ABSTRACT
Despite recent advances in diagnostic and therapeutic methods in antifungal research, aspergillosis still remains a leading cause of morbidity and mortality. One strategy to address this problem is to enhance the activity spectrum of known antifungals, and we now report the first successful application of Candida antarctica lipase (CAL) for the preparation of optically enriched fluconazole analogues. Anti-Aspergillus activity was observed for an optically enriched derivative, (−)-S-2-(2′,4′-difluorophenyl)-1-hexyl-amino-3-(1‴,2‴,4‴)triazol-1‴-yl-propan-2-ol, which exhibits MIC values of 15.6 μg/ml and 7.8 μg/disc in broth microdilution and disc diffusion assays, respectively. This compound is tolerated by mammalian erythrocytes and cell lines (A549 and U87) at concentrations of up to 1,000 μg/ml. When incorporated into dextran nanoparticles, the novel, optically enriched fluconazole analogue exhibited improved antifungal activity against Aspergillus fumigatus (MIC, 1.63 μg/ml). These results not only demonstrate the ability of biocatalytic approaches to yield novel, optically enriched fluconazole derivatives but also suggest that enantiomerically pure fluconazole derivatives, and their nanotized counterparts, exhibiting anti-Aspergillus activity may have reduced toxicity.
INTRODUCTION
Aspergillosis remains a significant threat to public health, and in spite of continuous efforts to improve timely diagnosis and clinical therapies, mortality caused by this disease remains unacceptably high (1, 2). Current therapeutic options for treating Aspergillus-induced disorders include antifungal agents such as polyenes, azoles, and echinocandins (3, 4). Thus, the discovery of new antifungal compounds remains important given the need to address the development of drug resistance in pathogenic fungi (5–7). One approach to accomplishing this goal is to prepare new derivatives of existing drugs with broad-spectrum activity and enhanced pharmacokinetic properties. As part of our ongoing efforts to use lipases (8–12), which catalyze reactions with high degrees of chemo-, regio-, and stereoselectivity in organic synthesis, we became interested in preparing new antifungals using biocatalysis.
Fluconazole, introduced in 1990, is a bis-triazole antifungal drug which possesses interesting pharmacokinetic properties, such as low plasma binding affinity, good water solubility, low first-pass metabolism, high oral bioavailability, and a long half-life, all of which should make it a drug of choice for treating fungal infections (13, 14). On the other hand, fluconazole has been reported to exhibit only limited activity against Aspergillus infections (15), which has led to many reports concerning the synthesis of various types of fluconazole derivatives and their chiral separation/resolution into constituent enantiomers (16–18; H. B. Borate, S. P. Sawargave, S. P. Chavan, M. A. Chandavarkar, R. R. Iyer, V. V. Nawathye, G. J. Chavan, A. C. Tawte, and D. D. Rao, 20 September 2012, international patent application WO 2012123952 A1). We now report the use of Candida antarctica lipase (CAL-B) in catalyzing the addition of amines to an achiral epoxide to yield optically enriched fluconazole analogues in which one of the triazole rings is replaced by n-alkylamino and cycloalkylamino substituents. To the best of our knowledge, the work reported here is the first direct synthesis of optically enriched fluconazole analogues using biocatalytic methods. In vitro assays show that the optically enriched analogues exhibit more potent antifungal activity than the corresponding racemic mixtures. Interestingly, this bioactivity can be enhanced by their encapsulation in dextran-based nanoparticles (19).
RESULTS AND DISCUSSION
Synthesis of fluconazole analogues.A series of linear and cyclic alkylamines was screened for reaction with the epoxide ring of (±)-1-[2-(2,4-difluorophenyl)-oxiranylmethyl]-1H-[1,2,4]-triazole (compound 1, Fig. 1) in a number of different organic solvents. Three different immobilized lipases were evaluated for their ability to catalyze this reaction: Candida rugosa lipase (CRL), porcine pancreatic lipase (PPL), and CAL-B. Although the ring-opening reactions catalyzed by CRL and PPL were of no practical utility, when the reaction was performed in the presence of CAL-B in tetrahydrofuran (THF) as solvent, the desired products (3a to -j, 5a, and 5b) were obtained with good yields in optically enriched forms (Fig. 1; also see Tables S1 and S2 in the supplemental material). Very importantly, all of the 12 novel fluconazole analogues formed in the lipase-catalyzed reactions were optically active, showing that aminolysis of the racemic starting epoxide (±)-1 had proceeded in an enantioselective fashion (Table S1). These 12 compounds could also be prepared in racemic form, as viscous oils in 75 to 80% yields, by direct reaction of the alkylamines with the racemic epoxide precursor (±)-1 in THF at 55°C. The time taken for complete consumption of aliphatic amines 2a to -j, 4a, and 4b in the CAL-B-catalyzed reaction varied between 18 h and 28 h, which was considerably shorter than the 48 to 56 h required for the chemical addition of the amines (Table S2). The structures of all 12 fluconazole analogues were unambiguously established on the basis of spectroscopic data (infrared [IR], 1H and 13C nuclear magnetic resonance [NMR], and mass spectra) and by comparison to literature data for known compounds 3b, 3c, 5a, and 5b (20, 21).
CAL-B-catalyzed epoxide ring opening with open chain and cyclic aliphatic amines. Note that samples of each compound could also be prepared in racemic form by heating the epoxide and amine at 55°C in THF (see the supplemental material). The new stereogenic center is indicated by an asterisk.
Although the enantiomeric enrichment of the fluconazole analogues prepared by lipase-catalyzed addition was not established, we were able to assign the absolute configuration of the major enantiomer using the optical activity of the unreacted epoxide isolated from the reaction mixture. These samples rotated polarized light in a positive (+) direction, meaning that the recovered, unreacted epoxide was enriched in the enantiomer for which the stereogenic center has the (S) configuration (Table S1) (22). CAL-B therefore preferentially employs (−)-R-1 in the aminolysis reaction, and assuming a standard SN2 mechanism for reaction of the amine with the epoxide, we can deduce that the fluconazole analogues must be enriched in the (−)-S-enantiomer (Fig. 1).
Antifungal activities of the fluconazole analogues.Pathogenic Aspergillus strains (Aspergillus fumigatus ITCC 6604, Aspergillus flavus ITCC 5192, and Aspergillus niger ITCC 0004) were used to determine the in vitro antifungal efficacy of the fluconazole analogues, in both their optically enriched and racemic forms. These experiments used standard broth microdilution (MDA), disc diffusion (DDA), and percent spore germination inhibition (PSGI) assays (23, 24). We note that the MDA is based on the same basic principle as that used in the CLSI microdilution protocol. The only difference between the two assays is that the CLSI protocol uses RPMI medium to prepare diluted drug solutions rather than the Sabouraud dextrose broth (a medium used to culture aspergilli in the laboratory) used by us to determine the MIC of the fluconazole derivatives. As recommended in CLSI protocols, we carefully monitored MDA parameters with respect to preparation of the test compounds, medium preparation, temperature, inoculum size, incubation time, MIC/endpoint determination, data recording, and interpretation of results to ensure the validity and quality of our results. On this point, we note that a previous study from our laboratory (25) showed that results with RPMI 1640 or RPMI 1640 containing glucose were not different from those obtained by using Sabouraud dextrose broth.
On the basis of their MIC values, all the compounds exhibited moderate to good anti-Aspergillus activities, with the analogue (−)-S-3d being more potent than the commercially available fluconazole (Table 1). We also observed that optically enriched mixtures of (−)-S-3a, (−)-S-3c, (−)-S-3d, (−)-S-3e, and (−)-S-5b were more active than the corresponding racemates. These data also confirm that introducing a linear aliphatic alkyl side chain is important for imparting antifungal activity, as reported previously (22, 23). On the other hand, when additional, “distal” N-substituted alkyl groups were present, as in compounds (−)-S-3g and (−)-S-3 h, antifungal activity was completely lost (Table 1). Compounds 3j and 5a exhibited no biological activity in broth microdilution assays and were not studied further. Our work also shows that the length of the alkyl side chain is an important factor in determining activity, i.e., the compound (−)-S-3d, containing an n-hexyl moiety, has higher activity than (−)-S-3a, (−)-S-3b, and (−)-S-3c, which contain ethyl, n-propyl, and n-butyl groups, respectively (Table 1). Decreasing the linker chain length also led to higher activity. Optically enriched (−)-S-3d was the most potent compound against Aspergillus fumigatus (Table 1) and was therefore used to examine how encapsulation in dextran nanoparticles might impact antifungal activity.
In vitro activity of selected, optically enriched fluconazole analogues against Aspergillus species
Characterization of (−)-S-3d release from O-alkylated dextran nanoparticles.Dextran nanoparticle-based drug delivery systems are biocompatible and biodegradable, possess low immunogenicity (19), and can be used for controlled release of pharmacologically active substances (26). We therefore encapsulated optically enriched (−)-S-3d into three types of dextran nanoparticles, derivatized with O-hexadecyl, O-decyl, and O-heptyl chains to ensure amphiphilicity, and examined their effect on anti-Aspergillus activity. After trapping (−)-S-3d within each of the nanoparticles by self-assembly (encapsulation efficiencies for the O-hexadecyl, O-decyl, and O-heptyl nanoparticles were 50% ± 4%, 22% ± 2%, and 30% ± 2%, respectively), the resulting particle size distributions were determined using dynamic light scattering (DLS). These measurements showed that the sizes of the O-hexadecyl-, O-decyl-, and O-heptyl-derivatized nanoparticles were 140 ± 16 nm, 187 ± 13.16 nm, and 183 ± 14.73 nm, respectively, and that all of the samples had a low polydispersity index (<0.3) (see the supplemental material). Examination of the rate at which the fluconazole analogue (−)-(S)-3d was released from each of the three types of nanoparticles showed an initial burst for the O-hexadecyl- and O-decyl-derivatized nanoparticles (Fig. 2).
In vitro release of (−)-S-3d from O-hexadecyl- (gray squares), O-decyl- (orange circles), and O-heptyl-derivatized (blue triangles) dextran nanoparticles.
Anti-Aspergillus activity and cytotoxicity of (−)-S-3d encapsulated in O-alkylated dextran nanoparticles.We next examined the effect of nanoparticle encapsulation on the activity of (−)-S-3d against Aspergillus fumigatus using a broth microdilution assay (Fig. 3). After 48 h of incubation (approximately 80% release), (−)-S-3d encapsulated in O-decyl-derivatized nanoparticles inhibited the growth of Aspergillus fumigatus at an effective concentration of 3.16 μg/ml. Perhaps more importantly, when the optically enriched fluconazole analogue was encapsulated in O-hexadecyl nanoparticles, complete inhibition of Aspergillus fumigatus growth was achieved at an effective concentration of 1.63 μg/ml (41.3% release at an initial concentration of 3.95 μg/ml). In addition, nanoparticle-encapsulated (−)-S-3d exhibits activity at a lower concentration than both fluconazole and free (−)-S-3d. Although we believe that this effect is associated with sustained release of the compound over time, it is also possible that drug uptake is more efficient because the drug in its encapsulated form is more efficiently captured by the cells. The general importance of this observation is also evident from the fact that the MIC of amphotericin B was decreased from 1.95 μg/ml to 0.97 μg/ml when the drug was encapsulated in O-heptyl nanoparticles.
In vitro antifungal activity of (−)-S-3d, amphotericin B, and their dextran nanoparticles. Lanes: a, negative control; b, empty O-alkyl dextran nanoparticles; c, amphotericin B; d, fluconazole; e, (−)-S-3d; f, O-heptyl nanoparticles containing (−)-S-3d; g, O-decyl nanoparticles containing (−)-S-3d; h, O-hexadecyl nanoparticles containing (−)-S-3d; i, O-heptyl nanoparticles containing amphotericin B; j, O-decyl nanoparticles containing amphotericin B; k, O-hexadecyl nanoparticles containing amphotericin B.
The cytotoxicity of (−)-S-3d and amphotericin B when encapsulated in derivatized nanoparticles was also evaluated using hemolysis and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT)-based assays (Fig. 4). Perhaps unsurprisingly, given that erythrocytes and cell lines treated with empty dextran nanoparticles (>90% cell viability at concentrations of 2 mg/ml) remained completely viable up to 1 mg/ml, the encapsulated, optically enriched fluconazole analogue (−)-S-3d exhibited cytotoxicity similar to that of the free compound. Thus, essentially no toxicity to two human cell lines (Fig. 4b and c) was seen when the compound was present at concentrations similar to the MIC values observed for its antifungal activity. The optically enriched fluconazole analogue (−)-S-3d was also considerably less cytotoxic than free amphotericin B in all assays (Fig. 4). It is therefore interesting that encapsulating amphotericin B into O-hexadecyl-derivatized nanoparticles lowered the cytotoxicity of this antifungal agent in both the hemolysis and MTT-based assays. Nevertheless, cell viability was reduced for amphotericin B-containing nanoparticles relative to derivatized nanoparticles containing the fluconazole analogue (−)-S-3d.
In vitro cytotoxicity assays for optically enriched (−)-S-3d and amphotericin B both in the free form and when encapsulated into dextran nanoparticles. (a) Hemolytic assay; (b and c) MTT-based assay using A459 (b) and U87 (c) cell lines.
Conclusion.Reacting alkylamines with a racemic epoxide precursor (Fig. 1) in the presence of immobilized lipase CAL-B in THF provides a simple approach for the preparation of optically enriched fluconazole analogues, which appear to exhibit better antifungal activity against Aspergillus than fluconazole. Although the extent to which the enzyme catalyzes the coupling reaction in an enantioselective manner remains to be determined, we have been able to assign the (S) configuration to the stereogenic center of the enantiomer that exhibits biological activity, assuming that (i) the aminolysis reaction proceeds with its usual chemical mechanism and (ii) only one enantiomer has antifungal activity. Given the difficulty of single-step chemical strategies for the preparation of chiral fluconazoles in optically enriched form, we anticipate that the enzymatic methodology reported here will have significant impact on this approach to obtaining novel variants of existing antifungal drugs.
The most active analogue prepared in this study, (−)-S-3d, is more potent against Aspergillus fumigatus than fluconazole, having MIC values of 8 to 16 μg/ml in a series of in vitro assays. Perhaps more importantly for drug discovery, the anti-Aspergillus potency of this compound is enhanced (MIC, 1.6 to 4.0 μg/ml) by encapsulation in derivatized nanoparticles, with minimal in vitro cytotoxic effects at concentrations of up to 2 mg/ml against human erythrocytes and cell lines of human origin.
MATERIALS AND METHODS
General procedure for the CAL-B-catalyzed synthesis of optically enriched fluconazole analogues.CAL-B immobilized on Accurel beads (300 mg) was added to a solution of the epoxide (±)-1 (5.0 mmol) and the appropriate amine (2a to -j, 4a, or 4b; 2.5 mmol) dissolved in THF, and the mixture was incubated at 55°C. The extent of the reaction was monitored by thin-layer chromatography (TLC), and the enzyme was removed by filtration when the amine was consumed. After removal of THF at reduced pressure, the residue was subjected to column chromatography using chloroform-methanol as an eluent to afford optically enriched samples of pure fluconazole analogues (−)-S-3a to -3j, (−)-S-5a, or (−)-S-5b and the unreacted epoxide (+)-S-1.
(−)-S-2-(2′,4′-Difluorophenyl)-1-hexylamino-3-(1‴,2‴,4‴)triazol-1‴-yl-prop-an-2-ol (3d) was obtained as a viscous oil in 80% yield. [α]D20 −20.3 (c 0.01, CHCl3); IR spectrum (film) μmax: 3,315 (OH and NH), 2,979, 1,620, 1,508, 1,415, 1,267, 1,145, 960 cm−1. 1H NMR (400 MHz, CDCl3): δ 0.83 (3H, t, J = 7.63 Hz), 1.17 to 1.33 (8H, m), 2.43 (2H, t, J = 6.87 Hz), 2.81 (1H, d, J = 12.97 Hz), 3.12 (1H, d, J = 12.21 Hz), 4.49 (1H, d, J = 14.50 Hz), 4.58 (1H, d, J = 13.73 Hz), 6.74 to 6.82 (2H, m), 7.50 to 7.55 (1H, m), 7.77 (1H, s), and 8.10 (1H, s). 13C NMR (100 MHz, CDCl3): δ 13.92, 22.46, 26.55, 29.83, 31.49, 50.00, 54.14 (d, JCF = 3.83 Hz), 55.98 (d, JCF = 4.79 Hz), 72.96 (d, JCF = 5.75 Hz), 104.12 (d, JCF = 26.84 Hz), 111.38 (d, JCF = 20.61 Hz), 125.05 (d, JCF = 13.42 Hz), 129.79 (d, JCF = 6.71 Hz), 144.60, 151.09, 158.92 (d, JCF = 237.78 Hz), and 162.29 (d, JCF = 249.20 Hz). High-resolution mass spectrometry (HRMS): m/z 339.1991 ([M + H]+, C17H25F2N4O calculated 339.1969).
Broth microdilution assay.Various concentrations of different derivatives in the range of 0.24 to 1,000.0 μg/ml were prepared in 96-well culture plates (Nunc, Roskilde, Denmark) by serial dilution in Sabouraud dextrose broth. Wells were inoculated with 1 × 106 spores (conidia) of Aspergillus in 10 μl of spore suspension. Negative controls were solvent in medium and spores only, with amphotericin B and fluconazole being used as positive controls. Plates were incubated at 37°C using a bio-oxygen demand (BOD) incubator (Calton; NSW India) and examined macroscopically after 48 h for the growth of Aspergillus mycelia. The activity of the analogues was defined as positive if the medium appeared clear without any growth of Aspergillus mycelia, and the minimum concentration of compounds inhibiting growth was reported as MIC (Table 1).
Disc diffusion assay.Autoclaved Sabouraud dextrose agar (SDA) was poured into radiation-sterilized petri dishes (10.0-cm diameter). A suspension of conidia of Aspergillus was prepared and overlaid on the agar plates. Different concentrations of the fluconazole analogues were impregnated on 5.0-mm-diameter sterilized discs (Whatman no. 1) and placed on the agar. Control discs containing solvent, amphotericin B, or fluconazole were also included in the assay. Plates were incubated at 37°C, and the zone of inhibition was determined after 72 h. MICs reported for this assay (Table 1) correspond to fluconazole analogue concentrations giving a zone of inhibition of at least 6.0 mm in diameter from the center of the plate.
Percent spore germination inhibition assay.Serial dilutions, ranging from 0.24 to 1,000.0 μg/ml, of each fluconazole analogue dissolved in Sabouraud dextrose broth were placed in radiation-sterilized petri dishes (10.0-cm diameter), with each dish then being inoculated with 100 ± 5 Aspergillus conidia. After incubation for 16 h at 37°C, wells were examined for spore germination using an inverted microscope (Nikon Diaphot; Japan), and the number of germinated, and nongerminated, spores was recorded. MICs in this assay (Table 1) correspond to fluconazole analogue concentrations resulting in inhibition of spore germination.
In vitro cytotoxicity assays.Two approaches were performed to assess the cytotoxicity of the fluconazole analogues. First, using a standard hemolytic assay (27), erythrocytes from healthy individuals were suspended in phosphate-buffered saline (PBS) to give a 2% (vol/vol) suspension. These cells were then incubated with various concentrations of each compound for 1 h at 37°C before being pelleted by centrifugation at 3,000 × g for 10 min. The percent hemolysis was then calculated from the optical density at 450 nm of the supernatant (Fig. 4a). The effect of solvent and PBS on erythrocyte viability was also checked. Triton X-100 (Sigma Chemicals, USA) was used for complete hemolysis of the erythrocytes.
In an alternate approach, an MTT-based assay (28) was used to examine the cytotoxicity of the analogues against A549 (human pulmonary epithelial) and U87 (primary glioblastoma) human cell lines, obtained from the National Centre for Cell Science, Pune, India (Fig. 4b and c). Briefly, cells were cultured in RPMI 1640 medium supplemented with l-glutamine and fetal calf serum (10%, vol/vol) before being harvested at the log phase of confluence and resuspended in RPMI 1640 medium. Samples (2 × 104 cells in 100 μl) were seeded into culture plates and allowed to grow overnight at 37°C under 5% (vol/vol) CO2. Fluconazole analogues were added at a variety of concentrations, and the cells were incubated under the same conditions for 24 h. Equivalent amounts of solvent, amphotericin B, and fluconazole were used as negative and positive controls. The medium was removed from each well before the addition of 50.0 μg of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in PBS (100 μl). After incubation for a further period of 4 h at 37°C, the MTT solution was removed and the cells were lysed using isopropanol-HCl (100.0 μl). The absorption of each well (at 540 nm) was used to determine the percent cytotoxicity in a microplate reader (SpectraMax 384 Plus; Molecular Devices, USA).
ACKNOWLEDGMENTS
The Council of Scientific and Industrial Research (CSIR, New Delhi), the University of Delhi, and the University Grants Commission (UGC, New Delhi) provided research facilities and funding for this study. S.M. acknowledges financial support from CSIR, New Delhi, through the award of a Senior Research Associateship.
FOOTNOTES
- Received 10 March 2017.
- Returned for modification 18 April 2017.
- Accepted 19 May 2017.
- Accepted manuscript posted online 12 June 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.00273-17 .
- Copyright © 2017 American Society for Microbiology.
