Comparison of the effects of clozapine, chlorpromazine, and haloperidol on membrane lateral heterogeneity

Chemistry and Physics of Lipids 112 (2001) 151– 163 www.elsevier.com/locate/chemphyslip

Comparison of the effects of clozapine, chlorpromazine, and haloperidol on membrane lateral heterogeneity Arimatti Jutila a,1, Tim So¨derlund a,1, Antti L. Pakkanen a, Matti Huttunen b, Paavo K.J. Kinnunen a,* a

Helsinki Biophysics and Biomembrane Group, Institute of Biomedicine/Biochemistry, P.O. Box 63 (Haartmaninkatu 8), FIN-0014 Uni6ersity of Helsinki, Helsinki, Finland b Department of Psychiatry, Uni6ersity of Helsinki, Helsinki, Finland Received 25 January 2001; received in revised form 20 June 2001; accepted 11 July 2001

Abstract The interactions of three neuroleptic drugs, clozapine (CLZ), chlorpromazine (CPZ), and haloperidol (HPD) with phospholipids were compared using DSC and Langmuir balance. Main emphasis was on the drug-induced effects on the lateral organization of lipid mixtures of the saturated zwitterionic 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC) and the unsaturated acidic phosphatidylserine, brainPS. In multilamellar vesicles (MLV) phase separation was observed by DSC at XPS E0.05. All three drugs bound to these MLVs, abolishing the pretransition at Xdrug E0.03. The main transition temperature (Tm) decreased almost linearly with increasing contents of the drugs, CLZ having the smallest effect. In distinction from the other two drugs, CLZ abolished the phase separation evident in the endotherms for DPPC/brainPS (XPS =0.05) MLVs. Compression isotherms of DPPC/brainPS/drug (XPS = Xdrug =0.05) monolayers revealed the neuroleptics to increase the average area/molecule, CLZ being the most effective. Penetration into brainPS monolayers showed strong interactions between the three drugs and this acidic phospholipid (in decreasing order CPZ \ HPD\CLZ). Hydrophobic interactions demonstrated using neutral eggPC monolayers decreased in a different order, CLZ \CPZ\HPD. Fluorescence microscopy revealed domain morphology of DPPC/brainPS monolayers to be modulated by these drugs, increasing the gel-fluid domain boundary length in the phase coexistence region. To conclude, our data support the view that membrane-partitioning drugs could exert part of their effects by changing the lateral organization and thus also the functions of biomembranes. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Drug– lipid interactions; Fluorescence microscopy; Membrane organization; Monolayer; Neuroleptics

* Corresponding author. Fax: + 358-9-1918276. E-mail address: [email protected] (P.K.J. Kinnunen). 1 A.J. and T.S. contributed equally to this study. 0009-3084/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 0 9 - 3 0 8 4 ( 0 1 ) 0 0 1 7 5 - X

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

152

Nomenclature brainPS CLZ CPZ DPPC eggPC HPD MLV NBD-PC y Dy y0 Tm Tp

brain phosphatidylserine clozapine chlorpromazine 1,2-dipalmitoyl-sn-glycero-3-phosphocholine egg yolk phosphatidylcholine haloperidol multilamellar vesicle 1-palmitoyl-2-(N-4-nitrobenz-2-oxa-1,3-diazol)aminocaproyl-sn-glycero-3-phosphocholine surface pressure change in surface pressure initial surface pressure main phase transition temperature pretransition temperature

1. Introduction Schizophrenia is a common psychiatric disorder with an incidence of  1% (Kandel, 1991). The clinical symptoms associated with schizophrenia are diverse, and it has been proposed that schizophrenia should be divided into subgroups (Kandel, 1991; Thibaut and Petit, 1997). Many theories have been proposed for the pathogenesis of this disorder, including dysfunction of dopaminergic, serotonergic, and glutaminergic systems, maldevelopment, and membrane dysfunctions (Kandel, 1991; Horrobin et al., 1994). The currently used neuroleptic drugs have been shown to affect the functions of a variety of receptors, such as adrenergic, muscarinic, histamine, 5-hydroxytryptamine, and dopamine receptor families (Bymaster et al., 1999). The effects of the various neuroleptic drugs further depend on the differences between affinities towards the different receptor subtypes. Interestingly, changes in the cell membrane phospholipid compositions in the brain of schizophrenic patients have been related to the onset of clinical symptoms (Pettegrew and Minshew, 1992). Recently, v-3 fatty acid supplemented diet was shown to improve the course of illness in bipolar disorder (Stoll et al., 1999). Altered membrane properties and an effect of this modulation on signal transduction were suggested as the mechanism of action.

Studies on effects by neuroleptic drugs other than those mediated by the neurotransmitter receptors are scarce for clozapine (CLZ) and there are only a few reports available on haloperidol (HPD) as well. This is contrasted by chlorpromazine (CPZ), which is perhaps the most thoroughly investigated neuroleptic agent. CPZ has been reported e.g. to inhibit protein kinase C (Singh et al., 1992), and to alter the activities of enzymes of lipid metabolism (Heiczman and To´ th, 1995). The observed increase in the cellular content of acidic phospholipid and increased unsaturated/saturated lipid ratio induced by CPZ could represent adaptive responses (StuhneSekalec et al., 1987). Hydrophobicity is needed for blood–brain barrier permeation, and CPZ partitions effectively into biological and artificial membranes. CPZ binds to the headgroup region forming a 1:1 complex with acidic phospholipid (Stuhne-Sekalec et al., 1987) and also penetrates into the acyl chain region of phospholipid membranes (Ro¨ mer and Bickel, 1979). Depending on membrane lipid composition and phase state both an increase as well decrease in the acyl chain order in membranes have been reported to be caused by CPZ (Ro¨ mer and Bickel, 1979; Neal et al., 1976). In gel phase phospholipid membranes CPZ induces the formation of fluid domains (Hanpft and Mohr, 1985). Binding of HPD to phospholipid membranes increases disorder more

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

in the interfacial region than in the hydrophobic core of the membrane (Palmeira and Oliveira, 1992). CPZ and CLZ are good antioxidants and decrease membrane lipid peroxidation (Dalla Libera et al., 1998) whereas HPD has been reported to have an opposite effect (Sawas and Gilbert, 1985). These effects might be of importance as lipid peroxidation has been shown to affect the affinity or number of binding sites in membranes for 5-hydroxytryptamine, muscarinic, a-adenergic, and dopamine receptor ligands (Rego and Oliviera, 1995, and references therein). Current views on biomembranes emphasize coupling between their organization and function (Kinnunen, 1991, 2000), which are further strongly connected to the physical properties of the lipid bilayer (Mouritsen and Kinnunen, 1996). Accordingly, changes in the organization and dynamics of the membrane can lead to alterations in the functions of membrane proteins as shown for P-glycoprotein (Romsicki and Sharom, 1999), phospholipase A2 (Burack et al., 1997), and opioid receptor (Lazar and Medzihradsky, 1992), for example. The ligand affinity of the latter has been shown to be sensitive to the changes in the ‘fluidity’ in the interfacial region, but insensitive to changes in the hydrocarbon core (Lazar and Medzihradsky, 1992). A model involving coupling of the membrane lateral pressure profile to the conformation and function of integral membrane proteins has been recently forwarded by Cantor (1997) and could provide a mechanistic basis for the effects of membrane composition on opioid receptor function, for instance. While the above findings demonstrate the importance of the lipid environment to the function of proteins, they also reveal the importance of drug induced changes in membrane organization, dynamics and function, and further suggest that these properties could be considered as potential drug targets (Jutila et al., 1998). Reconstituted dopamine D2-receptor requires a mixture of PC, PE, and PS for restoration of its ligand binding (Srivastava et al., 1987), with PS being particularly important. The depletion of PS from dopamine D2-receptors could thus diminish the ligand affinity. HPD has been reported to reverse PS induced inhibition of phosphatidyli-

153

nositol formation (Bonetti et al., 1985). Strong interaction between brainPS and neuroleptic drugs could detach PS from neurotransmitter receptors, e.g., dopamine D2-receptor, thus leading to altered function of the protein, as shown for the inhibition of cytochrome c oxidase by doxorubicin (Goormaghtigh et al., 1982). We have suggested the possibility that drugs partitioning into the lipid bilayer could exert part of their effects by changing the lateral organization of cellular membranes (Kinnunen, 1991) and have provided evidence to support this notion (Jutila et al., 1998; So¨ derlund et al., 1999a). Accordingly, the fact that neuroleptics are amphiphilic raises the possibility that some of the pharmacological and/or adverse effects could arise from primary drug–lipid interactions which lead to changes in membrane lateral organization and subsequently to altered membrane protein catalyzed processes. These may further be reflected in brain lipid metabolism and functions, the latter thus reaching all the way to individual behavioural patterns. The current study compares the effects of two conventional neuroleptics (CPZ and HPD) and the atypical antipsychotic CLZ on membrane lateral heterogeneity (Fig. 1). 2. Experimental procedures

2.1. Materials Hepes, EDTA, brainPS, eggPC, CPZ, HPD, and CLZ were from Sigma, and 1,2-dipalmitoylsn-glycero-3-phosphocholine (DPPC) from Coatsome (Amagasaki, Hyogo, Japan). 1-Palmitoyl-2-(N-4-nitrobenz-2-oxa-1,3-diazol) × aminocaproylphosphocholine (NBD-PC) was from Avanti Polar Lipids (Alabaster, AL). No impurities were detected in the lipids upon thinlayer chromatography on silicic acid using chloroform/methanol/water/ammonia (65/20/2/2, by vol.) as the solvent system and examination of the plates after iodine staining. To assess the surface chemical purity of DPPC, its compression isotherm was measured on an air–water interface and was observed to be within experimental error ( 3%) of that in literature (Smaby et al., 1994).

154

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

Likewise, the compression isotherm measured for brainPS monolayers complied with those reported in the literature (Demel et al., 1987). Lipids and drugs were dissolved and stored in chloroform. Concentrations of DPPC, eggPC, and brainPS were determined gravimetrically using a high precision electrobalance (Cahn Instruments Inc., Cerritos, CA). Concentration of NBD-PC was determined spectrophotometrically using 21 000 per cm as its molar extinction coefficient at 465 nm. For CLZ 16 800 per cm at 261 nm, for CPZ 24 000 per cm at 254 nm, and for HPD 11 900 per

cm at 247 nm were employed. Deionized water was Millipore filtered (Millipore, Bedford, MA).

2.2. Differential scanning calorimetry The indicated amounts of drugs and lipids were mixed in chloroform. These mixtures were dried with a gentle flow of nitrogen and subsequently kept under reduced pressure for at least 2 h to remove traces of the solvent. The samples were hydrated in 20 mM Hepes, 0.1 mM EDTA, pH 7 buffer at 55 °C for 30 min to obtain multilamellar liposomes utilized in DSC measurements. The samples were equilibrated on an ice-water-bath for at least 10 h before recording the endotherms using VP-DSC microcalorimeter (Microcal Inc., Northampton, MA). Heating rate was 30 deg/h and the final lipid concentration in the DSC cell was 0.4 mM. All scans were repeated to assure their reproducibility. Deviation from the baseline was taken as the beginning of the transition and return to the baseline as its end. The endotherms were analyzed using the routines of the software provided by Microcal.

2.3. Compression isotherms

Fig. 1. The chemical structures of clozapine (CLZ), chlorpromazine (CPZ), and haloperidol (HPD).

Compression isotherms were recorded using 111.1 cm2 (width 55 mm, length 202 mm, subphase volume 22 ml) trough (mTrough S, Kibron Inc., Helsinki, Finland). Surface pressure (y) was monitored with a metal alloy probe hanging from a high precision microbalance (KBN129, Kibron Inc.) connected to a Pentium PC. For compression isotherms and fluorescence microscopy, the indicated mixtures of lipids and neuroleptic drugs were dissolved in a mixture of hexane/isopropanol/water (70/30/2.5, by vol.). These solutions were spread on the air–buffer (20 mM Hepes, 0.1 mM EDTA, pH 7.0) interface. After 5 min equilibration, the film compression was started using two symmetrically moving barriers. Compression rate was in all measurements one A, 2/(acyl chain)/min. Data is represented as y vs. A, 2/(acyl chain), where each lipid molecule consists of two acyl chains. In the calculations, one drug molecule is taken as equivalent to one acyl chain. All monolayer measurements were done at ambi-

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

ent temperature ( + 22 – 23° C). The mean molecular areas occupied by the drugs in the film at any given surface pressure were calculated using the following equation: AD =(AT − AL)/XD,

(1)

where AT is the mean molecular area of the molecules in the presence of the indicated drug, AL is the surface area of the lipids in the absence of the drug, AD is the surface area of the drug, and XD, its mole fraction in the film. The equation applies to the situation with either ideal mixing or complete immiscibility, i.e., not involving molecular interactions, condensing effects, or non-ideal partitioning of the drugs. However, the results allow for qualitative comparison of the drugs, and the assumptions made are further supported by measured surface areas.

2.4. Drug penetration into lipid monolayers Penetration of CLZ, CPZ, and HPD into monomolecular lipid films was measured using magnetically stirred circular wells with a surface area of 1.6 cm2 and a subphase volume of 300 ml (Multiwell plate, Kibron Inc.). Surface pressure was monitored as above when recording compression isotherms. The indicated lipids were mixed in chloroform (0.5 mg/ml) and spread on the air– water interface with a microsyringe. The monolayers were allowed to equilibrate for :5 min to reach the indicated initial surface pressure values (y0). The drugs were injected into subphase (20 mM Hepes, 0.1 mM EDTA, pH 7.0) using a microsyringe (2 ml of 200 mM drug in DMSO) to yield a final drug concentration of 1.3 mM. This amount of DMSO as such had no effect on the surface pressure. The increment in the surface pressure after addition of the drug was complete within two to 20 min (depending on the drug and the lipid composition). The difference between y0 and the final surface pressure after the addition of drug was taken as the increase in surface pressure (Dy). The data are represented as Dy vs. y0, thus revealing the effect of increasing lateral packing density on the penetration of drug into monolayer (Brockman, 1999).

155

2.5. Fluorescence microscopy of lipid monolayers For fluorescence microscopy a Langmuir through equipped with a quartz window on the bottom was placed on the stage of a Zeiss IM-35 inverted fluorescence microscope. Compression isotherms were recorded as described above, with slight modifications. Accordingly, after the desired target pressure was reached by continuous compression by two barriers the film was allowed to stabilize for 10 min before the image was recorded through a Nikon ELWD (20x) objective. The excitation and emission wavelengths were selected with filters transmitting in the range 420– 480 nm and \ 500 nm, respectively. Fluorescence images were viewed with a Peltier-cooled digital camera (Hamamatsu C4742-95, Hamamatsu, Japan) connected to a computer. During the 10 min equilibration time a small decrease in y was observed, reflecting the relaxation of the monolayer. It is to be emphasized that the images obtained are unlikely to represent true equilibrium. Yet, the results should be amenable to comparison as the equilibration times and compression rates were kept identical. The observed domain morphologies were reproducible. In these experiments, the subphase volume was 22 ml and the total amount of lipids and the drugs in the monolayer was 15 nmol. The molar ratio (X= 0.05) of the drug contained in the film would, thus, correspond to a subphase concentration of : 34 nM.

3. Results

3.1. Differential scanning calorimetry We first compared the interactions between a binary phospholipid mixture composed of the zwitterionic DPPC and an acidic phospholipid, brain phophatidylserine (brainPS), and three neuroleptic drugs, CLZ, CPZ, and HPD. Endotherms of DPPC multilamellar vesicles (MLV’s) containing increasing amounts of brainPS are shown in Fig. 2. Neat DPPC has a sharp main transition endotherm peaking at Tm = 41.3 °C and a broad pretransition at Tp = 34.2 °C, with the corre-

156

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

Fig. 2. Endotherms of multilamellar DPPC/brainPS vesicles at XPS (from top to bottom) 0.00, 0.02, 0.05, 0.10, 0.15, and 0.20. The calibration bar corresponds to 1.0 kJ/°C/mol. Total lipid concentration was 0.4 mM in 20 mM Hepes, 0.1 mM EDTA (pH 7). Scanning rate was 30 deg/h.

sponding enthalpies of 32.790.6 and 4.290.4 kJ/mol. Incorporating brainPS decreases both transition temperatures as well as broadens the peaks. At XPS =0.02, only one endotherm centered at 40.4 °C is evident, indicating miscibility. The pretransition is absent at XPS \0.05 and at XPS E0.05, the asymmetry of the main endotherm towards the low temperature side is augmented. This suggests the existence of microdomains composed of DPPC and brainPS with increasing XPS. Tm decreases almost linearly with increasing XPS and at the highest content of PS studied (XPS = 0.20) was at 35.7 °C. A separate small endotherm at 41.2 °C corresponding to the peak for pure DPPC remains in the presence of brainPS (Fig. 2), and suggests that these domains are practically devoid of PS. The above can be rationalized as follows. Electrostatic repulsion between the negatively charged PS headgroups resists the formation of domains enriched in this lipid, with even

distribution representing the free energy minimum. However, with increasing XPS and at XPS E 0.05 this repulsion is partly overcome by an attractive potential, most likely due to hydrogen bonding between the PS headgroups (Boggs, 1987), similarly to that suggested for another acidic phospholipid, phosphatidylglycerol (Subramanian et al., 1998). Enthalpy (DH) of the main transition is diminished to :13 kJ/mol at XPS \ 0.10 demonstrating the extent of trans“ gauche isomerization in the transition to be significantly reduced and suggesting chain disorder to be augmented in both gel state and fluid membranes. At these molar ratios also the small peak at 41.2 °C becomes distinct from the main peak and its enthalpy is not included in the reported value for the main endotherm. The main purpose of this study was to compare the atypical neuroleptic, CLZ, with two conventional neuroleptics, CPZ and HPD, in their interactions with phospholipid membranes. Both CPZ and HPD have been reported to partition into DPPC liposomes (Hanpft and Mohr, 1985; Sarmento et al., 1993). In order to investigate the contribution from electrostatic and hydrophobic interactions to the binding of CLZ to the bilayer, we first studied the effect of increasing contents of this drug on the thermal phase behavior of DPPC liposomes (data not shown). At the highest content used XCLZ = 0.2, the value for Tp was lowered from 34.3 to 28.7 °C and Tm from 41.3 to 40.5 °C. The respective DH values diminished from 4.29 0.4 to 0.99 0.3 kJ/mol and from 32.79 0.6 to 20.99 1.1 kJ/mol. Heat capacity scans thus revealed CLZ to bind to DPPC liposomes. As the latter is zwitterionic, the interaction is likely to be driven by hydrophobicity. In the next series of experiments DPPC/brainPS (XPS = 0.05) MLVs were used to explore the impact of the acidic PS to drug–lipid interactions. Interestingly, at X= 0.05 CLZ abolished the peak at  41.2 °C (Fig. 3), while in the presence of CPZ and HPD phase separation was observed up to the highest concentration (X=0.10) of the drug studied (data not shown). The latter could be related to charge neutralization of brainPS by the latter two compounds and enrichment of these drugs into the brainPS/DPPC domains. The dif-

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

157

ferent effect of CLZ suggests the hydrophobic interactions of this compound to be more important for membrane association than for CPZ and HPD. Moreover, our data indicate CLZ to partition into the higher melting DPPC enriched domains. The effects of these drugs on the temperature and enthalpy of the pretransition of DPPC/ brainPS MLVs are compiled in Fig. 4. All three drugs decreased gradually both Tp and DHp of pretransition, the effects being least for CLZ. At XCLZ \0.03, and at XCPZ and XHPD \0.02 pretransition was not observed. The effects of the three neuroleptics on Tm and DHm of the main transition of DPPC/brainPS MLV’s (XPS =0.05) are depicted in Fig. 5. With increasing drug concentrations values for Tm shift gradually from 40.2 °C towards lower temperatures, being centered at 39.2, 38.9, and 38.5 °C for CLZ, CPZ, and HPD, respectively, at Xdrug =0.10. Enthalpy of the main transition is increased up to Xdrug =

Fig. 4. Effects of increasing CLZ ( ), CPZ () and HPD () content on temperature (Panel A) and enthalpy (Panel B) of the pretransition of DPPC/brainPS MLV’s at XPS =0.05. Conditions were as in Fig. 2.

Fig. 3. Endotherms of DPPC, DPPC/brainPS (XPS = 0.05), and DPPC/brainPS/drug (XPS = Xdrug = 0.05) MLVs. Conditions were as in Fig. 2.

0.02 for all three drugs, this enhancement being largest (: 16%) for HPD. At higher Xdrug, the effects become dissimilar, as follows. For CLZ, the enthalpy remains approximately equal to that measured in the absence of the drug, the only exception being XCLZ = 0.05, i.e., when the acidic phospholipid: CLZ stoichiometry is 1:1, where DHm has a minimum of 26.5 kJ/mol (Fig. 5(B)).

158

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

At XCLZ =0.05, there is also a change in the dependence of Tm on Xdrug (Fig. 5(A)). For CPZ, further increase in XCPZ has only a minor effect on variation in DHm, with maximal increase of 4.4 kJ/mol ( 15%). For HPD, at X \ 0.02 DHm decreases almost linearly reaching a minimum of 30.3 kJ/mol at XHPD =0.10.

Fig. 6. Compression isotherms of DPPC/brainPS/drug monolayers at XPS =Xdrug =0.05 for CLZ ( ), CPZ (), and HPD (). The line without symbols corresponds to the DPPC/ brainPS (XPS =0.05) monolayer. The inset shows an enlargement of the surface pressure range in LE – LC coexistence region. Subphase was 20 mM Hepes, 0.1 mM EDTA, pH 7. Monolayers were compressed continuously at a rate of 1 A, 2/(acyl chain)/min.

3.2. Compression isotherms

Fig. 5. Effects of increasing CLZ ( ), CPZ () and HPD () content on temperature (Panel A) and enthalpy (Panel B) of the main transition of DPPC/brainPS MLV’s at XPS = 0.05. Conditions were as in Fig. 2.

Phospholipid monolayers provide a highly informative approach for studying drug–lipid interactions as the lipid composition has no effect on the surface curvature and the lateral packing can be precisely controlled (Brockman, 1999). Due to the high affinity for PS of the compounds used in the present study, we may assume their degree of dissociation from the lipid monolayers into the subphase to be insignificant. Compression isotherms of DPPC/brainPS (at XPS = 0.05) monolayer in the absence and in the presence of the three drugs at Xdrug = 0.05 are depicted in Fig. 6. The inset shows a view of the surface pressure range for the liquid-condensed region (for a recent review, see Kaganer et al., 1999). All three drugs increased the area/molecule. At 20 mN/m and in the absence of the drugs this value was 22.3 A, 2, and was increased up to 25.3, 24.1, and 23.4 A, 2 by CLZ, CPZ, and HPD, respectively. The approximate constant increase in surface area at higher values of y (15–35 mN/m) caused by the drugs is in keeping with their tight and pressure dependent association to the lipid films. Assuming

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

the partitioning of the drugs into the subphase to be negligible, we can calculate the mean molecular areas to be 114, 68, and 42 A, 2 for CLZ, CPZ, and HPD, respectively. Interestingly, in the liquid expanded-liquid condensed coexistence region (at y  8 – 9 mN/m) the isotherms for the drug containing monolayers were almost superimposable and the mean molecular areas were : 137 A, 2 for CLZ and CPZ, and 114 A, 2 for HPD. The values obtained from the molecular modelling are in qualitative agreement with the experimental results, and also support the conclusion that all three drugs are quantitatively bound to the monolayer (Pakkanen et al., unpublished data).

3.3. Binding of drugs to lipid monolayers In order to get more insight into the interaction of the drugs with phospholipids, we measured also their association with monolayers residing on an air–water interface. More specifically, we monitored the increase in surface pressure (Dy) due to their penetration into monolayers (eggPC and brainPS) at different initial surface pressures (y0), varied in the range of : 12–35 mN/m. In these studies, eggPC was used instead of DPPC in order to avoid interference due to phase transitions. For pure PC, the partitioning of the drugs is driven by hydrophobicity, and the most efficient penetration into eggPC monolayers was evident for CLZ which under the conditions used induced Dy of about 3 mN/m at y0 =10 – 25 mN/m. The other two drugs interacted more weakly with eggPC films, with Dy being 1 –2 mN/m. Penetration of the drugs into eggPC decreased with increasing y due to augmented lipid lateral packing. Critical surface pressures abolishing the penetration of CLZ, CPZ, and HPD were 40, 32, and 25 mN/m, revealing the atypical neuroleptic CLZ to be the most membrane active of these compounds. These results support the conclusion drawn from DSC and compression isotherms that for CLZ hydrophobic interactions are more important for its membrane association than for

159

CPZ and HPD. The effects of the drugs on brainPS monolayers were more pronounced than for eggPC (Fig. 7). At y0 = 12–25 mN/m, Dy for brainPS monolayer was : 5, 11, and 8 mN/ m by CLZ, CPZ, and HPD, respectively. At y0  26 mN, an abrupt decrement in Dy was observed. This could reflect augmented protonation of the PS headgroups at higher packing densities with subsequent hydrogen bonding resulting in enhanced PS–PS interactions (Boggs, 1987), and thus impeding the penetration of the drugs driven by electrostatic attraction.

3.4. Fluorescence microscopy Fluorescence microscopy images of DPPC/ brain PS/NBD-PC (XPS = 0.05, XNBD – PC = 0.02) monolayer are characterized by the presence of domains at surface pressure values above 12 mN/m (Fig. 8), revealing the two phase region of the monolayer. The latter is also evident as a discontinuity in the compression isotherms (Fig. 6). The differential ‘staining’ of the coexisting gel state and fluid domains is due to the efficient

Fig. 7. Increase in the surface pressure (Dy) of brainPS (solid symbols) and eggPC (open symbols) monolayers at different initial surface pressures (y0) upon injection of CLZ ( , ), CPZ ( , ), or HPD (, ) into the subphase. The final concentration of the drugs was 1.3 mM. Subphase was 20 mM Hepes, 0.1 mM EDTA, pH 7.

160

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

4. Discussion

Fig. 8. Fluorescence microscope images of DPPC/brainPS/ NBD-PC (XPS = 0.05, XNBD – PC = 0.02) monolayers at 15, 20, and 25 mN/m (from left to right) after 10 min stabilization in the absence of drugs (uppermost panels) and at Xdrug = 0.05 for CLZ, CPZ, and HPD (from top to bottom). Subphase was 20 mM Hepes, 0.1 mM EDTA, pH 7. For more details see Materials and Methods.

partitioning of the fluorescent probe NBD-PC into the fluid (i.e. liquid expanded) domains (Weis and McConnell, 1985). In the absence of the drugs, the domains are relatively large with a distinct roundish shape. The size of the domains is affected by all three drugs (Xdrug =0.05) and the domain morphologies become more complex in the presence of CPZ and HPD. Instead, only slight changes in domain morphology are caused by CLZ, while the average size of gel state domains decreases. Both of these effects increase the length of the boundary between gel and fluid domains, indicating that the drugs stabilize the boundary and thus enhance phase separation. This also suggests preferential partitioning of these drugs into the domain boundaries.

Understanding of drug–lipid interactions is important for a number of reasons. First, lipid membranes provide the major barrier against the passive diffusion of drugs into the intestinal cells and into specific tissues, such as the blood-brain barrier. Elucidation of the mechanisms affecting the passive diffusion of compounds through the lipid bilayer are thus of primary importance in drug development. Second, the overexpression of the P-glycoprotein is one of the major causes for multidrug resistance in human cancers. These transporters are integral membrane proteins, and the interaction between the protein and ligand requires the latter to be located in the membrane (Romsicki and Sharom, 1999). Third, binding of a drug to lipids can lead to alterations in the function of membrane proteins. This could be involved in the actual mechanism of action or in the adverse effects of drugs, as proposed for the lung toxicity of amiodarone (see review by Reasor and Kacew, 1996) and the cardiotoxicity of doxorubicin (Goormaghtigh et al., 1982). Displacement of peripheral membrane proteins by competition with drugs for the liganding lipids is also possible, as exemplified for cytochrome c (Jutila et al., 1998). Fourth, understanding of drug-lipid interactions is crucial in the design of liposomes for use as drug carriers. Finally, the lipid membrane could represent the actual target for the drug (Kinnunen, 1991; So¨ derlund et al., 1999a,b), as shown amphotericin B (Bolard, 1986), and antimicrobial peptides (Bechinger, 1997). Changes in the membrane lateral heterogeneity are caused by all three neuroleptic drugs of this study, as demonstrated by DSC, and fluorescence microscopy. Importantly, the effects of these drugs are not identical, indicating that the observed changes are not caused by non-specific interactions between these compounds and the lipids. It should be emphasized that drug induced changes in lipid membrane lateral heterogeneity are readily expected to be strongly dependent on membrane lipid composition, the compound itself, and factors such as temperature, pH, osmolarity, and ionic strength (Kinnunen, 1991). We have recently demonstrated cyclosporin A to cause

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

lipid composition dependent changes in lipid domain morphology (So¨ derlund et al., 1999a). Additionally, we have shown also the association of cationic drug, doxorubicin, to depend on the membrane lateral organization (So¨ derlund et al., 1999b). In the transition region, the coexisting phases undergo intense fluctuations, as recently shown by atomic force microscopy of phospholipid monolayers deposited on mica sheets (Nielsen et al., 2000). Domain morphology is determined mainly by the balance between line tension and dipole repulsion (Kaganer et al., 1999; McConnell, 1991; Brockman, 1994). In brief, line tension tends to make domains compact and circular, with minimal domain boundary length, while the opposite is true for dipole repulsion. Based on these mechanisms, it can be concluded that all these drugs decreased the line tension and/or increased dipole repulsion in the order CPZ \HPD \CLZ. In the presence of HPD, the domains appear already at y : 9 mN/m, at a pressure where the other two drugs have no effect. At y =25 mN/m the domains become more fuzzy and their boundaries more diffuse in the presence of the drugs, in keeping with the drug induced broadening of the enthalpy peaks in DSC. CPZ, CLZ, and HPD all increased the gel-fluid domain boundary length. The increment in boundary length was larger for the conventional neuroleptics, CPZ and HPD, than for the atypical neuroleptic, CLZ. Similar difference between CLZ and the two conventional neuroleptics was also observed in DSC measurements. Accordingly, CLZ abolished the phase separation in DPPC/brainPS MLVs, while in the presence of CPZ or HPD phase separation was retained. These findings indicate a striking difference in the effects of CLZ and the conventional neuroleptics (CPZ and HPD) on the heterogeneity of binary lipid mixture. To this end, an increase in domain boundary length has been shown to increase membrane permeability (for a review see Mouritsen and Kinnunen, 1996), which could provide a mechanistic explanation for the observed increase in liposome permeability by CPZ (Maoi et al., 1979). DSC is not as sensitive as fluorescence microscopy and requires the use of higher concentra-

161

tions of both lipids and drugs. However, to emphasize the potential pharmacological relevance of the present study, we want to point out that the concentrations of the drugs (: 34 nM) required to induce the described effects on the domain morphologies are within the range of their therapeutic plasma concentrations (Dollery, 1991a,b; Spina et al., 2000). Likewise, while varying for different receptors and their subtypes the dissociation constants for these neuroleptic drugs are in the range of 0.1–10 mM (Brody et al., 1998). Electrostatic attraction seems to be important for the lipid binding of all three drugs, whereas for CLZ also hydrophobic interaction is contributing. Accordingly, in keeping with DSC studies using DPPC MLVs CLZ penetrated into eggPC monolayers at significantly higher values of y0 than CPZ or HPD and produced the largest Dy. Likewise, the lateral expansion of DPPC/ brainPS (XPS = 0.05) monolayers was largest for CLZ. The abrupt decrease in Dy in the penetration of all three drugs into brainPS monolayers at  25–26 mN/m could represent an attenuated affinity of the drugs towards the lipid monolayer as a consequence of diminished electrostatic attraction due to enhanced protonation of PS headgroups at higher surface charge densities. The differences in the contribution of hydrophobic and electrostatic interactions on the membrane association of CLZ, CPZ, and HPD are likely to be contributing also to the observed changes in membrane heterogeneity caused by these compounds. The calculated mean molecular areas of these three compounds are similar ( 114–137 A, 2/molecule) at low surface pressure ( 9 mN/ m). Instead, at higher packing pressures (y= 15– 35 mN/m) the mean molecular area occupied by CLZ remains high, 114 A, 2 whereas for CPZ and HPD a decrement to 68 and 42 A, 2, respectively, is evident. This suggests CLZ to be similarly oriented in the membrane at both low and high lateral packing pressures, while CPZ and HPD may reorientate and/or become excluded from the membrane interior at higher lateral pressures. The latter possibility is indicated by the monolayer penetration measurements. Yet, the derived mean molecular areas as well as the compression

162

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163

isotherms suggest that in the surface pressure ranges generally considered to be relevant for biomembranes ( :30 – 35 mN/m) the degree of dissociation of the drugs from the monolayer should be negligible. The similar mean molecular area of 114–137 A, 2 for the three compounds in the transition region is interesting. Notably, as the compressibility of the film has a maximum in the two phase region, this suggests that under these conditions the orientation of the compounds is solely controlled by the lateral pressure. The current results on the interactions between neuroleptic drugs and phospholipid model membranes composed of binary lipid mixtures reveal qualitative differencies between the two conventional neuroleptics, (CPZ and HPD) and atypical neuroleptic, CLZ. Accordingly, the two former compounds have a significant impact on membrane lateral heterogeneity, causing more pronounced rearreangements of the membrane constituents than the atypical neuroleptic CLZ. In the light of the above, it is tempting to suggest that the observed effects of CLZ, CPZ, and HPD on membrane lateral heterogeneity and PS-drug interactions could represent additional mechanisms by which these compounds may exert (part of) their pharmacological activities. Yet, more extensive comparison between atypical and conventional neuroleptics is required. Our findings also suggest possible use of membrane lateral heterogeneity for screening purposes in drug discovery.

Acknowledgements This work was supported by Academy of Finland and TEKES (P.K.J.K.). The authors wish to thank Ms. Outi Tamminen for skillfull assistance in sample preparation and DSC measurements.

References Bechinger, B., 1997. Structure and functions of channel-forming peptides: magainins, cecropins, mellitin and alamethicin. J. Membr. Biol. 156, 197 –211.

Boggs, J.M., 1987. Lipid intermolecular hydrogen bonding: influence on structural organization and membrane function. Biochim. Biophys. Acta 906, 353 – 404. Bolard, J., 1986. How do the polyene macrolide antibiotics affect the cellular membrane properties? Biochim. Biophys. Acta 864, 257 – 304. Bonetti, A.C., Canonico, P.L., MacLeod, R.M., 1985. Brain cortex phosphatidylserine inhibits phosphatidylinositol turnover in rat anterior pituitary glands. Proc. Soc. Exp. Biol. Med. 180, 79 – 83. Brockman, H., 1994. Dipole potential of lipid membranes. Chem. Phys. Lipids 73, 57 – 79. Brockman, H., 1999. Lipid monolayers: why use half a membrane to characterize protein – membrane interactions? Curr. Opin. Struct. Biol. 9, 425 – 427. Brody, T.M., Lasner, J., Minneman, K.P., 1998. Human Pharmacology Molecular to Clinical, 3rd ed. Mosby, St Louis, MO, p. 341. Burack, W.R., Dibble, A.R.G., Allietta, M.M., Biltonen, R.L., 1997. Changes in vesicle morphology induced by lateral phase separation modulate phospholipase A2 activity. Biochemistry 36, 10551 – 10557. Bymaster, F.P., Nelson, D.L., DeLapp, N.W., Falcone, J.F., Eckols, K., Truex, L.L., Foremann, M.M., Lucaites, V.L., Calligaro, D.O., 1999. Antagonism by olanzapine of dopamine D1, serotonin2, muscarinic, histamine H1 and a1-adrenergic receptors in vitro. Schizophrenia Res. 37, 107 – 122. Cantor, R.S., 1997. Lateral pressures in cell membranes: a mechanism for modulation of protein function. J. Phys. Chem. B 101, 1723 – 1725. Dalla Libera, A., Scutari, G., Boscolo, R., Rigobello, M.P., Bindoli, A., 1998. Antioxidant properties of clozapine and related neuroleptics. Free Radical Res. 29, 151 – 157. Demel, R.A., Paltauf, F., Hauser, H., 1987. Monolayer characteristics and thermal behaviour of natural and synthetic phosphatidylserines. Biochemistry 26, 8659 – 8665. Dollery, C., 1991a. Therapeutic drugs, vol. I. Churchill Livingstone, New York, pp. C201 – C206. Dollery, C., 1991b. Therapeutic drugs, vol. I. Churchill Livingstone, New York, pp. H1 –H4. Goormaghtigh, E., Brasseur, R., Ruysschaert, J.M., 1982. Adriamycin inactivates cytochrome c oxidase by exclusion of the enzyme from its cardiolipin essential environment. Biochem. Biophys. Res. Comm. 104, 314 – 320. Hanpft, R., Mohr, K., 1985. Influence of cationic amphiphilic drugs on the phase-transition temperature of phospholipids with different polar headgroups. Biochim. Biophys. Acta 814, 156 – 162. Heiczman, A., To´ th, M., 1995. Effect of chlorpromazine on the synthesis of neutral lipids and phospholipids from [3H]glycerol in the primordial human placenta. Placenta 16, 347 – 358. Horrobin, D.F., Glen, A.I., Vaddadi, K., 1994. The membrane hypothesis in schizophrenia. Schizophrenia Res. 13, 195 – 207.

A. Jutila et al. / Chemistry and Physics of Lipids 112 (2001) 151–163 Jutila, A., Ryto¨ maa, M., Kinnunen, P.K.J., 1998. Detachment of cytochrome c by cationic drugs from membranes containing acidic phospholipids: comparison of lidocaine, propranolol, and gentamycin. Mol. Pharm. 54, 722 –732. Kaganer, V.M., Mo¨ hwald, H., Dutta, P., 1999. Structure and phase transitions in Langmuir monolayers. Rev. Modern Phys. 71, 779 – 819. Kandel, E.R., 1991. Disorders of thought: schizophrenia. In: Kandel, E.R., Schwartz, J.H., Jessel, T.M. (Eds.), Principles of Neural Science. Appleton and Lange, East Norwalk, CT, pp. 853– 868. Kinnunen, P.K.J., 1991. On the principles of functional ordering in biological membranes. Chem. Phys. Lipids 57, 375 – 399. Kinnunen, P.K.J., 2000. Lipid bilayers as osmotic response elements. Cell. Physiol. Biochem. 10, 243 – 250. Lazar, D.F., Medzihradsky, F., 1992. Altered microviscosity at brain membrane surface induces distinct and reversible inhibition of opioid receptor binding. J. Neurochem. 59, 1233 – 1240. Maoi, M., Suzuki, T., Yagi, K., 1979. Effects of chlorpromazine and other phenothiazine derivatives on the permeability of liposomes. Biochem. Phar. 28, 295 –299. McConnell, H.M., 1991. Structures and transitions in lipid monolayers at the air – water interface. Annu. Rev. Phys. Chem. 42, 171 – 195. Mouritsen, O.G., Kinnunen, P.K.J., 1996. Role of lipid organization and dynamics for membrane functionality. In: Merz, K. Jr., Roux, B. (Eds.), Biological Membranes. Birkha¨ user, Boston, pp. 463 – 502. Neal, M.J., Butler, K.W., Polnaszek, C.F., Smith, I.C.P., 1976. The influence of anesthetics and cholesterol on the degree of molecular organization and mobility of ox brain white matter. Lipids in multibilayer membranes: a spin probe study using spectral simulation ny the stochastic method. Mol. Pharm. 12, 144 – 155. Nielsen, L.K., Bjo¨ rnholm, T., Mouritsen, O.G., 2000. Critical phenomena: fluctuations caught in the act. Nature 404, 352. Palmeira, C.M., Oliveira, C.R., 1992. Partitioning and membrane disordering effects of dopamine antagonists: influence of lipid peroxidation, temperature, and drug concentration. Arch. Biochem. Biophys. 295, 161 –171. Pettegrew, J.W., Minshew, N.J., 1992. Molecular insights into schizophrenia. J. Neural Transmiss. Supp. 36, 23 –40. Reasor, M.J., Kacew, S., 1996. An evaluation of possible mechanisms underlying amiodarone-induced pulmonary toxicity. Proc. Soc. Exp. Biol. Med. 212, 297 –304. Rego, A.C., Oliviera, C.R., 1995. Influence of lipid peroxidation on [3H]ketanserin binding to 5-HT2 prefrontal cortex receptors. Neurochem. Int. 27, 489 –496. Romsicki, Y., Sharom, F.J., 1999. The membrane lipid environment modulates drug interactions with the p-glycoprotein multidrug transporter. Biochemistry 38, 6887 –6896.

163

Ro¨ mer, J., Bickel, M.H., 1979. Interactions of chlorpromazine and imipramine with artificial membranes investigated by equilibrium dialysis, dual-wavelength photometry, and fluorimetry. Biochem. Pharm. 28, 799 – 805. Sarmento, A.B., de Lima, M.C., Oliviera, C.R., 1993. Partition of dopamine antagonists into synthetic lipid bilayers: the effect of membrane structure and composition. J. Pharm. Pharmacol. 45, 601 – 605. Sawas, A.H., Gilbert, J.C., 1985. Lipid peroxidation as a possible mechanism for the neurotoxic and nephrotoxic effects of a combination of lithium carbonate and haloperidol. Arch. Int. Pharmacody Ther. 276, 301 – 312. Singh, I.N., Massarelli, R., Kanfer, J.N., 1992. Modulation of phosphatidylserine homeostasis by amphiphilic cations in a human neuronal cell line, LA-N-2. J. Lipid Mediators 5, 301 – 311. Smaby, J.N., Brockman, H.L., Brown, R.E., 1994. Cholesterol’s interfacial interaction with sphingomyelins and phosphatidylcholines: hydrocarbon chain structure determines the magnitude of condensation. Biochemistry 33, 9135 – 9142. Spina, E., Avenoso, A., Facciola, G., Scordo, M.G., Ancione, M., Madia, A.G., Ventimiglia, A., Perucca, E., 2000. Relationship between plasma concentrations of clozapine and norclozapine and therapeutic response in patients with schizophrenia resistant to conventional neuroleptics. Psychopharmacology 148, 83 – 89. Srivastava, L.K., Kazmi, S.M.I., Blume, A.J., Mishra, R.K., 1987. Reconstitution of affinity-purified dopamine D2 receptor binding activities by specific lipids. Biochim. Biophys. Acta 900, 175 – 182. Subramanian, M., Jutila, A., Kinnunen, P.K.J., 1998. Binding and dissociation of cytochrome c to and from membranes containing acidic phospholipids. Biochemistry 37, 1394 – 1402. Stoll, A.L., Severus, W.E., Freeman, M.P., Reuter, S., Zboyan, H.A., Cress, K.K., Marangell, L.B., 1999. Omega 3 fatty acids in bipolar disorder. A preliminary doubleblinded, placebo-controlled trial. Arch. Gen. Psychiatry 56, 407 – 412. Stuhne-Sekalec, L., Chudzik, J., Stanacev, N.Z., 1987. Effect of chlorpromazine on the synthesis, hydrolysis, and transfer of microsomal cytidine liponucelotides and mitochondrial polyglycerophosphatides. Can. J. Physiol. Pharm. 65, 377 – 384. So¨ derlund, T., Lehtonen, J.Y.A., Kinnunen, P.K.J., 1999a. Interactions of cyclosporin A with phospholipid membranes: effect of cholesterol. Mol. Pharm. 58, 32 – 38. So¨ derlund, T., Jutila, A., Kinnunen, P.K.J., 1999b. Binding of adriamycin to liposomes as a probe for membrane lateral organization. Biophys. J. 76, 896 – 907. Thibaut, F., Petit, M., 1997. The deficit syndrome of schizophrenia: towards heterogeneity. Psychopathology 30, 257 – 262. Weis, R.M., McConnell, H.M., 1985. Cholesterol stabilizes the crystal – liquid interface in phospholipid monolayers. J. Phys. Chem. 89, 4453 – 4459.