Steroid Interference with Antifungal Activity of Polyene Antibiotics

Home , Lanosterol, Stigmasterol

WALTER A. ZYGMUNT AND PETER A. TAVORMINA Department of Microbiology and Natural Products Research, Mead Johnson & Company, Evansville, Indiana Received for publication 21 April 1966 ABSTRACT ZYGMUNT, WALTER A. (Mead Johnson & Co., Evansville, Ind.), AND PETER A. TAVORMINA. Steroid interference with antifungal activity of polyene antibiotics. Appl. Microbiol. 14:865-869. 1966.-Wide differences exist among the polyene antibiotics, nystatin, rimocidin, filipin, pimaricin, and amphotericin B, with reference to steroid interference with their antifungal activities against Candida albicans. Of the numerous steroids tested, ergosterol was the only one which effectively antagonized the antifungal activity of all five polyene antibiotics. The antifungal activities of nystatin and amphotericin B were the least subject to vitiation by the addition of steroids other than ergosterol, and those of filipin, rimocidin, and pimaricin were the most sensitive to interference. Attempts to delineate the structural requirements of steroids possessing polyene-neutralizing activity in growing cultures of C. albicans are discussed. The ultraviolet absorbance of certain antibiotic steroid combinations was also studied.

It has been suggested (1, 9, 13) that the polyene While studying the effects of various steroids antibiotics become bound to the fungal cell mem- on the antimonilial activity of pimaricin, we brane and cause permeability changes with observed that ergostenol was almost as effective attendant depletion of essential cellular con- as the above A5-3/3-hydroxy steroids in antag- stituents. Loss of potassium and ammonium onizing pimaricin. In contrast, certain other ions, inorganic phosphate, carboxylic acids, A5-33-hydroxy steroids showed no neutralizing sugar phosphates, nucleotides, and protein have activity. In view of the sparsity of information been shown to occur in several fungal species on the reversal of the biological activity of polyene after exposure to nystatin, filipin, and ampho- antibiotics by certain steroids against Candida tericin B (1, 5, 9, 10, 13, 14, 16). Furthermore, it albicans, a dimorphic fungus, it seemed pertinent has been suggested that the binding site on the to undertake further studies with this organism membrane contains a sterol, possibly unesterified with use of several classes of polyene antibiotics ergosterol (13). The presence of a sterol appears and a wide variety of steroids as potential to be a prerequisite for sensitivity toward the neutralizing agents. polyene antibiotics. Thus far, growth inhibitory activities of the polyene antibiotics have been MATERIALS AND MerHoDs demonstrated against yeasts (6, 14), algae (6, Growth conditions. All studies were done in a 12), protozoa (2), and flatworms (8). In contrast, liquid medium (Sabouraud's dextrose broth) with bacteria which contain no significant amount of volumes of 10 ml in a 50-ml Erlenmeyer flask. Unless sterols are insensitive to these antibiotics. otherwise indicated, the medium had an initial pH of Hickey (6) reported that certain fatty acids 4.5, and all flasks were cultured under constant agita- depress the antifungal activity of ascosin. Gott- tion on a rotary shaker at 28 C. C. albicans (ATCC lieb et al. (3) found that cholesterol, ergosterol, 10231) was maintained on YM (Difco) slants and was transferred at intervals. The inoculum sitosterol, and stigmasterol had a neutralizing weekly used was derived from a suspension of an 18- to 24-hr effect on the of antifungal activity filipin against slant culture in sterile saline. The cells were centri- Hansenula subpelliculosa. The above A5-3,B- fuged, washed twice, resuspended in saline, and ad- hydroxy steroids also antagonized the inhibitory justed to an optical density reading of 0.100 0.010 activity of pimaricin against Saccharomyces with a Coleman Junior spectrophotometer (620 m,p). cerevisiae (15). Two drops ofthis inoculum were used per flask. 865 866 ZYGMUNT AND TAVORMINA APPL. MICROBIOL.

Polyene antibiotics. Nystatin and amphotericin B RESULTS were kindly furnished by E. R. Squibb & Sons, New York, N.Y.; rimocidin, by Chas. Pfizer & Co., Inc., Data in Table 1 clearly differentiate the ability Brooklyn, N.Y.; pimaricin, by American Cyanamid of various compounds to antagonize the anti- Co., Pearl River, N.Y.; and filipin, by The Upjohn fungal activities of the polyene antibiotics tested. Co., Kalamazoo, Mich. For example, ergosterol is the only compound of Stock solutions (100 pg/ml) of each antibiotic were the 27 tested which interfered with the inhibitory prepared aseptically by dissolving 5 mg of antibiotic in 0.5 ml of dimethyl sulfoxide, diluting promptly activity of nystatin. In contrast, cholesterol, with sterile distilled water, and adjusting the pH to cholesterol acetate, dihydrocholesterol, 5a-cho- 4.5. lestan-3-one, lanosterol, ergosterol, ergostenol, Potential neutralizing agents. Cholesterol, choles- stigmasterol, and ,B-sitosterol are all highly terol acetate, cholesterol propionate, ergosterol, es- effective in neutralizing filipin. Our results tradiol, estrone, deoxycorticosterone, pregnenolone, precisely parallel those of Gottlieb et al. (4), dehydroepiandrosterone, jS-sitosterol, dihydrocholes- with the exception that diosgenin was inactive in terol, stigmasterol, testosterone, progesterone, cholic our hands. Epiandrosterone was not tested. acid, deoxycholic acid, squalene, lanosterol, AW-cho- Interestingly, although cholesterol acetate showed lenic acid-30-ol, A5-pregnen-3,B-ol-20-one-3-methyl with ether, diosgenin, smilagenin, calciferol, 5a-cholestan- reversal activity filipin, the propionate (Table 1) and palnitate (4) derivatives were 3-one, oleic acid, and sodium acetate were tested as possible antagonists of the polyene antibiotics. All inactive. Amphotericin B approached nystatin in compounds were obtained commercially and were of the specificity of chemical structure required for analytical reagent grade. Unless otherwise indicated, vitiation of anticandidal activity in that only addition of these compounds was at 100 pg/ml. ergosterol and j3-sitosterol are active compounds. Ergostenol was prepared by hydrogenation of With the exception of dihydrocholesterol inter- ergosterol (PtO2 in absolute ethyl alcohol). Nuclear fering with the biological activity of rimocidin, magnetic resonance analysis of the product showed it the spectra of compounds antagonizing rimocidin of to be a mixture A8:14 (50%), A8:9 (35%), and and were identical. A7:8 (15%) ergosten-3-ol. pimaricin Addition of the five antibiotics at a twofold All steroids were emulsified with the aid of poly- oxyethylene sorbitan monooleate (Tween 80) and increase in concentration results in only minor propylene glycol, sterilized by free-flowing steam for changes from the steroid-neutralizing activities 10 min and diluted aseptically with distilled water to described above. f3-Sitosterol (100 ,g/ml) failed a concentration of 1 mg/ml. Oleic acid was dissolved to antagonize the growth inhibition of 10 ,ug/ in 95% ethyl alcohol, sterilized by Seitz filtration, and ml of amphotericin B, whereas it was fully active diluted aseptically with distilled water. All other com- at a lower antibiotic level. pounds were dissolved in distilled water, neutralized, The following compounds antagonized the and sterilized by filtration. Only freshly prepared anticandidal activities of all five antibiotics (5 solutions of antibiotics and neutralizing agents were used throughout these studies. and 10 ,ug/ml) by less than 20% and in most The Cary recording spectrophotometer, model 14, instances were totally inactive: calciferol, preg- was used for ultraviolet spectral data. nenolone, dehydroepiandrosterone, testosterone,

TABLE 1. Vitiation of the antifungal activities ofpolyene antibiotics by certain steroidsa Neutralizaton of growth inhibition (per cent)C Steroidb Nystatin Filipin Amphotericin B Rimocidin Pimaricn Cholesterol ...... 100 - 100 90 Cholesterol acetate.. 94 - - Dihydrocholesterol.- 100 _ 99

5a-Cholestan-3-one ...... 97 Ergosterol ...... 88 100 98 100 100 Ergostenol ...... 95 88 74

Stigmasterol.- ...... 96 22 84 92 B-Sitosterol ...... 96 70 99 100 Lanosterol ...... _..76 - - - a Addition of all antibiotics was 5 pg/ml, and of steroids, 100 pg/ml. b In the absence of steroids, all antibiotics resulted in complete inhibition of yeast growth. c A total of 27 compounds were tested individually with each of the antibiotics, but only those compounds in which the growth inhibition was reduced by at least 20% are indicated. VOL. 14,1966 STEROID INTERFERENCE WITH ANTIFUNGAL ACTIVITY 867

TABLE 2. Effects ofergosterol and cholesterol additions on the antifungal activities ofpolyene antibioticsa

Neutralization of growth inhibition (per cent) Addition Amt Rimocidin Filipin Pimaricin Nystatin Amphotericin B pg/mi Ergosterol ...... 5 0 0 0 0 0 10 1 1 1 1 1 25 95 22 22 38 98 50 100 100 100 92 98 100 97 97 97 90 97

Cholesterol ..... 5 0 100 0 0 0 10 94 100 0 0 0 25 100 100 77 0 0 50 100 100 100 0 3 100 100 100 78 0 7 a Addition of all antibiotics was at 10 Ag/ml, and, in each instance, caused complete inhibition of yeast growth. TABLE 3. Effect ofpH on the neutralization of the growth media with initial pH in the broad range antifungal activities of nystatin and from 4.5 to 7.5 (Table 3). amphotericin B by ergosterol Lampen, Arnow, and Safferman (11) explained the sterol interference in the antifungal activity Neutralization of growth inhibition (per cent)a of certain polyene antibiotics essentially on a Ergos- B physicochemical basis. With both filipin and terol Nystatin Amphotericin nystatin, addition of ergosterol or cholesterol in Ab B C D A B C D a cell-free system resulted in spectral shifts and a decrease in absorbance. They also found that the antifungal activity of nystatin was markedly less 5 0 1 1 9 00 0 0 affected by cholesterol than was that of anti- 25 87 92 96 95 100 97 91 90 mycoin or filipin in S. cerevisiae. No data were 100 88 100 94 100 100 96 100 100 presented, however, on the comparable neutraliz- a Nystatin and amphotericin B were added at 10 ing effects of ergosterol with nystatin with use of Ag/ml and resulted in complete growth inhibition. viable yeast cells. Their data would suggest, how- b A = Sabouraud's dextrose broth (SDB), pH ever, somewhat equivalent activity for both 4.5; B = SDB,pH 5.5; C = SDB,pH6.5; andD = steroids. SDB, pH 7.5. Our results with C. albicans indicated marked differences in the ability of ergosterol and progesterone, cholic acid, deoxycholic acid, cholesterol to negate the antifungal activities of squalene, estradiol, estrone, deoxycorticosterone, either nystatin or amphotericin B (Table 2). cholesterol propionate, A5-cholenic acid, A5- When the absorption spectra ofnystatin, nystatin- pregnen-3(3-ol-20-one-3-methyl ether, diosgenin, cholesterol, and nystatin-ergosterol solutions were smilagenin, oleic acid, and sodium acetate. compared, either steroid effectively decreased the Because both cholesterol and ergosterol absorbance of nystatin, although ergosterol was interfered with the activities of rimocidin, filipin, the more effective in this regard (Table 4). These and pimaricin, but only ergosterol antagonized findings confirmed the results of Lampen et al. nystatin and amphotericin B, it seemed pertinent (11), but offer no explanation for the apparent to explore this facet by use of graded levels of the inactivity of cholesterol in antagonizing the anti- two steroids. The results are summarized in Table fungal activity with growing yeast cells. 2, and show cholesterol to be a more effective With reference to the findings of Perritt, antagonist of rimocidin, filipin, and pimaricin Phillips, and Robinson (15) that A5-3g-hydroxy than ergosterol. Both steroids at low levels steroids antagonize the antifungal activity of antagonized the antifungal activities of rimocidin pimaricin, we found that pregnenolone and and fihipin more effectively than of pimaricin. dehydroepiandrosterone (both A5-3,3-hydroxy Ergosterol neutralized the biological activity of steroids) showed no neutralizing effect with any nystatin and amphotericin B equally well in of the five polyene antibiotics. Nor was the 868 ZYGMUNT AND TAVORMINA APPL. MICROBIOL. TABLE 4. Effect of various steroids on the ultra- effectively as ergosterol but, like cholesterol and violet spectra of nystatina dihydrocholesterol, is relatively inactive in antagonizing the inhibition of nystatin against Nystatin (10 pAg/ml) C. albicans, is difficult to reconcile on the basis of Steroid added (100 pg/ml) a solely physical phenomenon (Table 4). Addi- Ultraviolet Absorbance maxima tions of stigmasterol, f3-sitosterol, and dehydroepiandrosterone (also all 33-hydroxy steroids) None ...... 321,306 0.644,0 .779 resulted in only minor changes in ultraviolet Cholesterol ...... 321, 306 0.560, 0.729 absorbance of nystatin. Dihydrocholesterol. .. 321, 306 0.578, 0.719 In view of the proposal that the binding site Ergosterol ...... 323, 309 0.416, 0.610 for polyenes on the fungal membrane may be Ergostenol ...... 322, 308 0.495, 0.670 unesterified ergosterol, the binding of polyenes Stigmasterol ...... 321,306 0.620,0.794 with ergosterol may be more rigid and less ,B-Sitosterol ...... 321, 306 0.632, 0.842 susceptible to equilibrium changes than that en- Dehydroepiandros- countered with certain other steroids. The finding terone ...... 321,306 0.710,0.850 that ergosterol is the only steroid which antago- a Ultraviolet absorption spectra of nystatin and nizes the antifungal activity of all five polyene nystatin-steroid mixtures in Sabouraud's dextrose antibiotics is consistent with this hypothesis. broth (pH 4.5) after incubation at room tempera- ture for 2 hr. ACKNOWLEDGMENTS The technical assistance of E. Townsend is appreci- ated. We are grateful to R. Covington and W. Comer absorbance of pimaricin decreased significantly for the preparation and analysis of the ergostenol, by additions of pregnenolone. As expected, and to J. Schmidt for the ultraviolet spectra analyses. additions of cholesterol significantly decreased the absorbance of pimaricin. LITERATURE CITED Ergostenol is an example of a steroid which is 1. BRADLEY, S. G., AND L. A. JoNEs. 1960. Mecha- not a A5-3#-hydroxy steroid, but which never- nisms of action of antibiotics. Ann. N.Y. Acad. theless showed neutralizing activity with pimari- Sci. 89:122-133. cin, rimocidin, and filipin. Interestingly, dihydro- 2. GHOSH, B. K., AND A. N. CHATTERJEE. 1962. cholesterol (no A5-unsaturation) antagonized the Leishmanicidal activity of nystatin, a polyene inhibition of rimocidin and fihipin, but not of the antifugal antibiotic. I. The probable mecha- other polyenes. Lanosterol neutralized the anti- nism of action of nystatin on Leismania dono- candidal activity of filipin only. vani. Antibiot. Chemotherapy 12:204-206. 3. GoTrLEIB, D., H. E. CARTER, J. H. SLONEKER, The steroids which neutralized the growth AND A. AMmANN. 1958. Protection of fungi inhibition of two or more polyene antibiotics against polyene antibiotics by sterols. Science were ergosterol, ergostenol, cholesterol, di- 128:361. hydrocholesterol, f8-sitosterol, and stigmasterol. 4. GOTrLEIB, D., H. E. CARTER, L. C. Wu, AND J. These sterols have the following structural H. SLONEKER. 1960. Inhibition of fungi by features in common: a 3f3-hydroxy group on ring filipin and its antagonism by sterols. Phyto- A and an 8- to 10-carbon side chain on ring D. pathology 50:594-603. The nature of the side chain appeared to play a 5. GoTmLEB, D., H. E. CARTER, J. H. SLONEKER, role in the activity of the steroids in that such L. C. Wu, AND E. GAUDY. 1961. Mechanisms of inhibition of fungi by filipin. Phytopathology 33-hydroxy steroids as dehydroepiandrosterone, 51:321-330. pregnenolone, estrone, estradiol, and such sapo- 6. HIcKEY, R. J. 1953. The antagonism between the genins as diosgenin and smilagenin, were all antifungal antibiotic, ascosin, and some long- inactive. chain, unsaturated fatty acids. Arch. Biochem. Although the data in these studies are in general Biophys. 46:331-336. agreement with the physicochemical hypothesis 7. HUNTER, E. O., AND I. MCVEIGH. 1961. The of sterol-polyene antibiotic binding proposed by effects of selected antibiotics on pure cultures of Lampen et al. (11), the relatively poor neutraliz- algae. Am. J. Botany 48:179-185. ing activity of cholesterol with nystatin when 8. JOHNSON, W. H., C. A. MILLER, AND J. E. BRUM- BAUGH. 1962. Induced loss of pigment in pla- compared with that of ergosterol can be inter- narians. Physiol. Zool. 35:18-26. preted in part by our findings that cholesterol 9. KINSKY, S. C. 1961. Alterations in the permea- appears to be only one-third as effective as bility of Neurospora crassa due to polyene anti- ergosterol in decreasing the absorbance of biotics. J. Bacteriol. 82:889-897. nystatin. The observation that ergostenol, how- 10. KINSKY, S. C. 1962. Nystatin binding by proto- ever, can bind nystatin about two-thirds as plasts and a particulate fraction of Neurospora VOL.14,1966 STEROID INTERFERENCE WITH ANTIFUNGAL ACTIVITY 869

crassa, and a basis for the selective toxicity of 14. MARINI, F., P. ARNOW, AND J. 0. LAMPEN. 1961. polyene antifungal antibiotics. Proc. Natl. Acad. The effect of monovalent cations on the in- Sci. U.S. 48:1049-1056. hibition of yeast metabolism by nystatin. J. 11. LAMPEN, J. 0., P. M. ARNOW, AND R. S. SAFFER- Gen. Microbiol. 24:51-62. MAN. 1960. Mechanism of protection by sterols 15. PERRITr, A. M., A. W. PHILLIPS, AND T. ROBIN- against polyene antibiotics. J. Bacteriol. 80: SON. 1960. Sterol protection against pimaricin 200-206. in Saccharomyces cerevisiae. Biochem. Bio- 12. LAMPEN, J. 0., AND P. ARNOW. 1961. Inhibition phys. Res. Commun. 2:432-435. of algae by nystatin. J. Bacteriol. 82:247-251. 16. SurroN, D. D., P. M. ARNOW, AND J. 0. LAMPEN. 13. LAMPEN, J. 0., P. M. ARNOW, Z. BOROWSKA, 1961. Effect of high concentrations of nystatin AND A. I. LASKIN. 1962. Location and role of upon glycolysis and cellular permeability in sterol at nystatin binding sites. J. Bacteriol. yeast. Proc. Soc. Exptl. Biol. Med. 108:170- 84:1152-1160. 175.