Medical Mycology December 2004, 42, 525 /529

Biochemical characterization of terbinane-resistant Trichophyton rubrum isolates

BERTRAND FAVRE*%, MAHMOUD A. GHANNOUM$ & NEIL S. RYDER* *Novartis Research Institute, Vienna, Austria and $Center for Medical Mycology, Department of Dermatology, Case Western Reserve University and University Hospitals of Cleveland, Cleveland, OH, USA

Med Mycol Downloaded from informahealthcare.com by HINARI on 05/07/13 For personal use only.

We investigated the biochemical basis for resistance in six sequential clinical isolates of Trichophyton rubrum , from the same patient, which exhibited high-level primary resistance to terbinafine. Cellular ergosterol biosynthesis was measured by incorporation of [14C]acetate, and microsomal squalene epoxidase was assayed by conversion of [3H]squalene to squalene epoxide and lanosterol. Direct comparison was made with a terbinafine-susceptible reference strain of T. rubrum in which squalene epoxidase was previously studied. Resistant isolates displayed normal cellular ergosterol biosynthesis, although slight accumulation of radiolabeled squalene suggested reduced squalene epoxidase activity. Ergosterol biosynthesis in the resistant isolates was only inhibited by terbinafine concentrations above 1 mg/ ml (IC50 5 mg/ml). In the reference strain, ergosterol biosynthesis was eliminated by terbinafine at 0.03 mg/ml in accordance with historical data. There was no significant difference in sensitivity between the six resistant isolates. Squalene epoxidase from resistant strains was three orders of magnitude less sensitive than normal enzyme to terbinafine (IC50 of 30 mmol/l and 19 n mol/l respectively). The epoxidase in the resistant strains was also unresponsive to tolnaftate. Resistance to terbinafine in these T. rubrum isolates appears to be due to alterations in the squalene epoxidase gene or a factor essential for its activity. Keywords dermatophyte, resistance, Terbinane, Trichophyton

Introduction
The allylamine terbinafine and the related compounds naftifine and butenafine are selective inhibitors of fungal squalene epoxidase (SE) [1]. This property is shared by the thiocarbamates tolnaftate and tolciclate [2], which are chemically distinct from the allylamines. Accumulation of squalene appears to be toxic to filamentous fungi, especially in dermatophytes [3]. Despite the extensive use of terbinafine for the treat-

Received 8 September 2003; Accepted 16 December 2003 Correspondence: Neil S. Ryder, Infectious Diseases Biology, Novartis Institutes for BioMedical Research Inc., 100 Technology Square, Cambridge, MA 02139, USA. Tel.: '/1 617 871 3143; Fax: '/1 617 871 7047; E-mail: neil.ryder@pharma.novartis.com % Present address: Laboratory of Dermatology, University Hospital CHUV, Lausanne, Switzerland.

ment of dermatophytosis and onychomycosis, the first clinically confirmed case of terbinafine resistance in dermatophytes was only recently reported [4]. Six sequential clinical isolates of Trichophyton rubrum originating from nail of a single patient who failed on therapy with oral terbinafine, were found to be resistant to terbinafine, with MIC ]/4 versus 5/0.001 mg/ml for normal strains. They were cross-resistant to the other SE inhibitors tested, butenafine, naftifine, tolnaftate and tolciclate, but normally susceptible to antifungals with a different mode of action, such as azoles and griseofulvin [4], as well as amorolfine and amphotericin B (N. S. Ryder, unpublished data, 2000). These results suggested a target-specific mechanism of resistance. The aim of the present study was to determine at the biochemical level whether resistance was indeed linked to abnormalities in the ergosterol biosynthesis pathway and SE enzyme.
DOI: 10.1080/13693780410001661482

2004 ISHAM

526

Favre et al.

Materials and methods

Fungal isolates
Six consecutive clinical isolates of T. rubrum designated (Novartis Fungal Index) NFI5146 to NFI5151, including the baseline isolate NFI5146, were derived from nail material of a single patient with toenail onychomycosis sampled at intervals. Details of the clinical origin, culturing and mycological testing of these isolates have been provided previously [4]. For comparison, the reference strain T. rubrum NFI1895 was tested in parallel as this strain has been studied extensively in terms of cellular ergosterol biosynthesis and SE inhibition by terbinafine [5].

Antifungal drugs
Med Mycol Downloaded from informahealthcare.com by HINARI on 05/07/13 For personal use only.

Terbinafine (batch 97045) was synthesized at Novartis (Basel, Switzerland). Tolnaftate was purchased from Sigma (St Louis, MO, USA, catalogue T-6638) and amorolfine was obtained from Hoffmann La Roche (Basel, Switzerland).

Cellular ergosterol biosynthesis

The method for measurement of ergosterol biosynthesis inhibition was modified from a previously described procedure [6]. Inocula were prepared from stocks frozen at /808C. Cultures were grown in 500ml conical flasks containing 150 ml Sabouraud 2% dextrose broth pH 6.5, inoculated at 3 )/104 to 5 )/105 colony forming units (c.f.u.) per ml and incubated on a rotary shaker (LH Engineering, Stoke Poges, UK) at 100 r.p.m. and 308C for 4 to 6 days until adequate fungal mycelium was obtained. Because of the slow growth rate of dermatophytes in culture, the organisms were still in the growth phase when harvested. Cultures were harvested by gravity filtration on a glass sinter (porosity 1), washed in K-Na-PO4 buffer, pH 6.5, and resuspended in incubation medium. Incubation medium consisted of yeast nitrogen base medium without (NH4)2SO4 and amino acids, containing 2% glucose (w/v) and 25 mmol/l K-Na-PO4 buffer pH 6.5. Cell suspensions were adjusted to a density equivalent to 3 /4 mg dry weight per ml. Incubations were performed in 50-ml conical glass flasks in a shaking water bath at 308C. Inhibitors were dissolved in DMSO and added at 100-fold final concentration. Controls received solvent only. Cell suspension (5 ml) was pre-incubated 10 min with the test compound and the assay initiated by addition of 25 ml substrate mixture prepared from [U-14C]-acetate (Amersham Biosciences, Vienna, Austria, 1 mCi/5 ml,

50 /62 mCi/mmol), 1 mol/l Na-acetate and distilled water mixed in the ratio 2:1:7. After an incubation time of 2 h, assays were terminated and cells harvested by filtration on Whatman (Maidstone, UK) 2.5 cm GF/A glass fiber filter discs. The cell mats were transferred to glass tubes, to which was added 1.5 ml 30% (w/v) KOH in 90% (v/v) ethanol, 1.5 ml ethanol and 0.3 ml 2% (w/v) 1,2,3-benzenetriol (pyrogallol) in ethanol. Tubes were closed with screw caps and left overnight in the dark at room temperature, then incubated 30 min at 808C. After cooling, 1.5 ml water was added and the non-saponifiable lipids were extracted twice with 2 ml petroleum ether (bp 40 /608C). The extract was washed with 3 ml distilled water, evaporated to dryness and dissolved in cyclohexane. The non-saponifiable lipids were then separated by thin layer chromatography as described previously [6]. The TLC sheets were analyzed by electronic autoradiography using the Instant Imager (Canberra Packard, Vienna, Austria, with software Packard Image for Windows 2.10). The Instant Imager provided both an autoradiographic image of the TLC sheet and a quantitative determination of the distribution of radioactivity. The non-specific radioactivity which accumulated at the origin of the plates was not included in the calculations. After automatic background subtraction, the radioactivity was calculated for each band (squalene, sterols etc.) as a percentage of the sum of the designated products. These data were transferred from the Imager results file to a template in Excel 97 for calculation of percent inhibition of ergosterol biosynthesis at each drug concentration and the IC50 value. Each experiment was performed with duplicate incubations, and repeated so that each final value was derived from four separate determinations.

Squalene epoxidase assay

Trichophyton rubrum microsomes (170 000 g pellet) were prepared from mycelium frozen in liquid nitrogen [5]. Microsomal SE activity was assayed by measurement of incorporation of [4,8,12,13,17,21-3H]squalene (American Radiochemicals, St Louis, MO, USA) into squalene epoxide and lanosterol exactly as previously described [5], except for the final protein concentration of 3 mg protein/ml (instead of 2 mg/ml) and the incubation time of 60 min (instead of 45 min). Inhibitors were first dissolved in DMSO and then diluted 100-fold into the assay to achieve the desired final concentration. Experiments were performed in duplicate and repeated with each inhibitor twice. IC50 values were calculated with the program Origin 6.1
2004 ISHAM, Medical Mycology, 42, 525 /529

Terbinafine-resistant T. rubrum

527

(OriginLab Corporation, Northampton, MA, USA) using the curve fitting function.

Table 2 Sensitivity of cellular ergosterol biosynthesis to terbinane in reference and resistant strains Strain Visita / 0 2 3 4 7 8 Type Reference Resistant Resistant Resistant Resistant Resistant Resistant Terbinafine IC50 (mg/ml)b 0.0049/0.002 5.59/0.9 4.19/0.8 5.09/0.4 4.29/1.2 5.09/1.2 5.49/2.5

Results
Reduced sensitivity of cellular ergosterol biosynthesis
Analysis of the labeled non-saponifiable lipids in the absence of drug revealed that in the resistant isolate NFI5150 around 25% of the radioactivity was associated with squalene (Table 1). Similar results were observed with the other five resistant isolates (not shown). By contrast, in the reference strain NFI1895 no significant radioactivity could be detected comigrating with squalene or other ergosterol precursors (Table 1), in agreement with earlier studies. This difference indicates that the flux of sterol precursors from [14C]acetate to ergosterol was not identical in the susceptible strain NFI1895 and the resistant isolates, and suggests the possibility of lower SE activity in the latter. With the reference strain NFI1895, a terbinafine concentration of 0.004 mg/ml was sufficient to inhibit cellular ergosterol biosynthesis by 50%, while in the resistant isolates the IC50 was about 1000-fold higher (Table 2), confirming that inhibition of ergosterol biosynthesis underlies the antifungal activity of terbinafine in both susceptible and resistant isolates.

NFI1895 NFI5146 NFI5147 NFI5148 NFI5149 NFI5150 NFI5151

a

Clinical history: visit 0 was for screening and culture; therapy with terbinafine (250 mg once daily) lasting 24 weeks started at visit 1, and subsequent visits were at 6-week intervals [4]. bMean9/SD for four separate determinations.

Med Mycol Downloaded from informahealthcare.com by HINARI on 05/07/13 For personal use only.

the resistant isolates, the technically demanding studies with the SE enzyme were performed in only two of the resistant isolates. The specific activities of SE measured under the standard assay conditions were similar in the three isolates: 0.049/0.01 nmol/h/mg protein (9/SD, n 0/5) for reference strain NFI1895, 0.049/0.01 nmol/h/ mg protein (9/SD, n 0/4) for NFI5146, and 0.059/0.01 nmol/h/mg protein (9/SD, n 0/4) for NFI5150. However, a dramatic difference was found between the normal and resistant strains with respect to sensitivity
Table 3 Sensitivity of microsomal SE activity from reference and resistant strains to terbinane, tolnaftate and amorolneStrain IC50 (mg/ml)a Strain NFI1895 NFI5146 NFI5150 Terbinafine 0.006 11 9 Tolnaftate 0.05 /33 /33 Amorolfine 9 19 15

Low sensitivity of squalene epoxidase

A simple way of distinguishing target-based mechanisms of resistance from the involvement of membrane transporters is by testing the sensitivity of the target to the inhibitor in cell-free conditions, ruling out the action of transporters [7]. As the cellular studies revealed no apparent differences in sensitivity between

a Mean of IC50 values from two separate determinations, each of which did not vary by more than 30% from the mean value.

Table 1 Comparison of radiolabeling patterns of ergosterol and intermediates in cells of Trichophyton rubrum NFI1895 (reference strain) and NFI5150 (resistant isolate) incubated with [14C]acetate and increasing concentrations of terbinaneaStrain Percent total counts measured in: (g/ ml) Strain NFI1895 NFI1895 NFI1895 NFI1895 NFI1895 NFI5150 NFI5150 NFI5150 NFI5150 NFI5150 Terbinafine 0 0.001 0.003 0.01 0.03 0 0.01 0.1 1 10 Ergosterol 929/1 909/2 479/28 139/8 59/2 679/3 679/3 649/5 589/6 129/2 4a-methyl-sterol 29/0.5 29/0.4 29/1 19/0.3 19/0.2 29/0.4 39/0.4 29/0.4 39/0.4 19/0.3 Lanosterol 49/1 59/1 49/2 29/0.2 19/0.2 59/2 69/2 59/2 69/2 59/3 Squalene 29/0.3 39/1 479/31 859/8 949/3 259/3 249/4 279/5 339/6 819/2

a Results are given as mean9/standard deviation for four data points (two independent experiments each performed with duplicate incubations). Radiolabeling and analysis of lipids were performed as described in the Methods section.

2004 ISHAM, Medical Mycology, 42, 525 /529

528

Favre et al.

of SE to terbinafine. Microsomal SE activity from resistant isolates NFI5146 and NFI5150 was 1000-fold less sensitive to terbinafine than that from the reference strain NFI1895 (Table 3). A similar difference was observed with tolnaftate, in agreement with the crossresistance of the isolates to thiocarbamates [4]. On the other hand, amorolfine, which also weakly inhibits T. rubrum SE [5], showed roughly similar activity against microsomal SE from reference and resistant strains (Table 3).

Discussion
At present, very little is known concerning drugresistance mechanisms in dermatophytes. The recent discovery of the first terbinafine-resistant dermatophyte associated with clinical failure [4] raised the important issue of the molecular mechanism of resistance to this widely used drug. Several mechanisms of resistance to another class of ergosterol biosynthesis inhibitors, the azoles, have been established in Candida species [8]. These include modification of the target enzyme, upregulation of the ergosterol biosynthesis pathway and, (most commonly) induction of drug efflux pumps leading to resistance to multiple drugs, including terbinafine in some cases [9]. We previously showed that the terbinafine-resistant T. rubrum isolates were fully cross-resistant to several classes of SE inhibitors, but displayed normal susceptibility to azoles and other antifungals [4]. This suggested that resistance was associated with changes in the target enzyme rather than with overexpression of efflux pumps. The data from the present study support this hypothesis. Both cellular ergosterol biosynthesis and cell-free SE activity from resistant isolates displayed reduced sensitivity to terbinafine by several orders of magnitude. The resistant cells possessed a functioning ergosterol biosynthesis pathway, but with an unusually high level of radiolabeled squalene in the control cells, which would be consistent with a reduced activity of squalene SE although alternative explanations are possible. Direct enzyme assays confirmed the existence of a functional SE in the resistant isolates and its abnormally low sensitivity to terbinafine. The specific activity of the resistant microsomal enzyme was similar to that from reference strain NFI1895, thus contradicting the evidence for reduced SE activity in intact cells. However, in vitro enzyme assays containing excess concentrations of cofactors and substrates may not detect differences in activity within living cells where these factors may be limiting. Interestingly, the resistant SE was also unresponsive to tolnaftate suggesting that allylamines and thiocarbamates may share a common binding site. In

contrast, the sensitivity of the terbinafine-resistant SE to amorolfine was not very different from that of a normal enzyme suggesting that amorolfine has a mechanism of inhibition distinct from that of terbinafine and tolnaftate. Amorolfine is a weak inhibitor of T. rubrum SE [5], while the principal antifungal targets of the drug are sterol-D14-reductase and sterol-D7-D8isomerase [10]. In the original report of the resistant isolates [4], there was an apparent further decrease in susceptibility (detectable only by the broth macrodilution assay) over the course of terbinafine treatment. Since the sensitivity of ergosterol biosynthesis remained essentially unchanged over the sequence of isolates (see Table 2), this subtle change must have been due to a different, as yet unidentified, mechanism. The six resistant isolates, including one taken at baseline before start of therapy, were obtained sequentially from the same patient and displayed identical random amplified polymorphic DNA (RAPD) profiles [4], indicating that they represent a single genotypic strain. Several lines of evidence point to a stable genetic alteration, probably in the SE gene, as the primary cause of terbinafine resistance in this case. These include (i) the drastic reduction in enzyme sensitivity, (ii) the rare occurrence of resistance, and (iii) the stable nature of the resistance in cultures taken over 24 weeks of therapy and after subculturing in the absence of the drug. Previous studies have shown that clinical failure of terbinafine was not associated with development of in-vitro resistance during therapy [11]. However, it appears that strains expressing primary resistance to the drug occur at a low frequency in the T. rubrum population. Occasional instances of high terbinafine MIC values have also been reported, but without clinical information [12]. Fungal SE genes have been cloned and sequenced from the yeasts Saccharomyces cerevisiae [13] and Candida albicans [14]. The amino acid substitution Leu 0/Phe251, apparently associated with a terbinafine-resistant phenotype, was identified in the S. cerevisiae gene [14], and a second mutation leading to the substitution Pro 0/Ala430 was recently reported to confer partial loss of sensitivity to inhibition by terbinafine [15]. Investigations are in progress to determine whether amino acid substitution(s) are indeed fully responsible for terbinafine resistance in T. rubrum , and whether other mechanisms of resistance can also be detected.

Med Mycol Downloaded from informahealthcare.com by HINARI on 05/07/13 For personal use only.

Acknowledgements
We thank Wolfgang Mlineritsch for performing the cellular sterol biosynthesis assay.
2004 ISHAM, Medical Mycology, 42, 525 /529

Terbinafine-resistant T. rubrum

529

References
1 Ryder NS. Terbinane: mode of action and properties of the squalene epoxidase inhibition. Br J Dermatol 1992; 126(Suppl. 39): 2 /7. 2 Ryder NS, Frank I, Dupont MC. Ergosterol biosynthesis inhibition by the thiocarbamate antifungal agents tolnaftate and tolciclate. Antimicrob Agents Chemother 1986; 29: 858 /860. 3 Ryder NS, Mieth H. Allylamine antifungal drugs. Curr Top Med Mycol 1992; 4: 158 /188. 4 Mukherjee PK, Leidich SD, Isham I, Leitner I, Ryder NS, Ghannoum MA. Clinical Trichophyton rubrum strain exhibiting primary resistance to terbinane. Antimicrob Agents Chemother 2003; 47: 82 /86. 5 Favre B, Ryder NS. Characterization of squalene epoxidase activity from the dermatophyte Trichophyton rubrum and its inhibition by terbinane and other antimycotic agents. Antimicrob Agents Chemother 1996; 40: 443 /447. 6 Ryder NS. Specic inhibition of fungal sterol biosynthesis by SF 86-327, a new allylamine antimycotic agent. Antimicrob Agents Chemother 1985; 27: 252 /256. 7 Favre B, Didmon M, Ryder NS. Multiple amino acid substitutions in lanosterol 14a-demethylase contribute to azole resistance in Candida albicans . Microbiology 1999; 145: 2715 /2725.

Med Mycol Downloaded from informahealthcare.com by HINARI on 05/07/13 For personal use only.

8 Balkis MM, Leidich SD, Mukherjee PK, Ghannoum MA. Mechanisms of fungal resistance: an overview. Drugs 2002; 62: 1025 /1040. 9 Sanglard D, Ischer F, Monod M, Bille J. Cloning of Candida albicans genes conferring resistance to azole antifungal agents:characterization of CDR2, a new multidrug ABC transporter gene. Microbiology 1997; 143: 405 /416. 10 Polak A. Preclinical data and mode of action of amorolne. Dermatology 1992; 184(Suppl. 1): 3 /7. 11 Bradley MC, Leidich S, Isham N, Elewski BE, Ghannoum MA. Antifungal susceptibilities and genetic relatedness of serial Trichophyton rubrum isolates from patients with onychomycosis of the toenail. Mycoses 1999; 42(Suppl. 2): 105 /110. 12 Ferna ndez-Torres B, Carrillo AJ, Mart n E, et al . In vitro activities of 10 antifungal drugs against 508 dermatophyte strains. Antimicrob Agents Chemother 2001; 45: 2524 /2528. 13 Jandrositz A, Turnowsky F, Hogenauer G. The gene encoding squalene epoxidase from Saccharomyces cerevisiae : cloning and characterization. Gene 1991; 107: 155 /160. 14 Favre B, Ryder NS. Cloning and expression of squalene epoxidase from the pathogenic yeast Candida albicans . Gene 1997; 189: 119 / 126. 15 Klobucnikova V, Kohut P, Leber R, et al . Terbinane resistance in a pleiotropic yeast mutant is caused by a single point mutation in the ERG1 gene. Biochem Biophys Res Comm 2003; 309: 666 /671.

2004 ISHAM, Medical Mycology, 42, 525 /529