Published ahead of print on February 28, 2008, doi:10.1165/rcmb.2007-0217OC
© 2008 American Thoracic Society DOI: 10.1165/rcmb.2007-0217OC Amiodarone Alters Late Endosomes and Inhibits SARS Coronavirus Infection at a Post-Endosomal Level 1 Novartis Vaccines, Siena, Italy; 2 Cardiovascular Therapy Research Laboratory, Department of Internal Medicine, University Hospital of Zurich, Zurich, Switzerland; 7 Institute of Organic Chemistry, University of Zurich, Zurich, Switzerland; 4 Venetian Institute of Molecular Medicine, University of Padova, Padova, Italy; 5 Divisione di Chirurgia Toracica, and Departments of 3 Oncological and Surgical Sciences, 6 Clinical Medicine, 8 Pediatrics, and 9 Medical and Surgical Sciences, University of Padova, Padova, Italy Correspondence and requests for reprints should be addressed to Aldo Baritussio, M.D., Clinica Medica I, Department of Medical and Surgical Sciences, University of Padova, 35128 Padova, Italy. E-mail aldo.baritussio{at}unipd.it
Amiodarone interferes with the endocytic pathway, inhibits proteolysis, and causes the formation of vacuoles, but uptake and intracellular distribution of the drug, origin of vacuoles, and functional consequences of amiodarone accumulation remain unclear. Our objective was to study amiodarone uptake, clarify the origin of vacuoles, and investigate the effect of amiodarone on the life cycle of the coronavirus responsible for the Severe Acute Respiratory Syndrome (SARS), which, to enter cells, relies on the proteolytic cleavage of a viral spike protein by the endosomal proteinase cathepsin L. Using alveolar macrophages, we studied uptake of 125I-amiodarone and 125I-B2, an analog lacking the lateral group diethylamino-β-ethoxy, and analyzed the effects of amiodarone on the distribution of endosomal markers and on the uptake of an acidotropic dye. Furthermore, using Vero cells, we tested the impact of amiodarone on the in vitro spreading of the SARS coronavirus. We found that (1) amiodarone associates with different cell membranes and accumulates in acidic organelles; (2) the diethylamino-β-ethoxy group is an important determinant of uptake; (3) vacuoles forming upon exposure to amiodarone are enlarged late endosomes; (4) amiodarone inhibits the spreading in vitro of SARS coronavirus; and (5) trypsin cleavage of the viral spike protein before infection, which permits virus entry through the plasma membrane, does not impair amiodarone antiviral activity. We conclude that amiodarone alters late compartments of the endocytic pathway and inhibits SARS coronavirus infection by acting after the transit of the virus through endosomes.
Key Words: amiodarone • endocytic pathway • SARS coronavirus
Amiodarone, an antiarrhythmic agent used to treat supraventricular and ventricular arrhythmias, has peculiar pharmacokinetics. After intestinal absorption or intravenous administration, it leaves the serum compartment with a half-life of 10 to 20 hours and accumulates extensively in peripheral tissues (1). Two characteristics seem important in determining its distribution through the body: the extreme hydrophobicity, which may explain the preferential accumulation into structures rich of lipids, and the presence of a diethylamino-β-ethoxy group, which may favor the association with acidic organelles (2–4). In fact, amiodarone inhibits the degradation of surfactant protein A (SP-A) (a calcium-dependent lectin taken up by alveolar macrophages and degraded into lysosomes) (4); decreases the size of the lysosomal compartment (4); and causes the accumulation of lyso-bis-phosphatidic acid, a rare phospholipid associated with the internal vesicles of late endosomes/multivesicular bodies (2, 5). On the basis of these findings, it has been suggested that amiodarone may affect late compartments of the endocytic pathway, but the mechanism of uptake of the drug, its cellular distribution, and the role played by the diethylamino-β-ethoxy group remain unclear. A peculiar consequence of amiodarone administration is the formation of vacuoles, multilamellar inclusion bodies and crystalline-like deposits in different cell types; however, the origin of these structures is also unknown (3, 6). To clarify some of these points, using alveolar macrophages, we analyzed uptake and intracellular distribution of 125I-amiodarone and 125I-B2, an analog lacking the diethylamino-β-ethoxy group (7). Furthermore, using confocal microscopy, we analyzed the effects of amiodarone on the distribution of markers of early and late endosomes and on the uptake of an acidotropic dye. Finally, since infection of target cells by the coronavirus (CoV) responsible of the Severe Acute Respiratory Syndrome (SARS) relies on the endosomal cleavage of a viral surface protein, we also studied the impact of amiodarone on the in vitro spreading of this newly emergent infectious agent (8). SARS-CoV is an enveloped virus, with a positive-stranded RNA genome, presenting club-shaped protrusions made by trimers of the spike (S) protein. The first step in the viral life-cycle is the binding of the S protein to its receptor, the angiotensin-converting enzyme 2 (ACE2). Subsequently the virus is taken up by receptor-mediated endocytosis, ending into an acidic endosomal compartment where the S protein is proteolitically cleaved by the endosomal protease cathepsin L. Protein S then triggers the mixing of viral and endosomal membranes, causing the release of the viral genome into the cytoplasm (9–16). Having found that amiodarone perturbs late compartments of the endocytic pathway, we tested its impact on the life-cycle of the SARS-CoV.
Materials Antibodies were from Stressgen Biotechnologies, Eugene, OR (Rab4 and Fas); Ancell, Bayport, MN (CD63), Abcam, Cambridge, UK (EEA1); and Athens Research and Technology, Athens, CA (cathepsin L). Amiodarone was from Sigma (St. Louis, MO). B2 (2-butyl-3-(3',5'-diiodo-4-hydroxy)-benzofuran) was synthesized as reported (7). 125I-amiodarone (specific activity 34.7 Ci/mmol) and 125I-B2 (3.36 Ci/mmol) were prepared following the method of Bellen and coworkers (17). Water solubility and interaction with membranes of amiodarone and B2 were measured as reported (18, 19). Vero cells (ATCC CCL-81) were maintained in GlutaMAX I (Gibco, Carlsbad, CA), 1% penicillin-streptomycin, 5% heat-inactivated fetal bovine serum (HyClone, Logan, UT). Alveolar macrophages, obtained after institutional approval and written consent from lung lobes resected for lung cancer, were maintained in Ringer Buffered 0.1% Albumin (RBA) (4).
Binding/Uptake of 125I-Amiodarone and 125I-B2
Immunofluorescence
Uptake of Lysosensor
Effect on Cathepsin L
Effects of Amiodarone on SARS-CoV Infection
Trypsin Treatment
Statistical Analysis
Physicochemical Properties of Amiodarone and B2 Amiodarone was more water-soluble than B2, especially at acidic pH. In fact, as the pH was lowered from 7.4 to 5.0 the solubility in water of amiodarone increased from 0.069 ± 0.005 to 2.04 ± 0.11 µM, while that of B2 increased from 0.009 ± 0.0001 to 0.10 ± 0.01 µM (differences between amiodarone and B2 and between pH5.0 and 7.4 all statistically significant). The affinity of amiodarone for membranes, expressed as logkIAM, was significantly greater than that of B2 at pH 7.4 (respectively 4.23 ± 0.02 versus 3.20 ± 0.04), but smaller than that of B2 at pH 5.0 (3.63 ± 0.04 versus 4.57 ± 0.08). Recently Bergstrom and coworkers reported that the solubility in water of amiodarone increases by 5 orders of magnitude as the pH is lowered from 7.4 to 5.0 (18). The reason for the discrepancy with our findings is unclear. It could be due to the fact that we started with 4 µM amiodarone in methanol (the highest concentration possible), while Bergstrom and colleagues started with an excess of solid drug.
Binding/Uptake of Amiodarone and B2 by Alveolar Macrophages As shown in Figure 1A, amiodarone associated rapidly with alveolar macrophages. Binding/uptake was linear over time and was critically dependent upon the presence of the diethylamino-β-ethoxy group, since 125I-B2, which lacks this group, associated with macrophages 50 times more slowly than amiodarone (Figure 1).
Clathrin-mediated endocytosis, caveolae, the vectorial transfer of materials through the endocytic pathway, fluid phase endocytosis, or the cytoskeleton can be inhibited or perturbed without change in the net uptake of amiodarone. In fact, phenylarsine oxide and amantadine (inhibitors of clathrin-mediated endocytosis), filipin (inhibitor of endocytosis through caveolae), cytochalasin D (a disassembler of the cytoskeleton), and wortmannin (which inhibits fluid phase endocytosis, phagocytosis, and the vectorial movement of materials through the endocytic pathway) did not influence the association of 125I-amiodarone with macrophages (Figure 1B). On the other side, bafilomycin A1 and concanamycin A, which dissipate vacuolar acidity by inhibiting V-ATPase, decreased uptake without abolishing it (Figure 1B). Since the inhibitors used had no detrimental effect on cell viability, these results suggest that there is no preferential mechanism of uptake of amiodarone and that amiodarone, besides accumulating into acidic organelles, associates also with other cell structures. In fact, in a series of experiments in which we incubated macrophages with 125I-amiodarone for 0, 30, 60, and 120 minutes and then subjected the postnuclear supernatant to a series of differential centrifugations each lasting 10 minutes (3,000 x g, 6,000 x g, 10,000 x g, 20,000 x g, 100,000 x g), amiodarone accumulated in all pellets. Most of the radioactivity (over 85% of total) was associated with the 3,000 x g, 6,000 x g, and 10,000 x g pellets, while the 100,000 x g supernatant contained at any time less than 2% of total radioactivity (not shown), indicating that free iodine represented a minor fraction of total radioactivity.
Distribution of Amiodarone among Organelles of the Postnuclear Supernatant
As shown in Figure 2, when macrophages were incubated with 0.2 µM 125I-amiodarone, the radioactivity gradually accumulated into organelles migrating at the extremes of the gradient, collecting to a greater extent toward the bottom. Treatment with bafilomycin decreased markedly the radioactivity associated with both populations of organelles (Figure 2). Thus it appears that amiodarone, presented to cells at low concentrations, associates in a time-dependent manner with diverse organelles of the postnuclear supernatant, some of which can be identified as endosomes and lysosomes on the basis of sedimentability, distribution of markers, and sensitivity to treatment with bafilomycin A1. These results were obtained from macrophages incubated with 0.2 µM amiodarone, while patients are chronically exposed to concentrations ranging from 1 to 4 µM. To test the effect of higher concentrations of amiodarone, we incubated alveolar macrophages for 16 hours both with 0.2 and with 10.0 µM 125I-amiodarone, and then analyzed the distribution of radioactivity among organelles of the postnuclear supernatant. As shown in Figure 3, the concentration of amiodarone had marked effects on the distribution of radioactivity. In fact, when present at low concentration, amiodarone distributed mostly toward the bottom of the gradient, suggesting a preferential accumulation into lysosomes. At high concentrations amiodarone accumulated mostly toward the top of the gradient, suggesting a preferential accumulation into endosomes.
125I-B2 was taken up by alveolar macrophages to such a low extent that the radioactivity of the postnuclear supernatant was barely greater than background (not shown).
Vesicles Forming upon Exposure to Amiodarone
Effect of Amiodarone on the Uptake of Lysosensor DND-189 Amiodarone changes dramatically the distribution of Lysosensor Green DND-189, an acidotropic dye that accumulates into acidic organelles as the result of protonation and whose fluorescence intensity is proportional to acidity (24). In fact, control macrophages accumulated Lysosensor as small vesicles clustered near the nucleus and also displayed a diffuse staining of the cytoplasm, while macrophages treated with 10 µM amiodarone accumulated Lysosensor in large vacuoles located mostly at the cell periphery (Figure 5A). Besides changing the distribution of Lysosensor DND-189, amiodarone decreased the net uptake of the probe, as indicated by the decrease of fluorescence emitted by whole cells (Figure 5B). The effect was dose dependent and significant at 2 µM amiodarone. The lower uptake of Lysosensor in the presence of amiodarone appears to have been due to decreased uptake rather than to increased release of the probe, since cells loaded with Lysosensor and then exposed to amiodarone lost their fluorescence more slowly than cells loaded with Lysosensor and then exposed to plain medium (not shown).
Effect of Amiodarone on Cathepsin L As shown in Figure 6, 10 µM amiodarone, did not change the total amount of catehepsin L present in lysates obtained from alveolar macrophages and did not affect the maturation of this enzyme, since the ratio between the immature procathepsin L (38 kD), the intermediate 30-kD form and mature cathepsin L (24, 25 kD) remained unchanged (25).
Effect of Amiodarone on SARS-CoV Infectivity It has previously been shown that amiodarone inhibits the degradation of surfactant protein A, resulting in the accumulation of the uncleaved protein in an endosomal compartment (4). Since infection of target cells by the SARS-CoV is reliant on the proteolytic cleavage of the S protein by the lysosomal cysteine proteinase cathepsin L (13, 14), we explored the possibility of an effect of amiodarone on the life-cycle of the SARS-CoV. All work with live SARS-CoV was done in a biosafety level 3 facility at Novartis Vaccines, Siena, Italy. Preliminary experiments showed that amiodarone is taken up by Vero cells, the uptake being comparable with that of alveolar macrophages considering kinetics and sensitivity to inhibitors. Vero cells were incubated for 2 hours with 0 to 50 µM amiodarone and then were infected with SARS-CoV (m.o.i. of 1). One hour after infection the cells were washed and further incubated in growth medium in the presence of the drug. The next day the cell culture supernatant was harvested and the virus was quantified by titration. As shown in Figure 7A, which presents the results of four independent experiments, amiodarone affected the life-cycle of SARS-CoV in a concentration-dependent manner: 5 µM amiodarone induced a significant diminution of the virus titer; 10 µM amiodarone decreased the virus titer 10 times; 50 µM amiodarone brought the virus titer below the detection limit, which was 10 TCID50/ml in this assay.
Amiodarone can be toxic to cells (4, 7) and its antiviral activity could be due to decreased cell viability. To rule this out, Vero cells were incubated for 24 hours with 0 to 50 µM amiodarone and then their ability to incorporate propidium iodide was analyzed using a FACSCalibur (Becton Dickinson, San Jose, CA). We found that cell viability was greater than 97% up to 30 µM amiodarone, 91% at 40 µM amiodarone, and 83% at 50 µM amiodarone. Thus amiodarone has antiviral activity at concentrations at which it has no effect on viability. The antiviral effect of amiodarone could be due to a direct effect of the drug on the SARS-CoV. To clarify this point, SARS-CoV in growth medium was incubated at 37°C for 24 hours without or with 10 µM amiodarone. Before titration on Vero cells, the two admixtures were diluted 10 times, so that the amiodarone concentration decreased to 1 µM (which has no effect on the propagation of SARS-CoV). We found no significant difference in virus titer between the two deposits (data not shown). In a further experiment SARS-CoV was incubated at 37°C for 24 hours with 5 or 50 µM amiodarone. Then the 50-µM amiodarone-virus admixture was diluted with fresh medium to 5 µM and both admixtures were titrated on Vero cells. We found no significant difference in virus titer between the two deposits (data not shown). We conclude, therefore, that a direct interaction of amiodarone with the SARS-CoV is unlikely or not significant. SARS-CoV enters into target cells by binding through the S protein to the surface receptor ACE2. The virus is then taken up, ending in an acidic endosomal compartment where, after proteolysis by cathepsin L, protein S triggers the release of the viral genome into the cytoplasm. A mildly acidic pH environment in the endosome seems to be important, since infection can be blocked by lysosomotropic agents like NH4Cl or chloroquine (13). In principle amidarone could decrease the infectivity of SARS-CoV by decreasing the density of surface receptors, by interfering with organelle acidification or by interfering with synthesis, maturation, or activity of cathepsin L. These mechanisms, however, are unlikely. In fact, amiodarone does not modify the density of ACE2 receptors on the cell surface, as determined by flow-cytometry using a human monoclonal antibody against ACE2 (not shown); does not inhibit organelle acidification (Ref. 4 and Figure 5); and does not interfere with synthesis and maturation of cathepsin L (Figure 6). In a final experiment we tested whether amiodarone interferes with early stages of the virus life cycle, in particular those involving the endosomes, where the drug appears to be active. We reasoned that if the endosomes were the only site of interefence of amiodarone with the virus life cycle, then it could be possible to restore virus infectivity by trypsin treatment of SARS-CoV bound to target cells (which enables the virus to fuse directly with the plasma membrane and to infect cells without using the endosomal pathway). Accordingly, Vero cells were pre-incubated with amiodarone and then the virus was allowed to bind to the cells at 4°C. Afterward trypsin was added, to allow the fusion of virus and cell membranes and the direct release of the virus genome into the cytosol (15). Thereafter, cells were incubated in growth medium overnight and virus titer was determined. As shown in Figure 7B, trypsin treatment of cell-bound virus could overcome the infection block mediated by ammonium chloride, the control in this experiment, but not the block induced by amiodarone, strongly indicating that amiodarone interferes with the SARS-CoV life cycle after delivery of the viral genome in the cytosol.
Binding/Uptake and Distribution of Amiodarone Association of amiodarone with cells is greatly influenced by the lateral group diethylamino-β-ethoxy. In fact, at pH 7.4 amiodarone had greater solubility in water and higher affinity for membranes than B2, while at pH 5.0 amiodarone had greater solubility in water but lower affinity for membranes than B2, indicating that the diethylamino-β-ethoxy group may favor both the transfer of the drug from carrier proteins to the plasma membrane and accumulation into acidic organelles. It is clear, however, that association with acidic organelles is just one of the ways in which amiodarone accumulates inside cells, since the dissipation of vacuolar acidity decreases but does not abolish uptake. From the present evidence, we propose that amiodarone could associate with the plasma membrane and then distribute to different cell compartments according to the section of the plasma membrane in which it happens to be. Amiodarone ending into early endosomes will follow the fate of these organelles along the endocytic pathway and, as the lumenal pH drops, it could move from the limiting membrane to the lumen. This interpretation can explain the distribution of the drug to all cell membranes, the lack of a preferential mechanism of uptake, the association with acidic organelles, the decreased uptake after dissipation of the vacuolar acidity, and the preferential association with lysosomes (which have the lowest internal pH).
Effects of Pharmacologic Concentrations of Amiodarone Biogenesis and proper functioning of late compartments of the endocytic pathway depend on a highly conserved set of protein (Hrs, ESCRT I-III, Alix) that assemble on the cytoplasmic side of early endosomes and make them mature into late endosomes/multivesicular bodies (27, 28). Recently it has been found that the suppression of hVps34, which converts phosphatidylinositol to phosphatidylinositol 3-phosphate (the docking site for SCRTIII), generates changes similar, although not identical, to those induced by amiodarone (29). These changes include sparing of early endosomes, formation of enlarged late endosomes, impaired endo-lysosomal proteolysis, and subtle disturbances in the maturation of cathepsin D (29). Thus, considering that ESCRT III is involved in the formation of multivesicular bodies (27, 28), it is tempting to speculate that amiodarone might act on a related target and that enlarged late endosomes might derive from the decreased formation of intralumenal vesicles. We found previously that incubation of alveolar macrophages with 10.0 µM amiodarone decreases the size of the lysosomal compartment (4). This might explain the observation that, when presented to macrophages at low (0.2 µM) concentration, 125I-amiodarone distributed mostly to lysosomes, whereas when presented at higher (10.2 µM) concentration, it distributed mostly to endosomes. Shrinking of the lysosomal compartment could also explain the decreased uptake of Lysosensor by macrophages incubated with amiodarone.
Effects of Amiodarone on SARS-CoV Infection Endosomal compartments and their proper functioning play an important role in the life cycle of the SARS-CoV; therefore, we thought that amiodarone, which inhibits endo-lysosomal hydrolysis and interferes with late compartments of the endocytic pathway, could interfere with virus entry into target cells. We found that amiodarone inhibits SARS-CoV infectivity and that inhibition appears to take place at a post-endosomal level, since amiodarone does not influence the surface density of ACE2, does not dissipate luminal acidity, does not impair synthesis and maturation of cathepsin L, and cannot be antagonized by the treatment of cell-bound virus with trypsin, which allows virus entry through the plasma membrane. How amiodarone inhibits SARS-CoV amplification remains at present unclear; however, considering its effects on cell morphology (4, 7), it is tempting to speculate that it might interfere with steps of the virus life cycle involving cell membranes. For example, amiodarone could interfere with the assembly or function of virus-specific, flask-shaped cytoplasmic compartments bordered by a double membrane into which novel viral RNA is synthesized (32–34). Alternatively, amiodarone might interfere with the assembly of structural proteins and genomic RNA into new virions or might inhibit the release of progeny virus in the extracellular space, since the protein complexes necessary for the biogenesis of multivesicular bodies (the ESCRTs) also mediate virus budding (27). Furthermore, it is still possible that the antiviral effects of amiodarone could be occurring also in the endosomes. Clearly, the evidence of an activity of amiodarone against the SARS-CoV is preliminary and its applicability in the clinic is open to question, since the concentrations needed to obtain an antiviral effect in vitro are greater than serum levels found in patients treated for cardiac arrhythmias. We feel, however, that this subject deserves further scrutiny for several reasons. The first reason is that in vivo amiodarone is transformed into metabolites, which retain many of the activities displayed by the parent drug (7), and thus the antiviral activity of amiodarone would be much clarified by a test in an animal model of SARS, in the hope that amiodarone and its metabolites could cooperate against the virus. Another reason to study in vivo the antiviral activity of amiodarone stems from changes in virus tropism during the infection, since it has been suggested that at the start of the infection SARS-CoV may enter target cells through the endosomal pathway, while at later stages inflammation could liberate a variety of proteases that might induce virus entry through the plasma membrane, causing a further increase in the replication rate (15). We hypothesize that amiodarone could alleviate this mechanism of amplification, since in vitro its antiviral activity appears to be independent from the mechanism of virus entry into target cells. A further reason to perform protection studies in animals is that amiodarone could be tried in association with drugs that affect the virus life cycle by a different action mechanisms (16). We conclude that amiodarone perturbs late compartments of the endocytic pathway and inhibits SARS coronavirus infection, acting after the delivery of the viral genome into the cytosol.
The authors thank professor Federico Rea for his help with the isolation of human alveolar macrophages.
* Present affiliation: International Vaccine Institute, Seoul, Korea. 
This work was supported in part by Fondi ex 60% to A.B. and by a generous gift from Mr. Antonio Fiore. Originally Published in Press as DOI: 10.1165/rcmb.2007-0217OC on February 28, 2008 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 12, 2007 Accepted in final form November 17, 2007
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
CLINICAL RELEVANCE
Present affiliation: Department of Internal Medicine, Section of Digestive Diseases, Yale University, New Haven, Connecticut.