Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

SR31747 [cis-N-cyclohexyl-N-ethyl-3-(3-chloro-4-cyclohexylphenyl)propen-2-ylamine hydrochloride; or SR 31747A], is a drug that exhibits immunosuppressive and anti-inflammatory properties (Paul et al., 1994; Casellas et al., 1994; Derocq et al., 1995; Carayon et al., 1995; Bourrié et al., 1995). In Saccharomyces cerevisiae, this drug was shown to inhibit the activity of SI, an enzyme of the sterol biosynthetic pathway encoded by the ERG2 gene (Silve et al., 1996a). As a first step toward elucidating whether sterol isomerase could be an SR 31747 target in mammals as well, we cloned and expressed an mSI-encoding cDNA in yeast (Silve et al., 1996b). We found that this enzyme was identical to EBP, an endoplasmic reticulum-resident membrane protein structurally dissimilar to the yeast enzyme, whose function was unknown (Hanner et al., 1995). In this report, we address the question of whether SR 31747 is indeed an inhibitor of the mammalian enzyme. This was achieved by characterizing the recombinant enzyme pharmacology and studying the effect of this drug on cell proliferation and sterol composition.

Tamoxifen is a nonsteroidal antiestrogen of the triphenylethylene type. This chemical, which binds to the nuclear estrogen receptor with high affinity, is being widely used as a therapeutic agent in estrogen-dependent tumor therapy, principally in breast cancer. Tamoxifen and other structurally related molecules are also known to lower serum LDL cholesterol in women, either affected with breast cancer or not, suggesting that these compounds might reduce the risk of cardiovascular disease. Long-term treatment of breast cancer patients with tamoxifen is indeed associated with reduced cardiovascular mortality (see Mikhailidis and Spyropoulos, 1996, for a review). In addition to the estrogen receptor, tamoxifen binds to unidentified binding sites that are localized in the cell microsomal fraction and are known as AEBS (Sutherland et al., 1980; Faye et al., 1980; Sudo et al., 1983). It has been suspected for several years that AEBS mediates the tamoxifen-induced inhibition of cholesterol biosynthesis (Cypriani et al., 1988; Teo et al., 1992). Recently, a series of structurally diverse drugs, including tamoxifen and various steroids, was shown to arrest cell proliferation by inhibiting cholesterol biosynthesis and provoking cholesterol auxotrophy in mammalian cells in culture (Metherall et al., 1996a, 1996b). Drug-treated cells accumulated lanosterol and a number of other and as yet unidentified cholesterol precursors at the expense of cholesterol in a drug concentration-dependent manner. Because all these cholesterol biosynthesis-inhibiting drugs were also known as ligands of the MDR, Metherall et al. (1996b) proposed that the drug-induced cholesterol biosynthesis arrest was mediated by MDR inhibition. Independently, in a study on postmenopausal women with advanced breast cancer, Gylling et al. (1995) reported that tamoxifen provoked a 15% decrease in serum LDL cholesterol and a 55-fold increase in serum 8-choletenol. These authors concluded that the cholesterol-lowering effect of tamoxifen was exerted chiefly at the level of the 8-7 isomerization step. Taken together, these observations suggested that the cholesterol synthesis-blocking effect by some of, if not all, the drugs studied by Metherall et al. (1996a, 1996b) could be exerted by the direct inhibition of one or several enzymes of the sterol biosynthetic pathway, like 8-7 sterol isomerase in the case of tamoxifen. We have tested this hypothesis by studying the effects of tamoxifen and of several other drugs on yeast cell proliferation, 8-7 sterol isomerase activity and SR31747 binding in EBP-producing cells.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents. SR31747 and [³H]SR31747 were synthesized by Sanofi Recherche. Tamoxifen, ifenprodil and verapamil were obtained from Sigma Chimie (Paris, France). 7-Ketocholestanol was obtained from Steraloids (Wilton, NH). Tridemorph and fenpropimorph were kindly supplied by Dr. A. Akers (BASF AG, Limburgerhof, Germany). Emopamil was a gift of Dr. G. Gross (Knoll AG, Ludiwgshafen, Germany). All the other drugs were supplied by Research Biochemical (Natick, MA).

Strain. The host strain used in transformation experiment was Saccharomyces cerevisiae EMY43 (MAT, ura3, trp1-4, erg2::TRP1). This strain is a haploid congenic derivative of wt FL100 (American Type Culture Collection 28383, Rockville, MD). The EMY43 genome harbors a disrupted allele of the ySI-encoding ERG2 gene conferring ergosterol prototrophy and aerobic lethality.

Growth media. EMY43 cells depend on exogenously supplied sterol (generally ergosterol) to proliferate. Because they are unable to utilize exogenously supplied sterol in the presence of oxygen, EMY43 cells cannot proliferate under aerobic conditions and are consequently grown anaerobically in ergosterol-containing medium (Silve et al., 1996a). Yeast culture media were either rich or minimal (YPD and SD, respectively) adequately supplemented to fulfill the strain's auxotrophic requirement (Guthrie and Fink, 1991).

Plasmids. The construction and structure of the various sterol-isomerase expression plasmids for yeast used throughout this study were fully described elsewhere (Silve et al., 1996b). In summary, pEMR1023 is a S. cerevisiae-Escherichia coli shuttle vector that contains URA3 as the selectable marker, a yeast 2-µ plasmid fragment containing the ARS and STB sequences for plasmid maintenance at a high copy number in yeast and an empty expression cassette made of the strong promoter and terminator of the yeast PGK1 gene. Plasmid pEMR1235 and pEMR1336 are derived from pEMR1023 by the integration of an mSI- and hSI-encoding cDNA sequence, respectively. pEMR1200, pEMR1292, pEMR1348 and pEMR1349 are low-copy centromeric plasmids. pEMR1200 contains the S. cerevisiae ERG2 gene, which encodes ySI. pEMR1292 harbors the same sterol isomerase-expression cassette as in pEMR1235. pEMR1348 and pEMR1349 are hSI- and mSI-expressing vectors, respectively, in which the cDNA expression is under the control of the weak ERG2 promoter. Cells transformed by pEMR1292, pEMR1348 and pEMR1349 are low SI producers.

In vivo drug tests. The effect of drugs on transformed cell proliferation was tested as follows: 5 µl of each cell suspension (~10⁵ cells) were plated onto medium containing drugs at different concentrations, as indicated, and plates were incubated at 30°C for 24 hr.

Binding experiments. Receptor binding assays for [³H]SR31747 were performed as already described (Paul et al., 1994) with minor modifications. The binding buffer was 50 mM Tris-HCl, 1.5 mM EDTA, 0.01% NaN₃ and 0.1% BSA pH 7.4 (25°C). Yeast homogenates were resuspended in 50 mM Tris-HCl buffer, pH 8.0, (4°C) for membrane preparation. In the competition experiments, 15 µg of membrane proteins and 0.2 nM [³H]SR31747 were incubated for 2 hr 30 min at 30°C in 1 ml of the corresponding buffer containing the different drugs at different concentrations (0.1 nM to 10 µM). In the saturation experiments, [³H]SR31747 (0.1-10 nM) was incubated in the presence or absence of tamoxifen. Nonspecific binding was determined in the presence of 1 µM SR31747. The membranes were separated from the free radioligand by filtration on GF/B filters soaked with 0.5% polyethylenimide, and after washing, radioactivity was counted.

Sterol analysis. Sterols were extracted from lyophilized cells as already described (Silve et al., 1996a). Sterol analysis was carried out using a Variant 3300 chromatograph, using a 30 m-DBI column (inner diameter, 0.312 mm) with a Ross injector, under helium as the carrier gas and a column oven temperature programmed from 200°C to 250°C to 300°C at 10°C/min and 2°C/min, respectively. Detection was obtained by a FID with a temperature of 300°C.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

SR31747 arrests yeast cell proliferation even in the yeast cells that produce hSI in place of their own sterol isomerase. EMY43 is a sterol isomerase-deficient strain that cannot proliferate under aerobic conditions as a result of ergosterol auxotrophy (Silve et al., 1996a). Ergosterol prototrophy and aerobic proliferation are restored on transformation with vectors expressing a sterol isomerase-encoding sequence (Silve et al., 1996b). We studied the effect of SR31747 on the proliferation of EMY43 cells that were producing mSI or hSI instead of ySI. Interestingly, we found that these cells became hypersensitive to the antiproliferative effect of this drug (fig. 1). Yeast transformants producing mammalian SI at high level (Silve et al., 1996b) were arrested by SR31747 at concentrations of <1 µM, whereas the concentration required to block proliferation in wt cells was 2-fold higher. Low SI producers were arrested by SR31747 at concentrations of <0.1 µM, with 10 nM being even sufficient to arrest cell proliferation in the case of transformants harboring pEMR1349, a centromeric plasmid expressing mSI under the control of the weak ERG2 promoter. Sterol analyses of drug-treated cells confirmed the accumulation of the 8-sterol molecules that denote sterol isomerase inhibition (fig. 2).

View larger version (74K): [in this window] [in a new window]   Fig. 1.   Effects of various drugs on aerobic proliferation in EMY43 transformants. Plasmids used to transform EMY43 cells were pEMR1200, pEMR1235, pEMR1292, pEMR1349, pEMR1336 and pEMR1348, respectively (see Materials and Methods). About 105 cells were plated in the absence of drug () or in the presence of SR31747, tridemorph and tamoxifen, at various concentrations (µM) as indicated.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Analysis of sterols produced by EMY43 transformants expressing either ySI (EMY43 [pEMR1200]) or hEBP (EMY43 [pEMR1336]). EMY43 [pEMR1200] and EMY43 [pEMR1336] transformants were grown in YPD for 24 hr in the absence () or presence of drugs as indicated. Sterols were extracted and analyzed as described in the text. Peak a corresponds to ergosterol, and peaks b, c and d were identified as ergosta-5,8,22-trien-3-ol, ergost-8-en-3-ol and ergosta-8,14-dien-3-ol, respectively (Silve et al., 1996a, 1996b).

To evaluate the affinity of SR31747 to hSI and mSI, binding studies were performed on membrane extracts from cells that were producing mSI or hSI instead of the yeast enzyme. Scatchard analysis revealed only one population of high affinity sites in each case with similar K_d values (0.93 ± 0.06 and 1.09 ± 0.25 for mSi and hSI, respectively; fig. 3), whereas no specific binding was observed with an erg2 gene disruptant devoid of any sterol isomerase (data not shown). As expected, ySI also bound [³H]SR31747 with high affinity (K_d = ~5 nM) (our results to be published elsewhere). [³H]SR31747 binding inhibition studies revealed that ySI and mammalian SI exhibited distinct pharmacological properties (fig. 4, table 1). Some drugs, like cis-flupentixol, 7-ketocholestanol, trifluoperazine and tamoxifen, inhibited SR31747 binding only with the mammalian enzymes, whereas other drugs, like fenpropimorph and haloperidol, were much more effective with the yeast enzyme than with the mammalian ones. Interestingly, tamoxifen inhibited SR31747 binding in a competitive manner (fig. 3). In contrast, emopamil was not effective as SR31747 binding inhibitor, although this calcium channel blocker is a high affinity ligand of hSI (also known as emopamil-binding protein) (Hanner et al., 1995). Finally, mSI and hSI were quite similar from a pharmacological viewpoint; the only striking difference concerned SKF-525A, which selectively blocked [³H]SR31747 binding with hSI.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Scatchard plots of [3H]SR31747 binding to recombinant mSI. Saturation analysis of [3H]SR31747 binding to mSI produced by pEMR1235-transformed cells. Experiments were performed as described in Materials and Methods. Each value is given as the mean of triplicate determinations. The figure shows the Scatchard transformation of these data. Binding experiments were done in the absence () or presence of 2 and 5 nM cold tamoxifen ( and , respectively). KD values were 0.93 ± 0.06, 2.55 ± 0.16 and 6.21 ± 0.51 nM, whereas Bmax values did not change significantly, being estimated at 22,370 ± 493, 20,860 ± 623 and 25,230 ± 1307 pmol/mg protein, respectively. A similar Scatchard analysis of [3H]SR31747 binding to hSI produced by pEMR1336-transformed cells gave a KD value of 1.10 ± 0.25 in the absence of tamoxifen (not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Inhibition of [3H] SR31747 binding to mSI by various drugs. Drugs were as indicated. Similar experiments were done with hSI and ySI. Corresponding IC50 values are given in table 1. mSI, hSI and ySI were extracted from cells transformed by pEMR1235, pEMR1336 and pEMR1200, respectively.

Tamoxifen is an efficient sterol biosynthesis inhibitor only in mammalian SI-producing cells. In agreement with the binding results, the antiestrogen tamoxifen was not found to exert any antiproliferative effect on ySI-producing strains, at least at concentrations up to 90 µM (fig. 1). In contrast, pEMR1336- and pEMR1235-transformed cells that were producing mammalian SI stopped proliferating in the presence of tamoxifen at concentrations of 2 to 9 µM. This effect was still more pronounced in low SI-producing strains, with 0.2 µM tamoxifen sufficient to inhibit cell proliferation in pEMR1292- and pEMR1349-transformed cells, which are low producers of mSI and hSI, respectively. To ensure that the tamoxifen-induced proliferation arrest was indeed linked to a sterol biosynthesis blockade, sterols extracted from ySI- or hSI-expressing cells grown in the presence of drug were analyzed by gas chromatography. Results confirmed that tamoxifen induced the accumulation of the 8-sterols that characterize sterol isomerase inhibition (fig. 2). This effect was observed only with mammalian SI and not with ySI, whereas SR31747 and tridemorph inhibited both SI types, as expected from the binding inhibition results (table 1).

The N-substituted dimethylmorpholine tridemorph is more effective than SR31747 as an antiproliferative agent in mammalian SI-producing strains. Tridemorph is a potent sterol SBI that is widely used in agriculture as an antifungal agent. In yeast, this N-substituted dimethylmorpholine derivative is ~100-fold more effective against ySI vs. yeast C14-sterol reductase (Baloch and Mercer, 1987). Tridemorph inhibits SI in a competitive manner, supposedly by mimicking the high-energy C8-carbonium intermediate involved in this isomerization reaction (Baloch and Mercer, 1987). pEMR1200-transformed cells that expressed ySI were ~10-fold more sensitive to tridemorph than to SR31747 (fig. 1). Tridemorph also was powerfully effective in inhibiting cell proliferation in hSI- and mSI-producing strains. In the absence of any SI inhibitor, cells transformed by either pEMR1336 or pEMR1235, an hSI- and mSI-expressing vector, respectively, are producing 7-sterol and 8-sterol at levels similar to wild-type (Silve et al., 1996a). These strains stopped proliferating in the presence of 0.1 and 0.3 µM tridemorph, respectively, whereas up to 1 µM SR31747 was required to obtain a similar arrest effect on the same strains. Low mammalian SI producers (i.e., cells transformed by pEMR1292, pEMR1348 and pEMR1349, respectively) produce large amounts of 8-sterols, even in the absence of any inhibitor (Silve et al., 1996a), as a consequence of rate-limiting SI activity. Such strains were found to be hypersensitive to SI inhibitors (fig. 1; see also Silve et al., 1996b) and still remained more susceptible to tridemorph than to SR31747. These comparative results suggested that tridemorph was more potent than SR31747 as an SI-inhibiting drug, even when the enzyme was from a mammalian source. Indeed, this dimethylmorpholine derivative was found to be ~15- and ~2-fold more effective in displacing [³H]SR31747 binding to ySI and mammalian SI, respectively, than cold SR31747 itself (table 1). In contrast, fenpropimorph, another N-substituted dimethylmorpholine derivative, was ~400- and ~300-fold less effective with mSI or hSI than with ySI (table 1). These pharmacological results are consistent with our previous data showing that fenpropimorph was ~50-fold less efficient as an antiproliferating drug and an SBI in yeast cells that were producing mSI or hSI instead of ySI (Silve et al., 1996b). In mammalian SI-producing cells, fenpropimorph is more effective against C14-sterol reductase than against sterol isomerase, whereas the reverse is true in wild-type cells (Silve et al., 1996b; Marcireau et al., 1990).

Ifenprodil is not effective as an antiproliferation agent except in low mammalian SI-producing strains. Ifenprodil is known to be a high affinity ligand of both human emopamil-binding protein and ySI (K_i = 2 and 1 nM; see Moebius et al., 1997, for a review). However, this drug was found to be only moderately efficient as an inhibitor of SR31747 binding to any of the three enzymes (IC₅₀ = 15-50 nM; see table 1). Surprisingly, this drug was found not to arrest cell proliferation in our ySI-producing cells, at least at concentrations up to 125 µM (fig. 1). Exchange of ySI with mammalian SI did not significantly alter ifenprodil sensitivity in pEMR1235- or pEMR1336-transformed strains. In contrast, cells transformed by low mSI or hSI expression plasmids (pEMR1292, pEMR1348 or pEMR1349) were susceptible to ifenprodil concentrations exceeding 12 to 25 µM. Sterol analyses of mammalian SI producers confirmed that this drug moderately inhibited sterol biosynthesis at the 8-7 sterol isomerase and possibly the C14-sterol reductase steps, as judged by the relative accumulation of both 8 and 8-14 sterols in treated cells.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The comparison of drug susceptibilities in yeast strains that produce SI from various origins enables the detection of inhibitors that are specific for one type of enzyme or for another. Although fungal and mammalian sterol isomerases are not structurally related, our work indicate that they share several nonsterol inhibitors that presumably act by mimicking the carbocationic reaction intermediate. Our results show, for instance, that tridemorph and SR31747 are not selective for the yeast enzyme. Tridemorph is even more effective than SR31747 in inhibiting cell proliferation in mammalian SI-producing yeast strains; this high efficacy of this drug can be correlated with its strong capacity to displace [³H]SR31747 binding to mSI and hSI (table 1). In contrast, fenpropimorph, which is an SBI structurally related to tridemorph, more efficiently inhibits [³H]SR31747 binding to sterol isomerase when the enzyme is of a yeast source. This latter result is consistent with our previous observation (Silve et al., 1996b) indicating that this antifungal agent was a less potent SBI when ySI was replaced by mSI or hSI. In mammals, more precisely, in cultured Swiss 3T3 fibroblasts, fenpropimorph is indeed known as an SBI, but its main target appears to be P-450 lanosterol demethylase (Corio-Costet et al., 1988). It is worth noting that both tridemorph and fenpropimorph also bind with very high affinity to the so-called sigma receptor subtype 1 (Moebius et al., 1997), also known as SR 31747-binding protein, or SRBP (Jbilo et al., 1997). Unlike mammalian SI, SRBP shares considerable structural similarities with yeast sterol isomerase. Although the actual function of SRBP remains elusive, this protein is thought to mediate the psychotropic and neuroprotective effects that are typical of most sigma ligands (Moebius et al., 1997), as well as the immunomodulatory effect of SR 31747 (Jbilo et al., 1997). Recently, fenpropimorph was shown to exhibit long-term neuroprotective properties in vitro, a feature shared with other conventional sigma ligands (Lesage et al., 1995). Thus, the risks posed by the utilization of these sterol biosynthesis inhibitors as antifungal agents, if any, should be reexamined carefully.

Both tamoxifen and trifluoperazine selectively displace [³H]SR31747 binding to mammal enzymes but not to ySI. Here again, these data are consistent with our previous results showing that trifluoperazine inhibits cell proliferation and sterol biosynthesis only in the yeast strains that are producing mSI or hSI instead of ySI. Although tamoxifen inhibits [³H]SR31747 binding to hSI in a competitive manner with the same efficacy as cold SR31747 itself, it is nevertheless a 10-fold less efficient antiproliferating drug compared with SR31747 in recombinant yeast cells. This tamoxifen-resistance phenotype is likely due to the activity of ABC transporters. Indeed, tamoxifen is a well known ligand of these multidrug resistance pumps in yeast as in mammalian cells (Kolaczkowski et al., 1996). Moreover, FL100 was shown to harbor a mutated PDR1 gene that induces the high level expression of several ABC transporter-encoding genes in a constitutive fashion (Gilbert et al., 1993). This mutation confers a multidrug-resistance phenotype (Gilbert et al., 1993). Similarly, even though both haloperidol and ifenprodil are high affinity ligands of ySI with K_d values of 0.3 and 1.4 nM, respectively (Moebius et al., 1996, 1997), which efficiently inhibit [³H]SR31747 binding to the yeast enzyme (IC₅₀ = 1 and 16 nM, respectively), none of these drugs were found to exert any efficient inhibitory effect on sterol biosynthesis and cell proliferation in our hands (Silve et al., 1996b; the current study). However, Moebius et al. (1996) reported that both haloperidol and ifenprodil were indeed efficient ergosterol biosynthesis inhibitors in other wild-type strains of yeast. Here again, a strain-dependent difference in pump activity level could account for this apparent discrepancy. A FL100 congenic strain that harbors a disrupted version of the PDR1 gene, thus producing ABC transporters at lower levels, was shown to exhibit a ~5-fold hypersensitivity to the antiproliferative effects of tamoxifen, ifenprodil and haloperidol, whereas the sensitivity of such a pdr1 gene disruptant to the antiproliferative effect of SR 31747 was not affected (results to be published elsewhere). ABC transporter activities mediate resistance to tamoxifen, haloperidol and ifenprodil supposedly by preventing these drugs from accumulating in the endoplasmic reticulum membrane where sterol isomerization takes place.

[³H]Emopamil is a high affinity ligand of hSI if expressed in yeast or liver (Hanner et al., 1995). However, emopamil was quite ineffective in the inhibition of [³H]SR31747 binding to mammalian SI in that a 1 µM concentration of the former drug was not a sufficiently high concentration to produce 50% inhibition. Verapamil, another calcium channel-blocking agent, was shown to powerfully inhibit emopamil binding to hSI (Hanner et al., 1995), whereas SR31747 binding to the same enzyme was not considerably inhibited by this drug (current study). These results strongly suggest that the binding sites of emopamil or verapamil and of SR31747 on this enzyme do not overlap. The possibility that sterol isomerase contains a calcium-binding regulatory site remains to be investigated. Verapamil was shown to inhibit cholesterol biosynthesis at an undetermined postlanosterol step in Chinese hamster ovary cells; this effect was obtained at micromolar concentrations. No such effect of verapamil could be detected in our yeast tests (data not shown). Further experiments are therefore required to determine whether drugs of the verapamil family are indeed sterol isomerase inhibitors.

Our results show for the first time that tamoxifen is a high affinity ligand of mammalian SI (EBP) (IC₅₀ value in the nanomolar range). Tamoxifen is known to bind with high affinity to the estrogen receptor but also to binding sites localized in the cell microsomal fraction. It is clear from our results that mammalian 8-7 sterol isomerase (EBP) displays such a binding site. The observation that both tamoxifen and SR31747 inhibit the EBP-catalyzed sterol isomerase reaction in yeast is consistent with our data showing that the two drugs share the same binding site on this enzyme, which presumably corresponds to the sterol-binding pocket of the active site. Our results showing that tamoxifen is a high affinity inhibitor of mammalian SI support the hypothesis first raised by Gylling et al. (1995), who suggested that the main hypocholesteremiant effect of this antiestrogen was exerted through the reduction of SI activity. Because the overproduction of plasma LDL cholesterol constitutes a well known atherogenic risk factor, this SI activity inhibition likely contributes to the decrease in the ischemic heart disease frequency that is observed among long-term users of tamoxifen.