Fungal infections are common in patients with acquired immunodeficiency syndrome and pose a major health management problem. There is a need for identification of new antifungals to complement the limited current repertoire and to combat newly arising resistant fungal strains. We have identified a novel antifungal activity for the antiarrhythmic drug amiodarone. Extensive characterization of this activity shows that amiodarone exhibits a growth inhibition for several diverse fungi, including species of Cryptococcus,Saccharomyces, Aspergillus,Candida, and Fusarium. The antifungal activity was shown to be fungicidal; Cryptococcus neoformans treated with amiodarone lost viability within hours of drug exposure. Growth inhibition could be suppressed by addition of very high concentrations (10 mM) of calcium to the medium, suggesting that disruption of calcium homeostasis was involved in the antifungal activity. Direct measurement of radiolabeled calcium efflux showed that addition of amiodarone resulted in an immediate efflux of cellular calcium. In conclusion, amiodarone displays broad-based fungicidal activity and may be acting in part by perturbing the calcium balance.
AIDS is defined by the occurrence of at least one of more than two dozen opportunistic infections. Opportunistic fungal infections, such as candidiasis, cryptococcosis, and histoplasmosis, occur frequently in patients with AIDS. Among the opportunistic infections, fungal infections caused by Pneumocystis, Candida,Cryptococcus, or Histoplasma were the first to occur in more than 50% of persons with AIDS; at time of death, nearly 85% of decedents had a fungal infection (Sugar, 1991; Kwon-Chung and Bennett, 1992; Stansell, 1993; Mitchell and Perfect, 1995; Jones et al., 1999; Bastert et al., 2001). Certain systemic fungal infections have a high mortality rate among cancer patients as well (Viscoli et al., 1997).
There is intense interest in identifying new drugs with different modes of action against fungal infections (DiDomenico, 1999). The current repertoire of antifungals has limitations such as insufficient efficacy, the need for intravenous administration, serious side effects, or the appearance of resistant fungal strains (Graybill, 1996). Importantly, most of the current treatments are fungistatic and subsequent clearing of these fungi is inadequate in patients with defective immune systems. Thus, it is imperative to identify cellular targets that, when impaired, lead to fungal cell death.
We have identified a new antifungal activity displayed by amiodarone, a drug now in clinical use as an antiarrhythmic (Mason, 1987; Gill et al., 1992; Roden, 1996; Singh, 1996). Amiodarone had been reported to bind directly to mammalian heterotrimeric G proteins, thereby activating them (Hagelüken et al., 1995). While testing for amiodarone's ability to affect G protein-mediated processes in fungi under study in our laboratory, it was found that amiodarone dramatically affected the growth of the fungal cells. The growth-inhibitory effect was examined further, and amiodarone was found to have potent fungicidal activity against a broad range of fungi, including those of clinical importance.
Materials and Methods
Strains, Media, and Reagents.
The Cryptococcus neoformans strains used were JEC21 (MATα) and 271 (MATα; provided by Dr. T. Kozel, University of Nevada, Reno, NV). The Aspergillus fumigatus was American Type Culture Collection strain B-5233 (provided by Dr. R. Washburn, University of Nevada, Reno, NV). The Saccharomyces cerevisiae strains were BC159 (MATa ade2 ura3–52 leu2–3,112 his3) and FY70 (MATa ade1 trp1 leu2 his3 ura3; provided by Dr. S. Garret, University of Medicine and Dentistry of New Jersey, Newark, NJ).
Cells were grown in SD [0.17% w/v Difco yeast nitrogen base without amino acids and ammonium sulfate, 0.4% w/v ammonium sulfate, 2% (w/v) glucose]. Agar media contained 2% (w/v) Bacto-agar. Adenine (12 mg/l final concentration), uridine (40 mg/l), leucine (30 mg/l), histidine (20 mg/l), and tryptophan (20 mg/l) were added to supplement auxotrophies as needed.
Amiodarone (ICN, Costa Mesa, CA) was added from a 10 to 20 mM stock in dimethyl sulfoxide (DMSO). DMSO was added to the no-drug cultures as a control. No effect on cell proliferation was observed by the addition of DMSO at the concentrations used (data not shown).
Growth Rates and Viability.
For quantitation of cell proliferation, cells were grown in 5 ml of liquid medium in Klett test tubes with vigorous shaking in a water bath. Measurement of cell density was done in a Klett-Summerson colorimeter. Cells were grown overnight in the medium to be tested. Dilutions of the overnight cultures were made into a series of test tubes containing fresh medium and various concentrations of drug or DMSO. The cell densities were monitored at time 0 (typically about 1–2 × 106 cells/ml) and over time (typically 8–10 points for each growth curve) up to about 10 h. The increase in cell density was plotted versus time, and the resulting curves were used to determine the generation time for each culture (Stanier et al., 1976). Each experiment was repeated three to eight times, and the mean values for generation times are given in the text and tables along with their standard errors. This method for cell growth was used because it is more accurate than standard macro- or microbroth methods (e.g., the reliability values are typically greater than 0.99) and is highly reproducible.
The viability of cells treated with amiodarone was tested by a colony-forming assay. Cultures were prepared as just described for the growth rate experiments, and cells were treated with 10 μM amiodarone or DMSO. Aliquots were taken at time 0 (when cells were added to the drug-containing medium) and at various subsequent times. The aliquots were spread onto SD agar medium without drug and incubated at 30°C until colonies formed. Typically, microliter volumes were taken at early times and sample volumes were increased as the viability decreased. When few if any viable cells were expected, the entire culture was collected by centrifugation, resuspended in a small volume of H2O, and plated to allow the identification of any remaining live cells. The viability for DMSO-treated cells did not decline significantly during the course of the experiment (data not presented). Colonies were counted to determine the number of viable cells per milliliter in the aliquot. Viability is defined as the number of colonies formed at at given time point divided by the number of colonies at zero time × 100%.
S. cerevisiae (FY70) cells were grown to mid exponential phase overnight in SD medium to a density of about 1 to 2 × 107 cells/ml. Two milliliters of cell culture was mixed with 1 μl of45CaCl2 (19 μCi; ICN) and then incubated at 30°C for approximately 1 h to allow accumulation of the radionuclide by the cells. One milliliter of the radiolabeled cells was then filtered through an HAWP 0.45 μm membrane filter (Millipore Corporation, Bedford, MA) to collect cells. The cells were washed with 15 ml of H2O and resuspended in 2 ml of 1 mM sodium acetate, pH 5.0, 3 μM GdCl3. A 200-μl aliquot was collected for the zero time point by filtration with membranes presoaked in 5 mM nonradioactive CaCl2 and washed with 15 ml of 5 mM nonradioactive CaCl2 to remove any external45Ca. Amiodarone (in DMSO) was added to the remaining cells at concentrations from 5 to 20 μM, and 200-μl samples were collected over time by filtration and washed as was the zero time aliquot. Any radiolabeled calcium released during the amiodarone treatment would be removed by the washing. The filtered cells were air dried and counted in 1 ml of scintillation fluid (Universol; ICN) in a Beckman Coulter (Fullerton, CA) LS 1701 liquid scintillation counter to determine the remaining cell-associated counts. An additional 1 ml of the same labeled cells was handled identically except cells were treated with DMSO as a control.
Inhibition of Cryptococcal Growth by Amiodarone.
The effect of amiodarone on the growth of C. neoformans cells was measured in liquid cultures. JEC21 cells were grown overnight in SD minimal medium (30°C) to mid/late exponential phase. At zero time, aliquots were transferred into fresh SD medium containing either amiodarone or DMSO as a control. Cell densities (which began at about 1–2 × 106 cells/ml) were measured over an approximately 10-h period, and the results were plotted to produce growth curves. Figure 1 shows dose-response curves for cells treated with varying concentrations (1–10 μM) of amiodarone. Each curve is the mean of four independent experiments, and rates of increase were used to determine the generation times (Table1). During this period, the control cells (which received only DMSO) increased in density about 10-fold to around 1.5 × 107 cells/ml and grew at a rate of 136 min/generation (Table 1). Addition of 1 μM amiodarone did not alter the generation time of the JEC21 cells, whereas addition of 3 μM amiodarone doubled the length of the generation time (to 273 min/generation). Thus, the MIC50 for amiodarone against JEC21 at 30°C is 2 μg/ml. Addition of amiodarone at 5 μM or above prevented virtually any increase in cell density during this period. Growth of Cryptococcus at high temperature is an important virulence factor for this fungus; thus, we examined the effect of amiodarone on the growth of cells at 37°C. Whereas no effect on growth was seen in cultures with 1 μM amiodarone, cultures containing 3 μM amiodarone displayed growth rates of 432 min/generation, which is about 58% slower than that at 30°C, showing an exacerbating effect of higher temperature. As with 30°C, no increase in cell density occurred in cultures treated with 5 μM amiodarone or greater. Thus, amiodarone treatment dramatically inhibits cryptococcal proliferation.
JEC21 is a laboratory strain developed for genetic and molecular biological research. We tested whether the amiodarone effect occurred with a strain isolated from a patient. The effect of amiodarone treatment on the growth of the clinical isolate strain 271 was determined (Table 1). Amiodarone also very effectively inhibited the growth of the clinical isolate; strain 271 growth was only slightly affected by the presence of 3 μM amiodarone but was largely inhibited by 5 μM and, like JEC21, did not grow in the presence of 7 μM amiodarone. Moreover, growth of another clinical strain (J9D) was more severely inhibited by amiodarone than was JEC21 (data not presented); thus, the inhibitory effect of amiodarone is not specific to any particular Cryptococcal strain.
Amiodarone is Effective Against a Broad Range of Fungi.
The antifungal activity of amiodarone is effective against a range of different fungi. Amiodarone was tested on S. cerevisiae andA. fumigatus (Table 1). Amiodarone inhibited the growth ofS. cerevisiae strain BC159 but required a slightly higher concentration, in the 15 μM range, to completely stop proliferation. Amiodarone was also effective in preventing the growth ofAspergillus nidulans, although nearly 50 μM amiodarone was required to completely inhibit growth. The MIC50against this strain of Aspergillus is about 4 μg/ml. Similar growth-inhibitory activity by amiodarone was seen againstFusarium oxysporum and Candida albicans (data not shown). Thus, amiodarone blocked the proliferation of Heterobasidiomycetes (Cryptococcus), Ascomycetes (Saccharomyces), and Hyphomycetes (Aspergillus).
The lack of measurable growth in cryptococcal cultures treated with 10 μM amiodarone could have been due to a fungicidal or fungistatic effect. To differentiate between these possibilities, JEC21 cells were grown in SD minimal medium and treated with either 10 μM amiodarone or the DMSO control. Aliquots of the treated cultures were taken at the time that drug was added (zero hour) and at various times up to 50 h later (see Materials and Methods) and spread on SD minimal plates to allow viable cells to form colonies. The viability of cells treated with amiodarone decreased rapidly (Table 2), with a 10-fold decrease in viable cells after only 1 h of exposure and a 200-fold decrease after 10 h. After about 1 day of treatment, only 0.001% of the cells were viable, whereas after about 2 days, no viable cells remained.
Effect of Cations on Sensitivity to Amiodarone.
It has been suggested that amiodarone affects calcium, sodium, and potassium ion channels in mammalian cells. We tested whether addition of cations to the growth medium would affect the growth inhibition caused by amiodarone. Cryptococcal cells (JEC21) growing in SD minimal medium were treated with 4 μM amiodarone to cause a marked decrease (but not total inhibition) in their growth rate. In parallel cultures, cells were simultaneously given either 10 mM CaCl2, 10 mM MgCl2, 10 mM KCl, or 20 mM NaCl . Control cultures received the salt additions without amiodarone. The growth rates of each culture were followed over several hours and generation times were determined (Table 3). A 4-μM amiodarone treatment caused a lengthening of the generation time to 422.3 min compared with 162.4 min for the DMSO control. The addition of CaCl2 was very effective in antagonizing the amiodarone-induced inhibition, allowing a generation time (190.5 min) close to that of the control, whereas treatment with MgCl2 weakly suppressed growth inhibition (generation time of 216.7 min.). Addition of NaCl had a minor effect on the generation time (281.5 min.) compared with the amiodarone-only treatment, whereas addition of KCl did not alter the generation time. The addition of the salts by themselves (except KCl) resulted in slight lengthening of the generation times (Table 3). Thus, addition of divalent cations, in particular, calcium, to the growth medium antagonized the growth-inhibitory effects caused by amiodarone, suggesting that amiodarone affects cation metabolism in particular.
Efflux of Radiolabeled Calcium Promoted by Amiodarone.
The ability of high concentrations of calcium to protect cells from growth inhibition by amiodarone suggested effects on calcium homeostatis. Efflux of radiolabeled calcium was measured in the S. cerevisiae strain FY70. Cells were incubated in growth medium containing 9.5 μCi/ml 45Ca (as45CaCl2) for 1 h to allow accumulation of labeled calcium. Cells were subsequently sampled, washed free of external labeled calcium, and suspended in sodium acetate buffer. A zero time sample was taken, and then cells were treated with amiodarone or DMSO and aliquots were taken over time. All samples were filtered and washed with 5 mM nonradioactive CaCl2 to remove any labeled calcium that may have effluxed from the cells. The remaining cell-associated radioactivity was determined by measuring cell filtrates in a scintillation counter. Approximately 65% of the radiolabeled calcium remained associated with control cells treated with DMSO over the 5-min course of the experiment (Fig. 2A). In contrast, cells exposed to 10 μM amiodarone retained only about 35% of the labeled calcium. The efflux of calcium upon amiodarone addition was very rapid, with the majority of the loss occurring in less than 30 s. The efflux of the labeled calcium upon amiodarone treatment was dose-dependent. Cells treated with 2.5 μM amiodarone lost virtually the same percentage of labeled calcium as did the control cells (25% at 20 s), whereas cells treated with 5 and 10 μM amiodarone lost increasing amounts of calcium (50% and 60% loss at 20 s, respectively). Increasing the concentration of amiodarone to 20 μM did not further increase calcium loss.
We have shown here that amiodarone has antifungal activity against a range of different fungi, including those of clinical importance, such as C. neoformans. The growth-inhibitory effect was shown to be fungicidal against Cryptococcus, resulting in about 90% killing within approximately 1 h of treatment. Treatment with amiodarone caused immediate loss of calcium from cells, and this may be the cause for subsequent cell death. Consistent with a role that calcium loss may play, incubating cells in the presence of very high concentrations (10 mM) of calcium effectively suppressed the growth-inhibitory effects of amiodarone, whereas 10 mM MgCl2 had only a slightly protective effect. The presence of very high NaCl (20 mM) or KCl (10 mM) had little or no protective effect. Such suppressive effects of calcium and magnesium on growth inhibition by amiodarone are similar to the suppressive effects of these ions on killing mediated by the KP4 antifungal protein fromUstilago maydis (Gu et al., 1995). KP4 was shown to specifically inhibit voltage-gated Ca2+ channels in mammalian cells. These results point to a mechanism for amiodarone antifungal activity by the theory that amiodarone constitutively opens a calcium channel, thereby allowing efflux of a significant amount of cytoplasmic calcium. This calcium loss may be involved in growth inhibition and subsequent cell death, although alternative mechanisms for the effects of amiodarone are possible.
Calcium is known to be involved in controlling numerous cell processes and is essential for cell growth (for review see Brown et al., 1995;Clapham, 1995; Berridge et al., 1998). The role of calcium in microorganisms is still not well understood (O'Day, 1990). Mutations in numerous calcium-binding proteins, such as calmodulin, can affect cell proliferation (Ohya and Anraku, 1992). Some of these mutations can be suppressed by the addition of very high concentrations of calcium to the growth medium. There are several yeast proteins that are predicted to be calcium transporters, including some in the plasma membrane related to mammalian calcium channels (Nelissen et al., 1997). The availability of yeast strains with deletions of most of the genes encoding these proteins will allow a determination of a role, if any, in amiodarone sensitivity. Ion channels are now being recognized as components of growth-regulatory signaling networks. For example, Rane (1999) demonstrated that there is a calcium-activated potassium channel that is up-regulated by growth factors in fibroblasts. When this channel was blocked, cell proliferation was inhibited. Ion channels as targets for growth control of fungi is revealed by the finding that theU. maydis fungal toxin is capable of inhibiting a calcium channel (Gu et al., 1995).
Fungi, including C. neoformans, are among the most common primary opportunistic infections to occur in HIV-infected patients (Sugar, 1991; Kwon-Chung and Bennett, 1992; Stansell, 1993;Mitchell and Perfect, 1995). Familiar fungal pathogens have become increasingly common clinical problems, with new fungal pathogens also now being recognized (Perfect and Schell, 1996). Treatments for pathogenic fungi have been moderately successful but have been associated with side effects and resistant strains (Graybill, 1996). There has also been a rise in infections by fungi, such asAspergillus, that are innately resistant to common therapeutics, such as fluconazole. Furthermore, the commonly used therapeutics are fungistatic and rely on the host's immune system for clearing the growth-arrested cells, which is a problem in immunocompromised patients. Thus, there is an ongoing need to identify new fungal targets and therapeutics (DiDomenico, 1999), in particular those that have fungicidal activity. Although it is tempting to conclude from our data that amiodarone may be just such an antifungal, there are significant adverse side effects of long-term amiodarone use that preclude its systemic use.
Amiodarone was developed over 30 years ago as an antianginal but is now widely used as a potent antiarrhythmic drug (Mason, 1987; Gill et al., 1992; Roden, 1996; Singh, 1996). It has been successfully used in the treatment of symptomatic and life-threatening ventricular arrhythmias and severe supraventricular arrhythmias. The plasma concentrations of amiodarone that are used clinically for treatment of arrhythmia can be in the range of 1 to 4 μM during loading periods. However, in research studies, plasma concentrations in the 10 μM range have been achieved. These are concentrations that are marginally effective against Cryptococcus in vitro. It was originally described as a class III antiarrhythmic because of its ability to increase the action potential duration. However, class I antiarrhythmic effects and a propensity to noncompetitively antagonize α- and β-adrenergic receptors have also been seen. Although its precise mode of action remains uncertain, amiodarone has been shown to block potassium channels and inactive sodium channels (more so than activated ones), as well as calcium channels (Kodama et al., 1999; Nattel and Singh, 1999;Dorian, 2000). Amiodarone has also been shown to directly activate G proteins (Hagelüken et al., 1995). The results presented here are intriguing in that growth inhibition of amiodarone depends only on its effects on calcium homeostasis but not that of sodium or potassium. A significant difference between the effect of amiodarone in yeast versus mammalian cells is that it induces an immediate calcium efflux in yeast, whereas it blocks channel activity in mammalian cells. The ability of amiodarone to induce calcium efflux suggests that amiodarone may be a useful new tool to study calcium homeostasis in yeast. Moreover, the ability to study amiodarone effects in yeast may provide insights into its molecular activity in higher eukaryotes.
In summary, amiodarone has broad-spectrum antifungal activity and demonstrates a rapid fungicidal effect. Amiodarone promotes a rapid efflux of calcium, suggesting that calcium homeostasis is responsible, at least in part, for its inhibitory effects. The potent antifungal activity of amiodarone reveals a new mechanism for future antifungal development. Because amiodarone was developed as an antianginal, it is quite possible that derivatives may have increased antifungal activity and reduced toxicity. In addition, research into the molecular mechanism responsible for its antifungal activity may suggest alternative agents that affect the same target.
I thank John Sutko for enlightening discussions.
This study was supported by National Science Foundation Grant MCB-9513713 (to W.E.C.).
- acquired immunodeficiency syndrome
- 0.17% (w/v) Difco yeast nitrogen base without amino acids and ammonium sulfate, 0.4% (w/v) ammonium sulfate, 2% (w/v) glucose
- dimethyl sulfoxide
- human immunodeficiency virus
- Received September 10, 2001.
- Accepted October 8, 2001.
- The American Society for Pharmacology and Experimental Therapeutics