Immunomodulators

Antibody

The mechanism of antibody (Ab) protection is not fully elucidated, but it is believed that certain Ab contribute to protection through regulation of immune responses (reviewed in (Casadevall and Pirofski, 2003; Brady, 2005)). There is also evidence that some monoclonal Ab (mAb) can modulate C. neoformans metabolism and are synergistic with antifungal drugs (McClelland et al., 2010). Furthermore Ab against highly conserved structures such as β-glucans, glucosylceramides and Hsp90 could potentially have broad activity against many species of the fungal kingdom. Ab against β-glucans reduced fungal burden in mice infected with Cryptococcus (Rachini et al., 2007), Ab against glucosylceramides protected mice from a lethal dose of C. neoformans (Nimrichter and Rodrigues, 2011). A humanized Ab fragment against Hsp90, under the name Mycograb®, was fungicidal in vitro (Nooney et al., 2005), but manufacturing concerns prevented approval for human use.

The coupling of Ab to radioactive molecules, dubbed radioimmunotherapy, enhances the efficacy of Ab by converting them into a delivery system for microbicidal radiation to targeted microbe (Tsukrov and Dadachova, 2014). Radioimmunotherapy with a capsule-binding Ab is effective in treating cryptococcosis in mice (Dadachova et al., 2003; Dadachova and Casadevall, 2009).

IFN-γ

Interferon-γ is a key cytokine in cryptococcosis such that in human patients a strong IFN-γ response indicates a good prognosis (Jarvis et al., 2012; Jarvis et al., 2014). Two clinical trials administering recombinant interferon have been performed. The first found a trend towards better clinical outcome and earlier CNS sterilization but no differences in survival (Pappas et al., 2004). The second showed that administration of IFN-γ improved fungal clearance from the CNS but had no statistically significant decrease in patient mortality (Jarvis et al., 2012). The lack of statistical significance in mortality rates but significantly better clinical outcomes may be explained by the small sample size and/or heterogeneous response to therapy. Overall these results are encouraging and suggest that certain individuals might benefit from IFN-γ. Further evidence for its efficacy comes from the case-report of two cryptococcosis patients with idiopathic lymphopenia and low baseline IFN-γ who improved upon administration of IFN-γ (Netea et al., 2004).

Other immunomodulators

TNF-α and Il-10 levels in patients correlate with disseminated infection (Lortholary et al., 1999; Lortholary et al., 2001). In mice Il-10 was shown to be beneficial for the host survival (Blackstock et al., 1999), possibly by ameliorating immune-derived tissue damage, but so far there is no rationale to support administration of IL-10 to patients. An alternative immunomodulatory therapy is administration of mimics of pathogen associated molecular patterns (PAMPs), particularly TLR agonists. Administration of unmethylated CpG nucleotides (Kinjo et al., 2007) or a mimic of microbial RNA, poly-IC, (Sionov et al., 2015) delayed mouse death upon Cryptococcus infection. These microbial-molecule analogues elicit an IFN-γ response, and a formulation of poly-IC (Hiltinol ®) is used as adjuvant for boosting vaccine activity, but is not potent enough as monotherapy.

Cryptococcal drug targets

Polysaccharide capsule

A distinguishing feature of Cryptococcus is the presence of a polysaccharide capsule, whose importance for virulence is illustrated by the fact that acapsular mutants are avirulent (Sionov, Chang & Kwon-Chung, 2013). The capsule protects the fungal cell from a myriad of host microbicidal molecules and shed polysaccharide can alter the immune response to the detriment of the host (Vecchiarelli et al., 2013). Consequently, the capsule and its biosynthetic pathways (Almeida et al., 2015) are attractive drug targets.

Because of its polysaccharide nature, the capsule is poorly immunogenic. The lack of Ab response can be overcome by use of polysaccharide–conjugate vaccines, which elicit strong antibody responses and allowed the isolation of several anti-capsule mAb from mice (Casadevall et al., 1992; Casadevall et al., 1998; Shapiro et al., 2002). These mAb have been extensively studied for their immunization and therapeutic properties (reviewed in (Casadevall and Pirofski, 2005; Casadevall and Pirofski, 2012)). Their use in clinical trials phase I (Larsen et al., 2005) was promising. However, this approach was not further developed because of insufficient resources at a time when the prevalence of cryptococcosis in developed countries was falling on account of effective treatment of AIDS.

Melanin and laccase

Melanin is a pigment synthesized by the enzyme laccase that is important for virulence for many pathogenic microbes and Cryptococcus spp. in particular (Kwon-Chung et al., 1982; Liu and Nizet, 2009). Melanin can both absorb oxidative stress radicals and toxic compounds, even antifungal molecules (van Duin et al., 2002; Nosanchuk and Casadevall, 2006). As noted before, trifluoperazine, an anti-psychotic drug affected growth of melanized cells in vitro but did not affect growth of non-melanized cells (Wang and Casadevall, 1996) and reduced mortality in mice when administered daily post-infection (Eilam et al., 1987). Later, the herbicide glyphosate was shown to inhibit melanin production and prolong mouse survival (Nosanchuk et al., 2001). These strategies have not been investigated in human trials.

Analogous to the strategy used for capsular polysaccharide, mAbs to fungal melanin have been generated that are active against C. neoformans (Rosas et al., 2001). Passive immunization with such mAbs prolonged survival of mice. The development of melanin mAbs is attractive as these mAbs may react with other fungal species that utilize melanin and become broad-spectrum antifungals.

Proteins, proteases, breaching of the blood brain barrier and secreted vesicles

The first fungal molecule proven to contribute to brain invasion was urease because strains lacking urease and accessory proteins have decreased fungal burden in the brain and are defective in crossing cellular layers in transwell assays (Cox et al., 2000; Olszewski et al., 2004). Later, other cryptococcal molecules were associated with CNS dissemination: phospholipase B1, which interacts with host cell Rac1 (Maruvada et al., 2012); the hyaluronic acid-like product of Csp1 gene, which interacts with CD44 in endothelium (Kim et al., 2012) and Mpr1 metalloprotease (Vu et al., 2014). All of these proteins are secreted and disrupt host cells to trigger transmigration. Secretion is essential for virulence, yet it was a conundrum in the cell biology of C. neoformans how fungal molecules sequentially traversed cell membrane, melanized cell wall and fungal capsule. This problem was solved by a recently described vesicular transport system. Vesicular transport has been implicated in capsule assembly (Rodrigues et al., 2007), secretion of melanin, urease and many others (Rodrigues et al., 2008). A recent screen has identified two compounds causing accumulation of intracellular vesicles and inhibited surface glucosylceramides. This promising compounds protected mice from lethal challenges to C. neoformans and C. albicans (Mor et al., 2015). All things considered inhibition of fungal vesicular transport has the potential to become a viable drug target in Cryptococcus spp.

Future directions

In this review we have highlighted several attempts to target Cryptococcus spp. or host immune mechanisms to improve outcome of cryptococcal disease. While some attempts have reached clinical trials the last new class of successful compounds were the azoles, first approved over 25 years ago. Several novel compounds remain under investigation and there is hope that some will progress to become useful antifungal agents. Histone deacetylase inhibitors, a class of drugs with immunosuppressive and immunomodulatory properties, are active in vitro against Cryptococcus and synergy with triazoles (Pfaller et al., 2009; Brandão et al., 2015). High-throughput screens generated at least two new classes of compounds with anticryptococcal activity in vitro: hydroxyaldimines or Schiff bases, (Magalhães et al., 2013) and aminiothiazioles (Khalil et al., 2015). Natural extracts from seaweeds or herbal essential oils have anti-cryptococcal activity (Amraoui et al., 2014; Cavaleiro et al., 2015) and curcumin, a molecule present in turmeric, was beneficial in mouse models of cryptococcosis (da Silva et al., 2016). A recent forward genetic screen has identified previously unknown genes essential for cryptococcal viability and potentially targetable (Ianiri and Idnurm, 2015). Meanwhile the recently discovered phenomena of non-lytic extrusion, the Trojan-horse brain dissemination and titan cells provide potential new antifungal mechanisms to target Cryptococcus spp.