Antidepressant-induced lipidosis with special reference to tricyclic compounds

Progress in Neurobiology 60 (2000) 501±512

www.elsevier.com/locate/pneurobio

Antidepressant-induced lipidosis with special reference to tricyclic compounds Zhenlei Xia, Gu Ying, Ann Louise Hansson, HaÊkan Karlsson, Yi Xie, Anders Bergstrand, Joseph W. DePierre, Lennart NaÈssberger* Unit for Biochemical Toxicology, Department of Biochemistry, Wallenberg Laboratory, Stockholm University, S-106 91, Stockholm, Sweden Received 3 June 1999

Abstract Cationic amphiphilic drugs, in general, induce phospholipid disturbances. Tricyclic, as well as other antidepressants belong to this group. In experimental animals, antidepressants induce lipid storage disorders in cells of most organs, a so-called generalized phospholipidosis. This disorder is conveniently detected by electron microscopic examination revealing myelin ®gures. Myelin ®gures or myeloid bodies are subcellular organelles containing unicentric lamellar layers. The lipidotic induction potency during in vivo is related to the apolarity of the compound. Metabolism of phospholipids takes place within the cell continuously. Several underlying mechanisms may be responsible for the induction of the phospholipid disturbance. For instance, it has been suggested that the compounds bind to phospholipids and such binding may alter the phospholipid's suitability as a substrate for phospholipases. Free TCA or metabolites thereof may also inhibit phospholipases directly, as has been demonstrated for sphingomyelinase in glioma and neuroblastoma cells. Both these mechanisms might result in phospholipidosis. Interaction between drug and phospholipid bilayer has been investigated by nuclear magnetic resonance technique. There seems to be large di€erences in the sensitivities amongst di€erent organs. Steroid-producing cells of the adrenal cortex, testis and ovaries are in particular susceptible to drug-induced lipidosis. The so-called foam cells are lung macrophages located in the interstitium which become densely packed with myelin ®gures during TCA exposure. It requires about 3±6 weeks of treatment to develop this converted cell. In cell cultures however, phospholipidosis is demonstrated already after 24 h only. It appears that the cells that undergo TCA-induced lipidosis may recover after withdrawal of the drug. The time required to achieve complete recovery ranges from 3±4 weeks to several months, depending on the organ a€ected. Little is known about the functional signi®cance of lipidosis. Even if TCA and other antidepressants show other e€ects, it has not been possible to exclusively relate it to phospholipidosis. However, few attempts have been made to correlate the physiological e€ects of TCAs in experimental animals to the morphological changes associated with phospholipidosis. There is an increasing evidence however, that cationic amphiphilic drugs may have e€ects on immune function, signal transduction and receptor-mediated events, e€ects that to some extent might be related to disturbances in phospholipid metabolism. # 2000 Elsevier Science Ltd. All rights reserved.

Contents 1.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502

2.

Phospholipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 2.1. Phospholipidosis lipidosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502

Abbreviations: CAD, Cationic Amphilic Drugs; TCA, Tricyclic Antidepressant Drugs; SSRI, Selective Serotonin Release Inhibitor; NMR, Nuclear Magnetic Resonance; IC50, Inhibitory Concentration (concentration that causes a 50% inhibition). * Corresponding author. Fax: +46-8-153024. 0301-0082/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 1 - 0 0 8 2 ( 9 9 ) 0 0 0 3 6 - 2

502

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

3.

Tricyclic antidepressants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Other antidepressants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Underlying mechanisms of TCA-induced lipidosis . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Inhibition of phospholipases by TCAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Organ distribution of TCA-induced phospholipidosis in experimental animal models

. . . . .

504 504 504 505 506

4.

Phospholipidosis in cell cultures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507

5.

Reversibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507

6.

Functional implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507

7.

TCA-induced lipidosis in human peripheral blood cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . 509

8.

Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511

1. Introduction The ®rst report on drug-induced phospholipidosis appeared in 1966, when Greselin described induction of an increased number of foam cells in the rat lung (Greselin, 1966). Since then, many reports on druginduced phospholipid storage disorders have appeared. This phenomenon has been related speci®cally to a typical group of so-called cationic amphiphilic drugs. To date, more than 50 cationic amphiphilic drugs (CADs) have been reported to induce this type of lipid storage disorder. The tricyclic antidepressants (TCAs), which are widely used in patients presenting depressive diseases, belong to the CAD group, as do many neuroleptic compounds. It has been found that TCAs are potent inducers of lipidosis in experimental models, as described in the early 1970s by the Kiel group (Lullmann-Rauch et al., 1973). The tricyclic antidepressive drugs interact with phospholipids and cause a generalized lipidosis in experimental animals. Similarly, the new generation of antidepressive drugs, the so-called SSRI compounds also induce phospholipidosis.

2. Phospholipids Phospholipids are important and dynamic components of the plasma and intracellular membranes. The phospholipids are divided into various classes. The fatty acid composition of each class of phospholipids di€ers widely from one tissue to another within a given species. In contrast, the corresponding fatty acid composition in a given organ or tissue in one species is

remarkably similar to that of the same tissue in a di€erent species. This fact could be important to keep in mind when evaluating results from experimental animal models. The characteristics of a membrane are to a large degree dependent on the proportions of cholesterol, galactolipids and phospholipids and in particular, of the fatty acid compositions and the chain length of the respective phospholipids. For instance, a high proportion of unsaturated fatty acids renders the membrane to increase its ¯uidity. This property may be responsible to the di€erences in the structure of membrane layers seen in myeloid bodies.

2.1. Phospholipidosis lipidosis The term phospholipidosis or lipidosis describes intracellular accumulation of various phospholipids, re¯ecting a disorder in phospholipid storage. Usually, there are increased levels of several classes of phospholipid, although there may be a predominant accumulation of a single phospholipid. The most easily observed indicator of lipidosis is the occurrence of a particular kind of subcellular organelles called myeloid bodies or myelin ®gures. Most likely they originate from primary lysosomes. At electron microscopic examination, two main morphological patterns are observed. One of these involves structures containing unicentric lamellar layers (myelin ®gures), usually with a distance of 45 AÊ between the layers. The other type of pattern is crystalline-like or crystalloid inclusions with a hexagonal pattern. This pattern can be adopted spontaneously by phosphatidyl-ethanolamine. In the presence of fusogenic com-

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

Fig. 1. The morphology of a myelin ®gure (typical membrane layers and whorls are clearly illustrated Ð 54,000).

503

504

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

Fig. 2. (A) Imipramine, an example of a tricyclic antidepressive compound. (B) Zimelidine, an example of a bicyclic antidepressive compound. (C) Maprotiline, an example of a tetracyclic antidepressive compound. (D) Citalopram, a so-called SSRI compound with a typical tricyclic structure.

pounds e.g. retinol, phosphatidylcholine also forms hexagonal assemblies (Howell et al., 1973). A third type of structure observed in connection with phospholipidosis consists of lamellar membranes with areas of amorphous material in between. Fusion of smaller myeloid bodies results in large multicentric myeloid bodies, so-called cytosegresomes. In these structures, abundant clear matrix is present. The formation of such vacuolated myeloid bodies has been attributed to osmotic swelling. These organelles demonstrate an increased matrix pH which is due to a heavy drug accumulation. Fig. 1 illustrates typical myelin ®gures composed of lamellar layers. This is the most common type of myelin ®gure observed in connection with treatment of experimental animals with TCAs. 3. Tricyclic antidepressants Tricyclic antidepressants contain three 5- or 6-membered rings, forming a bulky lipophilic structure, and a hydrophilic side-chain consisting of a short carbon chain with a positively charged amino group at its end. The following TCAs imipramine, clomipramine, amitryptyline, iprindole, triparanol, doxepin, trimipramine and noxiptylin, as well as metabolites thereof, i.e. desmethylimipramine have been thoroughly investigated in experimental animal models. All of these TCAs induce a generalized lipidosis, i.e. phospholipid accumulation in most of the organs of the experimen-

tal animal. Fig. 2A illustrates the molecular structure of imipramine, as a representative for the whole group of TCA compounds. 3.1. Other antidepressants During the past decade, antidepressants with molecular structures other than the classic tricyclic one have been introduced continuously into clinical practice. Several of these have been investigated experimentally and found to induce lipidosis. For instance, Zimelidine, a bicyclic molecular compound (Fig. 2B), but still a potent inducer of phospholipidos, as demonstrated by Lullmann-Rauch (Bockhardt and LullmannRauch, 1980). This compound was, however, withdrawn from the market due to serious neurological side-e€ects. Maprotiline is a tetracycline compound (Fig. 2C), and found to be less cardiotoxic compared to the TCA-group. The newer generation of antidepressants, the socalled SSRI (Selective Serotonin Release Inhibitors) compounds generally have molecular structures which di€er completely from those of the classic TCA drugs. Citalopram, however, has the most look alike molecular structure compared to the old TCAs (Fig. 2D), and has been found to induce a generalized lipidosis in rats (Lullmann-Rauch and NaÈssberger, 1983). 3.2. Underlying mechanisms of TCA-induced lipidosis In general, biological activities of drug molecules often involve interactions with biomembranes. Studies

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

Fig. 3. Accumulation of phospholipids in human monocytes during imipramine treatment. Monocytes were incubated with 25 mM imipramine up to 120 h. Accumulation of phospholipids was determined by measuring, (32 P) phosphate-incorporation lipid-soluble material compared to untreated monocytes.

on the underlying mechanisms of TCA-induced phospholipidosis have been carried out. Before reaching their target site(s) TCA must cross lipid membrane barriers. Therefore, the interactions between TCAs and phospholipids and phospholipid containing membranes play critical roles in their pharmacokinetics. In general, cationic amphiphiles have been demonstrated by cell electrophoresis to interact with cellular membranes (Ashman et al., 1986). Because TCAs contain lipophilic as well as hydrophilic moieties, their interactions with lipid membranes tend to be complex. Overall, it can be said that the relationship between induction of phospholipidosis and inhibition of phospholipases and/or drugbinding to phospholipids remains mechanistically puzzling. The proposed mechanism underlying TCA-induced lipidosis is complex and involves several steps. Firstly, the drug may undergo metabolism. If metabolism results in a high degree of polar metabolites, these are cleared from the body. On the other hand, non-polar metabolites are taken up and entrapped within lysosomes. If the TCA binds to phospholipids, such binding would most likely alter the phospholipid's suitability as a substrate for phospholipases. This might result in a decreased breakdown of phospholipids, a process which occurs continuously within the cell. On the other hand, free TCA or metabolites thereof may also inhibit phospholipases directly. Both of these mechanisms might result in phospholipidosis. Interaction between CAD:s and a phospholipid bilayer has been investigated employing nuclear magnetic resonance (NMR), (Seydel and Wassermann,

505

1976). This study demonstrated that both electrostatic, as well as hydrophobic forces are involved in the binding of such drugs to a phospholipid bilayer. Later studies have shown that there are di€erences in the capacities of di€erent TCAs to bind to lamellar bodies. For the three TCAs investigated, the order of capacity for binding of the hydrophobic moiety to the membrane was trimipramine > imipramine >> clomipramine. On the other hand, the binding capacity to the hydrophilic moiety was in the order clomipramine> trimipramine >> imipramine (Joshi et al., 1989). Although the hydrophobic moiety of clomipramine interacts weakly with membranes, this compound is a potent inducer of phospholipidosis. In the case of arti®cially prepared phospholipid vesicles consisting of L-a-dipalmitoylphosphatidylcholine, imipramine showed a higher capacity for binding compared to trimipramine, both in the hydrophobic and hydrophilic regions than did trimipramine (Joshi et al., 1988). Further evidence that one of the mechanisms by which TCAs cause phospholipidosis is through inhibition of phospholipid degradation was provided by studies on the accumulation of 14 C-glycerollabelled phospholipids, especially phosphatidylinositol and phosphatidylethanolamine (Fauster et al., 1983). This phospholipid accumulation appeared after a 12 h incubation by desimipramine and was dose-dependent. In a study involving treatment of human monocytes with imipramine, we found a continuous increase of 32 P incorporation into the phospholipids of treated cells, compared to controls, after 12±14 h of treatment (Fig. 3) (unpublished data). This ®nding is in accordance with what has been reported by Leli and Hauser. They observed in C6 glioma cells that incorporation of 32 P into total phospholipids increased upon exposure to desmethylimipramine reaching a plateau at about 300 mM (Leli and Hauser, 1987). 3.3. Inhibition of phospholipases by TCAs A set of enzymes designated phospholipases regulate phospholipid turnover. The most important phospholipases are A1, A2 and C. Imipramine has served as a model for TCA in investigations of phospholipase inhibition and this drug inhibits both phospholipases A and C. Imipramine exhibits the same IC50 value for phospholipases A as does chlorphentermine, a wellknown cationic amphiphilic drug, and besides, a potent inducer of phospholipidosis. However, imipramine was somewhat more ecient in inhibiting phospholipase C compared to chlorphentermine (IC50 values of 0.25 and 0.45 mM, respectively) (Hostetler and Matsuzawa, 1981). Sphingomyelinase is a phospholipase C-type enzyme that participates in the ®rst

506

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

hydrolytic cleavage step in sphingomyelin degradation, yielding phosphocholine and ceramide. This enzyme is known, for instance, to be de®cient in the well-known Niemann±Pick syndrome. Imipramine, desimipramine and amitrypline have been reported to inhibit sphingomyelinase activity in glioma and neuroblastoma cells and in human ®broblasts as well (Albouz et al., 1983a, 1983b). Increasing the polarity of the molecule may reduce the enzyme activity, as may insertion of a hydroxyl group. The order of sphingomyelinase inhibition by these three TCAs investigated was desimipramine> imipramine> amitryptiline. However, 2-OH-imipramine and 10-OH-imipramine did not inhibit sphingomyelinase activity to the same degree as did the parent drug (Albouz et al., 1983a, 1983b). 3.4. Organ distribution of TCA-induced phospholipidosis in experimental animal models Over a number of years, tricyclic antidepressantinduced phospholipidosis has been extensively studied in several experimental models. In the early 1970s, the Kiel group, consisting of Heinz Lullmann, Renate Lullmann-Rauch and their coworkers reported that TCAs caused the same type of ultrastructural alterations that had been described earlier for the anorectic drug chlorphentermine (Lullmann et al., 1973). In this context, the most common experimental model is the rat. In both the Wistar and SpragueDawley strains, TCAs induce phospholipidosis in almost all organs, a so-called generalized lipidosis. However, there seem to be large di€erences in the sensitivities amongst di€erent organs. In the case of the endocrine organs, steroid-producing cells of the adrenal cortex, testis and ovaries (Yates et al., 1967) are particularly susceptible to drug-induced lipidosis. Furthermore, TCAs have been described to induce a heavy accumulation of myeloid bodies in lymphocytes localized in lymph nodes (Lullmann-Rauch, 1974a, 1974b) as well as in circulating lymphocytes in Sprague-Dawley rats (Lullmann-Rauch and NaÈssberger, 1983). The TCAs examined and found to induce phospholipidosis in several organs of experimental animals include iprindole, doxepin, triparanol, imipramine, clomipramine, citalopram, 1-chloroamitriptyline and 10,11 dehydroamitriptyline. The latter two compounds have to our knowledge never appeared in the market. Experimental lipidosis usually appears after chronic intake, i.e. a period of administration longer than 3 weeks. However, the latency period prior to the appearance of this e€ect may vary from a few days to months. The time required for the development of TCA-induced phospholipidosis depends on the dosage and animal species, as well as on the molecular struc-

ture and the partition coecient between octanol and water of the compound. Species di€erences may be due to the di€erences in the capacity for metabolic elimination of CADs, either in the target or in the nontarget cells. Drug-induced lipidosis has also been described as occurring in di€erent parts of the central nervous system. Phospholipidosis induced by iprindole, 1-chloroamitryptyline and clomipramine in rat choroid plexus epithelial cells requires 4±10 weeks of treatment (Frisch and Lullmann-Rauch, 1979). On the ultrastructural level, most of the inclusions observed in this case were lamellated bodies but occasionally inclusions with a crystalloid pattern were also found. The same type of morphological alterations have been observed in retinal pigment epithelium and in the neuroretinal cells during TCA treatment (Drenckhahn and LullmannRauch, 1978). Chronic imipramine treatment of virgin Wistar rats for as long as 9 weeks caused a pronounced lipidosis in the uterine and vaginal epithelium (Geist and Lullmann-Rauch, 1994). A characteristic phenomenon associated with oral administration with TCAs is the appearance of lipidosis in pulmonary tissue. The most striking feature found here is the appearance of so-called foam cells (Lullmann-Rauch and Scheid, 1975; Vijeyaratnam and Corrin, 1972). One explanation for this ®nding is the pronounced accumulation of TCAs in the lungs. In experimental studies on rats imipramine is heavily accumulated in the lungs. Increasing the dose of this drug from 10 to 100 mg/kg body weight resulted for instance, in only a three-fold increase in phospholipidosis in the liver, but a 20-fold increase in phospholipid accumulation in pulmonary tissue (Drew et al., 1981). Thus, there is a clear relationship between the intraorgan concentration and the development of lipidosis. It is thought that pulmonary accumulation of imipramine is the result of passive di€usion rather than facilitated transport. The foam cells demonstrated in the lungs of TCAtreated animals are macrophages derived from interstitial macrophage and heavily packed with myelin bodies of varying sizes. About 3±6 weeks of treatment is required to develop densely packed foam cells in alveolar lumina in rats (Hruban, 1984). Such cells were ®rst demonstrated in 1966 by Greselin, in experimental animals treated with AY-9644, a compound which inhibits cholesterol synthesis (Greselin, 1966). If administration of the TCA is withdrawn, the number of foam cells will quickly normalize within 2±5 weeks (Wold et al., 1976). A heavy phospholipidosis was induced in alveolar macrophages by iprindole during in vivo treatment for 4 weeks with a 25-fold increase of phospholipid content compared to macrophages from untreated animals (McNulty and Reasor, 1981).

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

4. Phospholipidosis in cell cultures In isolated cell cultures, lamellar bodies can appear within only a few hours of exposure to TCAs. In our laboratory, for example, we have found that TCAs induce a massive accumulation of myelin ®gures in cultured renal epithelial cells. We incubated LLC-PK1, a conventional cell line originating from pig kidney and RPTEC, a human primary kidney cell line with TCAs and observed the formation of myelin ®gures within 12 h (unpublished data). In human monocytes incubated with imipramine or clomipramine on the other hand, 8±16 h of incubation were required before the ®rst myelin ®gures appeared (Hansson et al., 1997). The same ultrastructural pattern as observed in our laboratory in monocytes has been demonstrated in cultured rat peritoneal macrophages upon exposure to a number of psychotropic drugs (Drenckhahn et al., 1976). In these cases, myelin ®gures appeared after a 24 h incubation period. Abnormal lamellated inclusions have also been observed in cultured rat peritoneal macrophages exposed to those local anaesthetics exhibiting strong lipid-binding properties and a CAD molecular structure (JaÈgel and Lullmann-Rauch, 1984). Type II alveolar epithelial cells can be successfully isolated and maintained in monolayer culture. However, during long-term culture these cells lose their characteristic lamellar bodies. L-2 cells represent a clonal cell line derived from rat Type II cells. These cells normally fail to demonstrate lamellar body inclusions. However, when incubated with iprindole, L-2 cell cultures develop typical lamellar cytoplasmic inclusions within 8 h (Martin and Kachel, 1987). After 24 h of exposure, imipramine induces the formation of myleoid bodies in human ®broblasts in vitro (Palmeri et al., 1992). 5. Reversibility Binding-studies have only been carried out for the classic TCAs. The binding between a TCA and di€erent phospholipids involves primarily electrostatic forces and is therefore, reversible. The reversibility of the TCA/phospholipid complex will depend entirely on the dissociation rate constant under the given intracellular conditions and on the actual concentration of the drug in question. Lowering the TCA concentration in tissues by discontinuing the administration will eventually lead to complete dissociation of the TCAlipid complex. The time course of this process is determined by the rate of elimination of the respective TCA and by the dissociation rate constant of the complex. By dissociation, the phospholipids regain their normal

507

properties and can be metabolized or utilized by the cell in the usual fashion. To date, it appears that cells which undergo TCAinduced lipidosis or, indeed lipidosis of other kinds, do not undergo secondary changes that are irreversible. The time required for the recovery after withdrawal of the TCA ranges from 3±4 weeks to several months, depending on the organ studied. Anyhow, to our knowledge TCA-induced lipidosis in experimental animals is reversible in all organs a€ected. In the case of L-2 cells in culture, withdrawal of iprindole resulted in the progressive loss of lamellar bodies in the cytoplasm with eventual return of the L-2 cells to the morphology of these cells to that which they exhibited prior to drug treatment (Martin and Kachel, 1987). 6. Functional implications Although the phenomenon of lipidosis has been well-known for many years, little is yet known concerning the functional signi®cance of lipidosis induced by TCA and other drugs. It seems unlikely that the extensive morphological alterations seen in di€erent kinds of cells are not associated with some kind of cellular dysfunction. It may, however, be dicult to determine whether an eventual dysfunction is due to the lipidosis itself or is a separate e€ect of the TCA administered. Surprisingly, we have found no studies on possible dysfunctions related speci®cally to TCA-induced lipidosis in the literature. However, a puzzling ®nding is that phospholipidotic alveolar macrophages release less reactive oxygen species compared to normal macrophages when exposed to zymosan particles, though they exhibit a greater oxygen consumption (McNulty and Reasor, 1981). It may be that the oxygen consumed in the respiratory burst is being utilized di€erently by phospholipidotic alveolar macrophages making less oxygen available for production of oxygen reactive species. Alternatively, the oxidant species could be consumed within phospolipidotic cells. It may then be accompanied by an increase in lipid peroxidation within the cell, which has been demonstrated for in chlorphentermine-induced alveolar macrophages (Reasor et al., 1980). On the other hand, antidepressants have, of course, been reported to exhibit e€ects other than alleviation of depression. One functional implication that has been reported to be associated with TCA treatment, but probably not directly related to lipidosis, is the reduced paw edema in rats caused by administration of iprindole or 1-chloroamitriptyline (Lullmann-Rauch, 1975). Furthermore, it has been reported that imipramine inhibits release of histamine from rat peritoneal mast cells (Ferjan and Erjavec,

508

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

Fig. 4. Appearance of myelin ®gures in human monocytes after clomipramine (30 mM) treatment for 72 h (27,000).

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

1996). In addition, the tricyclic antidepressants amitriptyline, doxepin and imipramine reduce the plasma histamine levels observed during treatment with the compound 48/80 (Irman-Florjanc and Stanovnik, 1998). Further examples of non-antidepressant e€ects are inhibition of synthesis or release of interleukins in mono- and lymphocytes (Xia et al., 1996a, 1996b). Recently, we have demonstrated that TCAs induce apoptosis speci®cally in human T-lymphocytes in vitro and that this process is not related to phospholipidosis. However, in experimental animals it has been reported that chronic treatment with chlorphentermine reduces the ability of the kidneys to concentrate and excrete a water load, while elevating the blood urea level. TCAs cause exactly the same morphological alterations as chlorphentermine in certain types of kidney cells. The cells most a€ected are the podocytes and the epithelial cells of the proximal tubuli and of collecting ducts (Lullmann-Rauch, 1975). However, a physiological study, such as that described above for chlorphentermine has not been carried out for the TCAs. It has been demonstrated that iprindole causes cataract and corneal alterations in rats. These conditions were associated with the appearance of lamellated inclusions in the lental epithelium and in corneal epithelial and stromal cells. Furthermore, retinal lipidosis has been demonstrated in albino rats receiving high oral doses of iprindole, 1-chloro-amiptriptyline, imipramine and clomipramine for several weeks. Abnormal cytoplasmic inclusions with a crystalloid substructure were identi®ed in retinal pigment epithelium. Cell inclusions with a typical laminar pattern were found in retinal ganglion cells as well (LullmannRauch, 1976). Recovery from all of these e€ects was complete within 10 weeks after termination of drug treatment. Prolonged treatment of rats with iprindole, clomipramine or 1-chloro-amiptriptyline has been found to cause phospholipidosis in dorsal root ganglion cells (Lullmann-Rauch, 1974a, 1974b). The same is a wellknown phenomenon seen in humans su€ering from Niemann±Pick syndrome, an in-born lipidosis disorder (Robb and Kuwabara, 1973). It has also been found that administration of TCAs causes myopathy and lipidosis in the rat leg muscle. However, the muscle ®bre necrosis observed, appears not to be a direct result of lipidosis, since in short-term experiments ®bre necrosis occurred long before lipidosis had fully developed. Moreover, no correlation between the number of lipidosis-speci®c inclusions and the frequency of ®bre necrosis was observed (Dreckhahn and LullmannRauch, 1979). Such pathological changes as those described above might, in fact, correspond to a very late stage in an apoptotic process. However, it has been observed that non-symptomatic carriers and patients with an inher-

509

ited lipidosis (GM2 gangliosidosis) may deteriorate during, antidepressant treatment (Navon and Baran, 1987; Federico et al., 1991). The worsening of the clinical condition of these patients might be related to a reduction in hexosamindase activity. In recent years, a few patients on antidepressant treatment have been reported to demonstrate peripheral neuropathy, myoclonus and a Creutzfeldt±Jacob like syndrome consistent with lesions resembling lipidosis in di€erent tissues (Federico et al., 1989, 1992). It is, however, doubtful if the lipidosis induced is involved in the pathogenesis of these disorders. 7. TCA-induced lipidosis in human peripheral blood cells TCAs fail to induce lipidosis in isolated human Tas well as in B-lymphocytes under in vitro conditions. However, in monocytes well-developed myeloid bodies appear after 48±72 h of incubation with imipramine, clomipramine, or citalopram (Hansson et al., 1997) Fig. 4. After a corresponding 12 h incubation, no myeloid bodies were detected in granulocytes, but disorganization of the cytoplasm was observed. In this context, it can be mentioned that chlorphentermine doesn't induce lipidosis in human lymphocytes either. No clinical studies involving examination of peripheral blood cells from patients undergoing treatment with TCAs have appeared. We have developed a simple method based on ¯ow cytometry for the quantitation of myelin ®gures in cells. This procedure should make it possible to monitor the occurrence of inclusions in peripheral blood cells from individuals on TCA treatment (Xia et al., 1997a, 1997b). 8. Concluding remarks TCAs have been in clinical use for several decades now. They possess anticholinergic e€ects and their most serious adverse e€ects are cardio- and neurotoxicity. They are also well-known as potent inducers of lipidosis in experimental animals. Interesting questions which remain are: (1) to what extent such phospholipidosis occurs in humans on TCA treatment and (2) whether phospholipidosis is responsible for any of the side-e€ects described for TCAs. TCA-induced phospholipidosis displays the ultrastructural and histochemical features similiar to those present in patients with inborn lipid-storage diseases. These in-born diseases are primarily due to de®ciencies in certain lysosomal enzymes. These storage diseases are characterized by poor outcome in most cases. The morphological alterations have been directly linked to

510

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

the severe clinical symptoms associated with these types of in-born diseases. However, it is important to point out that in the case of TCA-induced phospholipidosis, such a linkage has not yet been established. One should be cautious when attempting to elucidate the cause of TCA-induced lipidosis in vivo, since the pharmacokinetics of indidivual TCAs may be an important determinant of their ability to induce phospholipidosis. There are examples of other cationic amphipathic drugs known to bind to phospholipids and to accumulate in lung slices in vitro which do not induce pulmonary lipidosis. This may be due to a lack of drug accumulation in the lungs in vivo because of pulmonary metabolism and elimination. Obviously, it is necessary to be careful when comparing ®ndings observed in experimental models with the clinical situation. When the phenomenon of lipidosis is discussed, potential deletrious e€ects are always in focus. However, this phenomenon may perhaps function as a defense for the cell. Upon exposure to certain compounds, the cell may sequester the drug in their lysosomes, which can be further expanded into myelin bodies in order to minimize its free intracellular concentration. This might provide an explanation for the di€erence in the sensitivities of lympho- and monocytes towards TCA-induced apoptosis. The resistance of monocytes may simply re¯ect the fact that in these cells TCAs are neutralized in myelin bodies, so that the free intracellular concentration is too low to induce apoptosis. The drug-phospholipid complexes that accumulate in secondary lysosomes are probably metabolically inert (?). However, TCAs highly concentrated within lysosomes may also interfere with the functions of this organelle either by interaction with lysosomal hydrolases and/or by raising the intra-lysosmal pH. Thus, the ability of TCAs to produce lipidosis re¯ects their ability to inhibit phospholipid catabolism. Although TCAs have been in clinical use for a very long time, to our knowledge no cytoplasmic inclusion bodies corresponding to those observed in in vitro and in vivo experimental models have been described in humans. This might, of course, simply re¯ect a lack of knowledge and/or interest in the fact that these compounds may cause disturbances in phospholipid homeostasis. An alternative explanation is that TCAs may undergo more rapid metabolism to forms which are less potent as inducers of lipidosis in humans compared to experimental animals. At present, there is no clear evidence that TCA-induced phospholipidosis might be involved in the development of the common cardio- and/or neurological side e€ects often observed during treatment with these drugs. Diethylaminoethoxyhexestrol, a drug developed to reduce serum cholesterol levels, and amiodarone are the only drugs presently known to induce lipidosis in

humans (Yamamoto et al., 1971). Amiodarone is a potent antiarrhythmic agent with a molecular structure similar to that of TCAs and is known for its pronounced pulmonary toxicity. Patients su€ering from this adverse e€ect demonstrate inclusion bodies containing whorls of membrane in their type II pneumocytes and interstitial cells, as well as in extrapulmonary tissue (Dake et al., 1985). Few attempts have been made to correlate the physiological e€ects of TCAs in experimental animals to the morphological changes associated with phospholipidosis. However, there is an increasing evidence that cationic amphiphilic drugs may have e€ects on immune function, signal transduction and receptormediated events, e€ects which might be related to disturbances in phospholipid metabolism. In our in vitro studies on the e€ects of TCA on human monocytes, we observed suppression of phagocytosis by these cells at an early stage in the exposure. During this initial phase, no myelin bodies are demonstrated. After 2 days of incubation, lipidosis could be detected morphologically and in parallel, an enhancement in phagocytic activity occurred. Treatment of rats with TCAs for a longer period (4 weeks) also leads to an enhancement in monocyte/macrophage phagocytosis (McNulty and Reasor, 1983). Thus, it would appear that induction of lipidosis prevents impairment of the phagocytotic activity of monocytes by TCAs. Monocytes exposed to TCAs in vitro did not demonstrate a reduction in the numbers of their plasma membrane receptors. The initial reduction in phagocytotic activity might thus re¯ect interactions between the TCA and structures in the plasma membrane. This ®nding may be related to the observation that macrophages obtained from rats receiving acute chlorphentermine treatment display an impairment in Fc-mediated binding (Lehnert and Ferin, 1983). However, processes which can alleviate the e€ects of TCA-membrane interactions, e.g. an increased expression of receptors, may take place after a time, resulting in normalization of phagocytosis. Another example where myelin bodies may play a role in preventing a detrimental biological process concerns apoptosis in monocytes. Human lymphocytes have been shown to be very sensitive to TCA-induced apoptosis (Xia et al., 1996a, 1996b, 1997a, 1997b) whereas monocytes/macrophages are resistant to this process (unpublished data). A striking di€erence between these two cell types is that TCAs induce large myelin bodies in human monocytes/macrophages, but fail to produce lipidosis in the lymphocytes. Our interpretation of these observations is that lack of these structures in lymphocytes allows accumulation of a much higher intracellular concentration of free TCA than occurs in monocytes/macrophages. We have found previously that TCAs inhibit both

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

PHA- and con A- induced mitogenesis in lymphocytes (MaÊrtensson and NaÈssberger, 1993). In context, the same results have been observed also in an in vitro experimental model. In a study of mouse splenic lymphocytes, tricyclic antidepressants at high concentrations inhibited mitogen-induced proliferation of splenic Tand B- lymphocytes from treated mice (Audus and Gordon, 1982). The same phenomenon has been described for chlorphentermine, which has been found to suppress the blastogenic response (Sauers et al., 1986). It has also been demonstrated that chlorphentermine inhibits hydrolysis of phosphatidylinositol to diacylglycerol and inositol triphosphate (Sauers et al., 1988). This is an early event in blastogenesis, but it is not known whether TCAs also regulate this biochemical process or whether this conversion is related to phospholipidosis. An increase in the apolarity of a TCA enhances its potency in suppressing blastoid transformation and, in parallel, renders the molecule a more potent inducer of lipidosis, indicating a possible linkage between these two phenomena. In conclusion, TCAs as well as several of the newer generation of antidepressants are potent inducers of phospholipidosis. Despite extensive studies over a period of many years, it still remains unclear to what extent disturbances in phospholipid metabolism may be related to any of the adverse e€ects associated with this group of drugs. These compounds also have other interesting biological e€ects especially on immunocompetent cells, that may suggest additional clinical applications for these drugs in the future.

Acknowledgements The work from the author's laboratory referred to here was funded by the Lundbeck Foundation, Copenhagen, Denmark.

References Albouz, S., LeSaux, F., Wenger, D., Hauw, J.J., Baumannn, N., 1983a. Modi®cations of sphingomyelin and phosphatidylcholine metabolism by tricyclic antidepressants and phenothiazines. Life Sciences 38, 357±363. Albouz, S., Vanier, A.T., Hauw, J.J., LeSaux, F., Boutry, J.M., Baumann, N., 1993b. E€ect of tricyclic antidepressants on sphincomyelinase and other sphingolipid hydrolases in C6 cultured glioma cells. Neuroscience Letters 36, 311±315. Ashman, R.B., Blanden, R.V., Ninhan, B.W., Evans, D.F., 1986. Interaction of amphiphilic aggregates with cells of the immune system. Immunology Today 7, 278±283. Audus, K.L., Gordon, M.A., 1982. Tricyclic antidepressant e€ects on the murine lymphocyte mitogen response. J. Immunopharmacol 4, 13±27. Bockhardt, H., Lullmann-Rauch, R., 1980. Zimelidine-induced lipidosis in rats. Acta Pharmacol et Toxicol 47, 45±48.

511

Dake, M.D., Madison, J.M., Montgomery, C.K., Shellito, J.E., Hinchcli€e, W.A., Winkler, M.L., Bainton, D.F., 1985. Electron microscopic demonstration of lysosmal inclusion bodies in lung, liver, lymph nodes and blood leukocytes of patients with amiodarone pulmonary toxicity. Am. J. Med 78, 506±512. Dreckhahn, D., Lullmann-Rauch, R., 1979. Experimental myopathy induced by amphiphilic cationic compounds including several psychotropic drugs. Neuroscience 4, 549±562. Drenckhahn, D., Kleine, L., Lullmann-Rauch, R., 1976. Lysosomal alterations in cultured macrophages exposed to anorexigenic and psychotropic drugs. Lab Invest 35, 116±123. Drenckhahn, D., Lullmann-Rauch, R., 1978. Di€erential susceptibilities of pigment epithelium and neuroretina toward several amphiphilic cationic drugs. Exp. Mol. Pathol 28, 360±371. Drew, R., Siddik, Z.H., Mimnaugh, E.G., Gram, T.E., 1981. Species and dose di€erences in the accumulation of imipramine by mammalian lungs. Drug Metabolism and Dispos 9, 322±326. Fauster, R., Honegger, K., Wiesman, U., 1983. Inhibition of phospholipid degradation and changes of the phospholipid-pattern by desimipramine in cultured human ®broblasts. Biochem. Pharmacol 32, 1737±1744. Federico, A., Palmeri, S., Malandrini, A., Fabrizi, G., Mondelli, M., Guazzi, G.C., 1991. The clinical aspects of adult hexosaminidase de®ciencies. Dev. Neurosci 13, 280±287. Federico, A., Palmeri, S., Malandrini, A., Mangano, L., Ciacci, G., Scarpini, C., Tiacci, G., 1992. Dementia, myoclonus, peripheral neuropathy and lipid-like material in skin biopsy during psychotropic drug, treatment. Biol. Physchiatry 32, 721±727. Ferjan, I., Erjavec, F., 1996. Changes in histamine and serotonin secretion from rat peritoneal mast cells caused by antidepressants. In¯am. Res 45, 141±144. Foerstl, J., Hohagen, F., Hewer, W., Pohlman-Eden, B., 1989. Another case of Creutzfeldt-Jacob like-syndrome due to antidepressant toxicity. J. Neurol. Neurosurg. Psychiatry 52, 920± 926. Frisch, W., Lullmann-Rauch, R., 1979. Di€erential e€ects of chloroquine and of several other amphiphilic cationic drugs upon rat choroid plexus. Acta. Neuropathol. (Berl.) 46, 203±208. Geist, S.H., Lullmann-Rauch, R., 1994. Experimentally induced lipidosis in uterine and vaginal epithelium of rats. Annals Anatomy 176, 3±9. Greselin, E., 1966. An inhibitor of cholesterol biosynthesis and the alveolar macrophages. Can. J. Comp. Med. Vet. Sci 30, 121±126. Hansson, A.L., Xia, Z., Berglund, M.C., Bergstrand, A., dePierre, J.W., NaÈssberger, L., 1997. Reduced cell survival and morphological alterations induced by three tricyclic antidepressants in human peripheral monocytes and lymphocytes and in cell lines derived from these cell types. Toxicol. in Vitro 11, 21±31. Howell, J.I., Fisher, D., Goodall, A.H., Verrinder, M., Lucy, J.A., 1973. Interactions of membrane phospholipids with fusogenic lipids. Biochim. Biophys. Acta 332, 1±10. Hostetler, K.Y., Matsuzawa, Y., 1981. Studies on the mechanism of drug-induced lipidosis cationic amphipathic drug inhibition of lysosomal phospholipases A and C. Biochem. Pharmacol 30, 1121±1126. Hruban, Z., 1984. Pulmonary and generalized lysosomal storage induced by amphiphilic drugs. Environmental Health Perspectives 55, 53±76. Irman-Florjanc, T., Stanovnik, L., 1998. Tricyclic antidepressants change plasma histamine kinetics after its secretion induced by compound 48/80 in the rat. In¯am. Res 47, s26±s27. JaÈgel, M., Lullmann-Rauch, R., 1984. Lipidosis-like alterations in cultured macrophages exposed to local anaesthetics. Arch. Toxicol 55, 229±232. Joshi, U.M., Kodavanti, S., Couderts, B., Dwyer, T.M., Mehendale, H.M., 1988. Types of interaction of amphiphilic drugs with phospholipid vesicles. J. Pharmacol. Exp. Toxicol 246, 150±157.

512

Z. Xia et al. / Progress in Neurobiology 60 (2000) 501±512

Joshi, U.M., Rao, P., Kodavanti, S., Lockard, V.G., Mehendale, H.M., 1989. Fluorescence studies on binding of amphiphilic drugs to isolated lamellar bodies relevance to phospholipidosis. Biochim. Blophys. Acta 1004, 309±320. Lehnert, B.E., Ferin, J., 1983. Particle binding phagocytosis and plastic substrate adherence characteristic of alveolar macrophages from rats acutely treated with chlorphentermine. J. Reticuloendothel. Soc 33, 293±303. Leli, U., Hauser, G., 1987. Modi®cations of phospholipid metabolism induced by chlorpromazine, desmethylimipramine and propanol in C6 glioma cells. Biochem. Pharmacol 36, 31±37. Lullmann, H., Lullmann-Rauch, R., Rell, G.H., 1973. A comparative ultrastructural study of the e€ects of chlorphentermine and triparanol in rat lung and adrenal gland. Virchows Arch. Cell. Pathol. Abt B 12, 91±103. Lullmann-Rauch, R., Reil, G.H., Scheid, D., 1973. LipidosisaÈhnliche ZellveraÈnderung bei der Ratte nach Behandlungmit Thymoleptika. Verh Dtsch Ges Path 57, 425±431. Lullmann-Rauch, R., 1974a. Lipidosis-like ultrastructural alterations in rat lymph nodes after treatment with antidepressants or neuroleptics. Naunyn-Schmiedeberg's Arch. Pharmacol 286, 165±175. Lullmann-Rauch, R., 1974b. Lipidosis-like alterations in dorsal root ganglion cells of rats treated with tricyclic antidepressants. Naunyn-Schmiedeberg's Arch. Pharmacol 283, 219±222. Lullmann-Rauch, R., 1975. Lipidosis-like renal changes in rats treated with chlorphentermine or with tricyclic antidepressants. Virchows Arch. Abt. B Cell Pathol 18, 51±62. Lullmann-Rauch, R., Scheid, D., 1975. Intraalveolar foam cells associated with lipidosis-like alterations in lung and liver of rats treated with tricyclic psychotropic drugs. Virchows Arch. Abt. B Cell Pathol 19, 255±268. Lullmann-Rauch, R., 1976. Retinal lipidosis in albino rats treated with chlorphentermine and with tricyclic antidepressants. Acta. Neuropath. (Berl.) 35, 55±67. Lullmann-Rauch, R., NaÈssberger, L., 1983. Citalopram-induced generalized lipidosis in rats. Acta. Pharmacol. et. Toxicol 52, 161± 167. MaÊrtensson, U., NaÈssberger, L., 1993. In¯uence of antidepressants on mitogen stimulation of human lymphocytes. Toxicol. in Vitro 7, 241±245. Martin, W.J., Kachel, D.L., 1987. Iprindole reverses the lamellar body de®ciency of cultured L-2 cells. Am. J. Pathol 129, 34±43. McNulty, M.J., Reasor, M.J., 1981. Iprindole-induced phospholipidosis in rat alveolar macrophages: alterations in oxygen consumption and release of oxidants. Exp. Lung. Res 2, 57±69. McNulty, M.J., Reasor, M.J., 1983. Enhanced phagocytotic and bactericidal activities of phospholipidotic rat alveolar macrophages. J. Reticuloendothel. Soc 30, 539±549. Navon, R., Baran, D., 1987. Depletion of cellular b-hexosaminidase by imipramine is prevented by dexamethasone; implications for treating psychotic hexosaminidase-A de®cient patients. Biochem. Biophys. Res. Commun 148, 1098±1103.

Palmeri, S., Mangano, L., Battisti, C., Malandrini, A., Federico, A., 1992. Imipramine induced lipidosis and dexamethasone e€ect: morphological and biochemical study in normal and chronic GM2 gangliosidosis ®broblasts. J. Neurol. Sci 110, 215±221. Reasor, M.J., Koshut, R.A., McNulty, M.J., 1980. Chemiluminescence from rat alveolar macrophages following induction of phospholipidosis with chlorphentermine. Toxicol. Appl. Pharmacol 52, 497±506. Robb, R.M., Kuwabara, T., 1973. The ocular pathology of Type A Niemann±Pick disease. Invest. Opthalmol 12, 366±373. Sauers, L.J., Wierda, D., Walker, E.R., Reasor, M.J., 1986. Morphological and functional charges in mouse splenic lymphocytes following in vivo and in vitro exposure to chlorphentermine. J. Immunopharmacol 8, 611±631. Sauers, L.J., Wierda, D., Reasor, M.J., 1988. Chlorphentermine suppresses the phosphatidylinositol pathway in concanavalin A activated mouse splenic lymphocytes. Immunopharmacol. Immunotoxicol 10, 1±9. Seydel, J.K., Wassermann, O., 1976. NMR-studies on the molecular basis of drug-induced phospholipidosis II. Interaction between several amphiphilic drugs and phospholipids. Biochem. Pharmacol 25, 2357±2364. Vijeyaratnam, G.S., Corrin, B., 1972. Pulmonary histiocytosis simulating desquamative interstitial pneumonia in rats receiving oral iprindole. J. Pathol. Bacteriol 108, 105±111. Wold, J.S., Joost, R.R., Gring, W.J., Marroquine, P., Harris, P.N., 1976. Phospholipid accumulation in rats produced by ¯uoxetine and chlorphentermine. Toxicol. Appl. Pharmacol 37, 118±119. Xia, Z., dePierre, J.W., NaÈssberger, L., 1996a. Tricyclic antidepressants inhibit IL-6. IL-1b and TNFa release in human blood monocytes and IL-2 and interferon g in T cells. Immunopharmacol 34, 27±37. Xia, Z., dePierre, J.W., NaÈssberger, L., 1996b. The tricyclic antidepressants clomipramine and citalopram induce apoptosis in cultured human lymphocytes. J. Pharmacy. Pharmacol 48, 115±116. Xia, Z., Appelkvist, E.L., dePierre, J.W., NaÈssberger, L., 1997a. Tricyclic antidepressant-induced lipidosis in human peripheral monocytes in vivo, as well as in a monocyte-derived cell line, as monitored by spectro¯uorimetry and ¯ow cytometry after staining with Nile red. Biochem. Pharmacol 53, 1521±1532. Xia, Z., Karlsson, H., dePierre, J.W., NaÈssberger, L., 1997b. Tricyclic antidepressants induce apoptosis in human T-lymphocytes. Int. J. Immunopharmacol 19, 645±654. Yamamoto, A., Aduchi, S., Imoto, T., Nishikawa, N., 1971. Studies on drug-induced lipidosis III. Lipid composition of the liver and some other tissues in clinical cases of ``Niemann±Pick like syndrome'' induced by 4,4-diethylethoxyhexestrol. J. Biochem. (Tokyo) 70, 775±784. Yates, R.D., Arai, K., Rappoport, D.A., 1967. Fine structure and chemical composition of opaque cytoplasmic bodies of triparanol treated Syrian hamsters. Exp. Cell Res 47, 459±478.