Introduction

Top Summary Introduction Procedures Results Discussion References

Cyclosporin A (CsA) is a key element in the struggle against rejection in organ transplantation; its introduction into the immunosuppression in 1980s resulted in a major improvement in graft survival (Cohen et al., 1984). This fungal protein is composed of a 11-residue cyclic peptide containing two uncommon amino acids: (4R)-4-[(E)-2-butenyl]-4,N-dimethyl-L-threonine and L--aminobutyric acid, as well as seven peptide bond N-methylated residues. Two conformations have been demonstrated for CsA: one for the unbound (free) peptide crystallized form in an organic solvent and the other for CsA-immunophilin complex or CsA bound to the phosphatase calcineurin (O'Donohue et al., 1995). The binding of CsA to calcineurin is thought to be an important step in its mechanism of action, leading to the decreased synthesis of interleukin-2 (Hemar and Dautry-Varsat, 1990). The clinical use of CsA is limited by its nephrotoxicity, with the molecular level mechanisms of this side effect remaining unresolved (Kopp and Klotman, 1990). CsA has been recently reported to block the opening of mitochondrial permeability transition pore and to interfere with the induction of apoptosis (Bernardi, 1996). Contrasting with these findings, CsA has been shown to also induce apoptosis in some cell lines (Kitagaki et al., 1997).

Due to its profound hydrophobicity, CsA avidly partitions into membranes. CsA has been shown to abolish the pretransition for saturated phospholipids and to decrease both their main transition temperature and enthalpy (O'Leary et al., 1986; Wiedmann et al., 1990). CsA decreases acyl chain order at temperatures below main phase transition, whereas increased order at temperatures above the transition is evident (Wiedmann et al., 1990). CsA increases the lamellar-to-hexagonal phase transition temperature of dielaidoylphosphoethanolamine at low drug-to-lipid molar ratios, whereas a decrease is seen at higher contents of CsA (Epand et al., 1987). CsA also inhibits membrane fusion (Epand et al., 1987).

Coupling between membrane organization and function is emphasized in the modern view of membrane biology, and several molecular level processes generating dynamic order have by now been established in model membranes. Both peripheral and integral membrane proteins, as well as metabolites, ions and drugs, and changes in the temperature and degree of phospholipid hydration, for instance, can alter membrane organization (Kinnunen, 1991; Mouritsen and Kinnunen, 1996). On the other hand, changes in membrane organization can also influence the association of different ligands to the membrane, as shown for doxorubicin (Söderlund et al., unpublished observations), cytochrome c (Mustonen et al., 1987), and histone H1 (Rytömaa and Kinnunen, 1996). Accordingly, there is a reciprocal interplay between membrane organization and binding of ligands to membranes.

Changes in membrane organization have been suggested to be induced by several drugs, such as tacrine, which is used in the treatment of Alzheimer's disease (Lehtonen et al., 1996); the anticancer drug doxorubicin (Goormaghtigh et al., 1982), local anesthetic agents (de-Paula and Schreier, 1996); and the antifungal drug amphotericin B (Lance et al., 1996). A recent approach to molecular-level mechanisms of action (or side effects) of drugs is to elucidate their interactions with membranes and effects on membrane organization in relation to their membrane-involving functions. Due to the high lipid solubility of CsA, it was of interest to study the interactions of this compound with phospholipid membranes. Likewise, because some of the effects of CsA reported so far resemble those exerted by cholesterol, we also investigated the influence of this sterol on the interaction of CsA with lipid membranes.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. HEPES and EDTA were purchased from Sigma (St. Louis, MO), and dimethyl sulfoxide (DMSO) was purchased from Merck (Darmstadt, Germany). 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), egg yolk phosphatidylcholine (eggPC; molecular weight, ~750), 5-cholesten-3-ol (-cholesterol), and 5-cholesten-3-ol (-cholesterol) were obtained from Sigma, and 1-palmitoyl-2-(N-4-nitrobenz-2-oxa-1,3-diazol)aminocaproyl-sn-glycero-3-phosphocholine (NBD-PC) was obtained from Avanti Polar Lipids (Alabaster, AL). CsA was a kind gift from Novartis (Basel, Switzerland). The purity of lipids was checked by thin-layer chromatography on a silicic acid-coated plates (Merck) using chloroform/methanol/water (65:25:4, v/v) as a solvent system. Examination of plates after iodine staining or fluorescence illumination revealed no impurities. To assess the surface chemical purity of DPPC, eggPC, and -cholesterol, their pressure/area (-A) isotherms were recorded using a computer-controlled Langmuir film balance (µTroughS; Kibron Inc., Helsinki, Finland). The isotherms obtained were in keeping with those in the literature. The concentrations of lipid and drug solutions were determined gravimetrically using a high precision electrobalance (Cahn Instruments, Inc., Cerritos, CA) or spectrophotometrically using  = 19,000 M¹ at 465 nm for NBD-PC. Molecular weights of 752 and 724 were used for DPPC and DMPC, respectively, corresponding to their monohydrates.

Penetration of CSA into Lipid Monolayers. Penetration of CsA into monomolecular lipid films was measured using magnetically stirred circular wells (surface area, 31 cm²; volume, 50 ml) drilled in Teflon. Surface pressure () was monitored with a platinum Wilhelmy plate attached to a microbalance connected to a 486 PC via a DT01-EZ data acquisition board. Aqueous phase was 5 mM HEPES and 0.1 mM EDTA (pH 7.4). Lipids were spread on the air-water interface in chloroform (approximately 1 mg/ml) and were allowed to equilibrate for 20 min so as to reach different initial surface pressures (₀) before the addition of CsA (5 µl, 2 mg/ml in DMSO) into the subphase. The increment in  after the addition of CsA was complete in approximately 20 min, and the difference between the initial surface pressure (₀) and the value observed after the intercalation of the drug into the film was taken as . All measurements were performed at ambient temperature (~+24°C). The data are represented as versus ₀, thus revealing the decrement in the intercalation of CsA into lipid monolayer on increasing lateral packing density of the film. These graphs also yield the critical surface pressure _c (i.e., lipid lateral packing density preventing the penetration of the drug into film).

Differential Scanning Calorimetry. Phospholipids, cholesterol, and CsA were codissolved in chloroform to obtain the desired lipid-to-drug ratios. Solvent was first evaporated under a stream of nitrogen, and the dry residues then were maintained under reduced pressure for a minimum of 2 h. Subsequently, samples were hydrated in a buffer (5 mM HEPES, 0.1 mM EDTA, pH 7.4) for 30 min at a temperature of approximately 10°C over the main transition to yield multilamellar vesicles (MLVs) at a total phospholipid concentration of 0.77 mM. Liposomes were maintained at +4°C overnight before recording heat capacity scans with a high-sensitivity differential scanning microcalorimeter (VP-DSC; Microcal Inc., Northampton, MA), operated at a heating rate of 0.5°C/min. All samples had the same thermal history. The calorimeter was connected to a Pentium PC, and data were analyzed using commercial software (Origin; Microcal Inc., Northampton, MA). Transition enthalpies are expressed as kilojoules per mole of phospholipid and were determined by intergration of the peaks, using the internal calibration of the instrument as a reference. The deviation from the baseline was taken as the beginning of the transition and the point of return to the baseline as the end of the transition. Data points represent the mean for triplicate analyses, and the error bars indicate S.D.

Fluorescence Microscopy of Lipid Monolayers. Effect of CsA on the lateral distribution of NBD-PC was studied using a Zeiss IM-35 inverted microscope combined with a computer-controlled monolayer apparatus (µTrough S). Total surface area of the trough is 120 cm², and the volume of the subphase is 25 ml. The trough was mounted on the microscope stage, and the quartz-glass window in the bottom of the trough was positioned over an extralong working distance 20× objective. A 450- to 490-nm bandpass filter was used for excitation, and a 520-nm long-pass filter was used for emission. Images were recorded with a Peltier-cooled 12-bit digital CCD camera (C4742-95, Hamamatsu, Japan) interfaced to a computer and operated by the software (HiPic 4.2.0a) provided by the manufacturer.

	Results

Top Summary Introduction Procedures Results Discussion References

Binding of CsA to Lipid Monolayers. Although the association of CsA with liposomes composed of neutral zwitterionic lipids is well established (O'Leary et al., 1986; Wiedmann et al., 1990), we are not aware of studies on the penetration of CsA into phospholipid monolayers as a function of their lateral packing density. Increment in surface pressure () as a function of time, after the addition of CsA into the subphase underneath a eggPC monolayer at an initial surface pressure (₀) of 15.4 mN/m, is illustrated in Fig. 1. As expected from the hydrophobicity of CsA, the penetration of this peptide into the film was rapid. Similar measurements were subsequently made at varying values for ₀, and the data are illustrated as versus ₀ (Fig. 2A). Interestingly, the data readily reveal the dependence of the interaction of CsA with the monolayer as a function of ₀ to be biphasic. More specifically, at surface pressures below 19 mN/m, the slope of the versus ₀ indicates a limiting value _c of 25 mN/m for the penetration of CsA into the film. However, penetration of CsA is clearly augmented for monolayers initially maintained at ₀ >19 mN/m, and for these films, a significantly higher limiting value of _c 35 mN/m, prohibiting the penetration of CsA into the film, is measured.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Time course of the increase in surface pressure () for an eggPC monolayer over an aqueous subphase (5 mM HEPES, 0.1 mM EDTA, pH 7.4) after the addition of 10 µg of CsA (final concentration, 167 nM). Initial surface pressure (0) was 15.4 mN/m. The contents of the subphase were mixed by magnetic stirring. The experiment was carried out at ambient temperature (+24°C). The addition of CsA is marked by an arrow.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Increase in surface pressure () due to the addition of CsA into the subphase and as a function of initial surface pressure (0). The monolayers used were eggPC (A) and eggPC/-cholesterol (1:1, mol/mol; B). The lines drawn were calculated using least-squares linear fitting, with the following correlation coefficients for the different parts of the graphs: A, r = 0.96569 ( = 11.3-19.2) and r = 0.98175 ( = 19.2-32.5); B, r = 0.99737 ( = 11-21.1) and r = 0.83944 ( = 21.1-29.2). Conditions were as described in the legend for Fig. 1.

Differential Scanning Calorimetry. Previous DSC studies have revealed CsA to decrease the enthalpy (H) of the main phase transition of DPPC, and a location of the drug in the hydrophobic membrane interior was suggested (O'Leary et al., 1986; Wiedmann et al., 1990). A similar effect is observed for cholesterol (Demel and de Kruyff, 1976) and is shown at X_chol = 0.1 for a binary mixture with DMPC (Fig. 3). CsA further broadens the main transition peak of DMPC/-cholesterol (10:1, molar ratio) MLVs (Fig. 3), suggesting possible intercalation of CsA within the hydrocarbon interior of the membrane. The values H for for DMPC/-cholesterol (10:1, mol/mol) with progressively increasing contents of CsA are depicted in Fig. 4A. The main transition enthalpy (denoted as H₀) for the DMPC/-cholesterol MLVs is approximately 11 kJ/mol. At drug-to-lipid ratios ranging from 1:100 to 2:100, CsA increases H by approximately 2 kJ/mol. At drug-to-lipid stoichiometry of 3:100, a sharp dip in H is observed, with H decreasing to 9.1 kJ/mol, yet on further increasing drug-to-lipid stoichiometry to 7:200, H increases. The value for H remains above H₀ at drug-to-lipid ratios of 4:100 and 9:200, whereas a sharp minimum in H is observed at drug-to-lipid ratios of 5:100 and 11:200, with the H being approximately 50% of H₀. At drug-to-lipid molar ratios of 6:100 and 13:200, H again exceeds H₀, where after a 7:100 drug-to-lipid ratio, a smaller dip in H is observed. As the content of CsA is further increased up to drug-to-lipid stoichiometry of 10:100, the value for H increases, becoming nearly twice the value of H₀. Less dramatic changes are evident in T_m (Fig. 4), yet on exceeding X_CsA = 0.06, there is a decrement in T_m by approximately 0.6 degree.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Broadening of the main phase transition peak of DMPC: -cholesterol (10:1, mol/mol) MLVs on increasing XCsA, measured by DSC. Total lipid concentration was 0.77 mM in 5 mM HEPES and 0.1 mM EDTA, pH 7.4. The heating rate was 0.5°C/min.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   A, Changes in the main H of DMPC/-cholesterol (10:1, mol/mol) MLVs due to increasing concentrations of CsA. Total lipid concentration was 0.77 mM in 5 mM HEPES and 0.1 mM EDTA, pH 7.4. All samples had the same thermal history. Scanning rate was 0.5°C/min. Each data point represents the mean of triplicate analyses, with the error bars indicating ±S.D. B, Changes in the peak of the maximum heat capacity (Tm) of DMPC/-cholesterol (10:1, mol/mol) MLVs due to increasing concentrations of CsA.

Fluorescence Microscopy of the Monolayers. Morphology of the two-dimensional domains of lipid monolayers is sensitive to their chemical composition (e.g., Weis and McConnell, 1985; Weis, 1991). To obtain further insight into the effects of CsA on lipid alloys, we studied the lateral distribution of the fluorescent lipid analog NBD-PC (X = 0.02) in DPPC monolayers as a function of . Representative images of a DPPC monolayer at surface pressures of 13.1 and 18.7 mN/m are shown in Fig. 5, A and C. These pressures correspond to the transition region, evident as the coexistence of the fluid and solid regions. As the fluorescent probe, NBD-PC readily partitions into the former domains; these appear as light and dark areas, respectively (Weis and McConnell, 1985). In the absence as well as the presence of CsA, dark, crystalline domains containing very low amounts of probe first appeared on compression to surface pressures between 9 and 10 mN/m (Fig. 5B). The presence of CsA (X = 0.05) diminishes domain size, whereas the length of the fluid/gel (liquid-expanded/liquid-condensed) domain boundary increases. At higher pressures, the effect of CsA on the domain boundaries becomes more pronounced (Fig. 5B). For example, for DPPC/NBD-PC film at 18.7 mN/m, the domain boundaries are diffuse, with a gradient of decreasing NBD-PC fluorescence on going from crystalline to fluid region (Fig. 6). This is contrasted by the sharp boundaries observed in the presence of CsA (Fig. 6). The shape of the domains observed with CsA somewhat resembles those reported for low mole fractions (X = 0.02) of cholesterol (Weis and McConnell, 1985).

View larger version (175K): [in this window] [in a new window]   Fig. 5.   Fluorescence microscopy images of DPPC/NBD-PC (98:2, mol/mol) monolayers in the absence (A and C) and presence of CsA (X = 0.05; B and D). The subphase is 5 mM HEPES and 0.1 mM EDTA, pH 7.4, at ambient temperature (~+24°C). The compression rate was 1 Å2/acyl chain/min. Values of surface pressure were (A) 13.1 mN/m, (B) 14.1 mN/m, (C) 18.7 mN/m, and (D) 17.8 mN/m. Scale bar is 20 µm. See text for more details.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of CsA on the lateral distribution profile of NBD fluorescence intensity within the domain boundaries for DPPC/NBD-PC (98:2, mol/mol) films. The data were taken from C (no CsA, graph A) and D (XCsA = 0.05, graph B) in Fig. 5 and were fitted by using sigmoidal function. For both graphs, the data were normalized to one measured for the fluid domain.

View larger version (117K): [in this window] [in a new window]   Fig. 7.   Changes in the NBD-PC (X = 0.02) distribution in DPPC/-cholesterol (88:10, mol/mol) monolayers caused by CsA, observed by fluorescence microscopy. The lipid alloy organization is shown in the absence (A-D) and the presence of CsA (X = 0.05, E-H). Surfaces pressures were (A) 13.3 mN/m, (B) 18.8 mN/m, (C) 19.6 mN/m, (D) 30.5 mN/m, (E) 13.8 mN/m, (F) 17.7 mN/m, (G) 22.1 mN/m, and (H) 31.4 mN/m. Subphase was 5 mM HEPES and 0.1 mM EDTA, pH 7.4. Scale bar is 20 µm. See text for further details.

	Discussion

Top Summary Introduction Procedures Results Discussion References

Conventionally, drugs are a priori assumed to exert their action in cells by more or less specific binding to sites in proteins. Interestingly, a novel concept has emerged from studies on the membrane binding of cytotoxic peptides such as magainins, cecropins, and defensins (Bechinger, 1997). No receptors have been demonstrated for these compounds, and they are presently believed to exert their activity by interacting with the lipid bilayer. Yet, the molecular mechanism or mechanisms of their action and the property of the bilayer being modified remain to be established. In this regard, the rich scale of different phases (i.e., membranes with distinct physicochemical properties) exhibited by different lipids is of interest (Kinnunen and Laggner, 1991). There is experimental evidence indicating a correlation between the physical state (i.e., the phase state of cellular membranes), determined by their lipid composition, and the physiological state of the cell (Kinnunen, 1991, 1996). There is no reason to believe that the modulation of specific properties of the lipid bilayer would be limited to the above cytotoxic peptides; it could contribute to both the therapeutic mechanism and side effects of membrane-partitioning compounds in general. To this end, a large variety of structurally dissimilar drugs are hydrophobic or amphiphilic and readily partition into lipid membranes; good examples are tacrine (Lehtonen et al., 1996), doxorubicin (Mustonen et al., 1993; Söderlund et al., 1999), and CsA (O'Leary et al., 1986; Wiedmann et al., 1990). Drugs may also modulate peripheral lipid-protein interactions, as shown for chlorpromazine, doxorubicin, lidocaine, and gentamycin (Ito et al., 1983; Mustonen and Kinnunen, 1991; Jutila et al., 1998).

The increase in acyl chain order in DPPC liposomes by CsA is similar to that caused by cholesterol (Wiedmann et al., 1990). Increasing the contents of cholesterol progressively decreases H of the main transition in a manner depending on the acyl chain composition of the matrix lipid (McMullen et al., 1993). Comparison of these data on the effects of cholesterol and CsA suggests the localization of these compounds in membranes to be similar. For cholesterol, the long axis of its sterol ring structure parallels the acyl chains of the membrane lipids, with the hydroxyl group residing vicinal to the phospholipid ester carbonyl groups. The side chain is buried in the membrane interior. CsA occupies a larger area in the interior of the membrane compared with the membrane surface (Wiedmann et al., 1990). Also, the present DSC data are consistent with the intercalation of CsA into the membrane interior (O'Leary et al., 1986; Wiedmann et al., 1990).

The thermal phase behavior of DMPC/-cholesterol (10:1, mol/mol) liposomes as a function of X_CsA is complex (Fig. 4). A likely explanation to these data could be provided by the same principles as forwarded for tacrine-induced changes in the thermal behavior of dimyristoylphosphatidic acid (Lehtonen et al., 1996). More specifically, the latter results were interpreted in the terms of formation of regular superstructured regions in the bilayer at well-defined drug-to-phospholipid ratios. In principle, all systems organize so as to minimize their free energy. In a bilayer alloy, this may require its components to respond to changes in composition by changes in organization. This is best implied by the alterations observed in the main transition enthalpy at CsA-to-cholesterol ratios of 3:10 and 1:2 (X_CsA = 0.03 and 0.05, respectively). The formation of superlattices in liposomes containing cholesterol has been proposed (Liu et al., 1997). CsA could exert effects similar although not identical to those of cholesterol (Wiedmann et al., 1990).

The biphasic, packing density-dependent interaction of CsA with the phospholipid monolayer is peculiar and reveals cholesterol to decrease the penetration of CsA into the lipid monolayer in a surface pressure-dependent manner (i.e., at pressures exceeding 22 mN/m) (Fig. 2). Because the inclusion of cholesterol leads to a increased lateral packing density of phosphatidylcholine monolayers (Gershfeld and Pagano, 1972), it also decreases the free volume within the hydrophobic part of the monolayer. Obviously, this would impede the penetration of CsA into the lipid, yet more specific lipid-drug interactions also could be involved and could reflect a pressure-induced change in the conformation and/or orientation of CsA. In this case, the conformation and/or orientation of CsA in the membrane would also be sensitive to cholesterol.

Fluorescence microscopy of lipid monolayers has been used to study the lateral organization of lipids (Nag et al., 1991), changes in lipid domains induced by phospholipase A₂ (Grainger et al., 1989), monolayer phase transitions (Weis, 1991), domain formation induced by Ca⁺⁺ (Eklund et al., 1988), and domain formation induced by electric fields (Lee et al., 1994). Lipid-protein interactions have also been studied by fluorescence microscopy of monolayers (Subirade et al., 1995). The determinants for domain morphologies has been a subject of intense research (Weis, 1991), and factors such as dipole fields and line tension have been shown to contribute. To our knowledge, this technique has not been used to investigate possible changes in domain morphology caused by drugs. The fluorescent probe NBD-PC used in this study preferentially partitions into the "fluid" (liquid-expanded) lipid (Weis and McConnell, 1985). Accordingly, fluorescence microscopy of phospholipid monolayers allows visualization of the coexistence of fluid and gel domains in the transition region (Weis, 1991). In the transition region, three phases with distinctly different fluorescence intensities are evident (Fig. 7). The dark and light areas represent gel-like and fluid domains, respectively, separated by a boundary region with a gray, intermediate, and diffuse gradient of fluorescence emission. Evidence for an "intermediate" lipid domain within the main transition has been recently forwarded for DMPC liposomes (Jutila and Kinnunen, 1997). More complex behavior is observed when -cholesterol is present. DPPC/-cholesterol/NBD-PC monolayers exhibit reticular domains at higher surface pressures. The effect of CsA on the lateral organization of this system is dramatic, and only gel- and fluid-like domains are evident. The change in the domain shape is also pronounced, from complex reticular networks to completely circular domains at all surface pressures. The effect of CsA on the lateral organization of pure DPPC monolayers suggests that it partitions into the boundaries between gel/liquid domains, with these boundaries becoming sharp in the presence of this drug.

The present results show that the interaction of CsA with membranes containing cholesterol are much more complex than those revealed in previous studies with neat PC bilayers (Wiedmann et al., 1990). In brief, we demonstrated that CsA not only changes the thermal phase behavior of the membrane but also alters dramatically the lateral organization in monolayers on a micrometer scale. Although direct comparison of the different membrane models, liposomes and monolayers, is ambiguous, we can conclude that both systems demonstrate that the interaction of CsA with membrane lipids depend on the presence of cholesterol. There is a large body of evidence showing that cholesterol affects a number of processes of diverse nature in different cells. Moreover, organization of cholesterol in membranes can be anticipated to be critical to its functions (see Schroeder et al., 1995, and references therein). Definitive conclusions regarding the pharmacological significance of our findings would clearly be premature at this stage, yet in conjunction with the importance of coupling between organization and function in biomembranes, the present results do indicate that efforts along these lines may provide novel insights to the understanding of the molecular mechanisms of action of membrane-associating drugs.