Biochimica et Biophysica Acta, 1011 (1989) 97-101 Elsevier

97

BBA 12439

Gadolinium chloride-induced shifts in intrahepatic distributions of liposomes

George Lfizhr *, Mieke van Galen and Gerrit L. Scherphof Laboratoo, of Physiological Chemist~. State University Groningen. Groningen (The Netherlands)

(Received 14 October 1988)

Key words: Gadolinium chloride: Macrophage blockade: Liposome: Kupffer cell: Phagocytosis: (Rat hepatocyte)

Intravenously administered gadolinium chloride caused only a slight decrease in the rate of elimination of small unilamellar liposomes from the blood and had no influence on the total hepatic uptake of these vesicles, but did alter their intrahepatic distribution substantially. Uptake by the non.parenchymal cells was substantially decreased, whereas uptake by the parenchymal cells showed a concomitant increase. Our earlier observations (Roerdin~. et al. (1981) Biochim. Biophys. Acta 677, 79-89) on the effect of lanthanides on the in vivo distribution of multilamellar liposomes have been extended, in that we demonstrate, in addition to the drop in elimination rate from the blood and in the over-all hepatic uptake, a shift of liposome distribution within the Kupffer cell population. While the larger Kupffer cells, which normally take up a major fraction of an injected liposome dose, were strongly inhibited in liposome uptake, the more numerous small macrophages showed a 3-4-fold increase in uptake.

Introduction

In a manner similar to the removal of colloidal particles or other particulate matter capable of undergo- ing phagocytosis, systemically administered multilamel- lar or large unilamellar liposomes are rapidly removed from the blood stream by the cells of the mononuclear phagocyte system (MPS). The major portion of the vesicles is taken up by the liver, although splenic and, to a lesser extent, lung and bone-marrow macrophages also contribute to the intravascular clearance of the liposomes [1,2]. Several studies have revealed that within the liver nearly all of the liposomes are taken up by the Kupffer cells, and only small portions can reach the hepatic parenchymal cells [3,4]. Previously, it was re- ported from this laboratory [3] that lanthanides, which effectively depress MPS activity [5,6], significantly re- duce the blood elimination rate and the uptake of the liposomes by Kupffer cells, but failed to influence the cellular distribution of the liposomes within the liver in favor of the hepatocytes. This is compatible with the

* Present address: Institute of Pathophysiology, University Medical School, P.O. Box 531, 6701 Szeged, Hungary.

Abbreviation: MPS, mononuclear phagocyte system.

Correspondence: G.L. Scherphof, Laboratory of Physiological Chem- istry, State University Groningen, Bloemsingei 10, 9712 KZ Gronin- gen, The Netherlands.

0167-4889/89/$03.50 © 1989 Elsevier Science Pul~I,.'2hers B.V. (Biomedical

micro-anatomical architecture of the liver [7]. The fenes- trations in the sinusoidal endothelium, with an average diameter of about 100 nm [8], will prevent access of the multilamellar or large unilamellar liposomes to the un- derlying parenchymal cells despite an increased overall availability due to blocked Kupffer cell uptake.

In contrast, small unilamellar iiposomes are small enough to reach the liver parenchymal cells through the endothelial fenestrations as was repeatedly shown [9-12]. The fraction of such vesicles that actually re- aches the hepatocytes was reported to depend on the liposomal lipid composition [11]. Vesicles composed of sphingomyelin, cholesterol and phosphatidylserine in a 4 : 5 : 1 molar ratio possess a sufficiently high affinity towards the liver macrophages to allow for a contribu- tion to total liver uptake of as much as 50~ of the total liver-associated fraction. Therefore, in the present study we set out to establish whether suppression of the endocytic capacity of the Kupffer cells by lanthanide administration will cause small unilamellar liposomes of this particular lipid composition to shift their intrahe- patic distribution in favor of the parenchymal cells.

We previously reported [13,14] the fractionation of the liver macrophage population into several sub-frac- tions by means of stepwise elutriation centrifugation. These Kupffer cell sub-fractions display different mor- phological and functional characteristics, e.g., dif- ferences in phagocytic and lipolytic activities both in vivo and in vitro [13,14]. It seemed of interest to estab-

Division)

98

fish whether the rare earth metal salts such as gadolinium chloride could differentiate between the various Kupffer cell sub-fractions with respect to their reported inhibi- tory action on Kupffer cell phagocytosis [5,6].

The values presented are means with standard devia- tion of three independent determinations.

Results

Materials and Methods

Cholesterol, egg phosphatidylcholine, bovine brain sphingomyeEn, L-a-phosphatidylserine, dicetyl phos- phate and collagenase type I were obtained from Sigma; pronase-E was from Merck; DNase type I grade II was from Boehringer; [cholesteryl-l,2(n)-3H]cholesteryl hexadecyl ether was purchased from New England Nuclear and gadolinium chloride (GdC! 3) was obtained from ICN Pharmaceuticals Inc.

Small uni!amcllar vesicles composed of cholesterol, sphingomyelin and phosphatidylserine or of cholesterol, egg phosphatidylcholine and phosphatidylserine, both in a molar ratio of 5 :4 :1 and labeled with a tracer amount of [3H]cholesterol ether, were prepared by sonification of a dispersion of these lipids in a Tris- buffered salt solution as described previously [9]. Multi- lamellar vesicles composed of egg phosphatidylcholine, dicetyl phosphate or phosphatidylserine and cholesterol in a molar ratio of 4 :1 :5 , were prepared in a buffer comprising 10 mM Hepes and 135 rnM NaCl buffer (pH 7.4) and were sized according to Olson et al. [15] by extrusion through a polycarbonate filter of 0.8/zm pore size (Nuclepore).

To depress the Kupffer cell activity, animals received gadolinium chloride intravenously, in a dose of 20/z mol per kg body weight 24 h prior to administration of the liposomes.

Both multilamellar and small unilamellar liposomes, in a dose of 50/zmol of total lipid per kg body weight, were injected into the penile vein of ether-anesthetized male Wistar rats varyi.ng in body weight from 190 to 230 g. Liposome elimination from the blood was de- termined by measuring radioactivity in tail vein samples following decolorization with H202. Blood volume was taken as 65 ml per kg body weight.

Liposomal uptake by liver and spleen was de- termined by measuring the radioactivity contents of homogenates of the organs. To determine the intrahe- patic distribution of the liposomes, the livers were per- fused with collagenase and the radioactivity contents of the isolated parenchymal and non-parenchymal cell fractions were measured, as described previously [3], assuming 450. l0 T hepatocytes and 194.107 non- parenchymal cells per kg body weight [16]. To measure the in vivo uptake of liposomes by the Kupffer cell sub-populations, the fiver parenchyma was digested with pronase, and the Kupffer cells were fractionated accord- ing to size by means of stepwise counter-flow centrifu- gallon [14] following pre-fractionation of the non- parenchymal cells on Metrizamide ® gradients [17].

Gadolinium chloride caused a slight decrease in the blood elimination of intravenously injected small un- ilamellar liposomes (Fig. 1). It did not influence the total hepatic uptake of such vesicles, but greatly altered their intrahepatic distribution in favor of the parenchymal cells (Table l). During the initial elimina- tion phase, gadolinium chloride decreased the blood clearance rate of the small unilamellar liposomes com- posed of cholesterol, sphingomyelin and phosphati- dylserine, but at the end of the observation period, 17½ h after the administration of liposomes, the effect of the lanthanide on the blood concentration of the iiposomes was negligible. Nonetheless, uptake of liposomes by the non-parenchymal cells was significantly decreased and that by the parenchymal cells concomitantly increased, due to the action of the lanthanide. This intrahepatic shift of liposomes from the non-parenchymal to the parenchymal cell fraction is even more clearly demon- strafed whe~ considering the relative contributions of these cells normalized to 100% t~,zal liver uptake; in the control animals, less than 60% of the total liver-associ-

100 r0diooctivity

~0 • I

suv MLV MlV-13d

I I I I I 0 0 30 ~ 120 180 240

time (mini Fig. ]. Elimination of small unilamellar liposomes and multilamellar liposomes from blood; effect of OdCi 3. Rats were injected in- travenously with radiolabeled small unilamellar liposomes (SUV) composed of cholesterol, sphingomyelin and phosphatidylserine (5 : 4:1), or multilameilar liposomes (MLV) composed of cholesterol, egg phosphatidylcholine and dicetyl phosphate (5:4:1), in a dose of 50/z mol of lipid/kg body weight. The percent radioactivity remaining in the blood was plotted against time. Gadolinium chloride was given intravenously in a dose of 20 /zmol/kg body weight 24 h before liposome administration. Data are means :l: S.D. of three different experiments. The points marked with an asterisk are significantly different from the corresponding control (P < 0.05). ®, SUV control;

o, SUV GdCI3; ×, MLV control; ,% MLV GdCI 3. f

99

TABLE i

Effect of GdCl ~ on the in vivo fate of small unilamellar iiposomes

[14C]Cholesteroi ether-labeled small unilamellar liposomes composed of cholesterol, sphingomyelin and phosphatidylserine (5:4:1) were injected intravenously, and 17.5 h later the radioactivity in the blood, spleen and liver was determined as described in the Materials and Methods. The total uptake by parenchymal cells (PC) and non-parenchymal cells (NPC) was determined by assuming the presence of 450-l0 T hepatocytes and 194.107 non-parenchymal cells per kg body weight. Gadolinium chloride (2 /~mol per 100 g body weight) was given intravenously 24 h before liposome administration. The values given are percent of injected dose±S.D, for three expe_rimcnts. * These values for parenchymal and non-parenchymal cells give the relative intrahepatic distributions, taking the sum of the two fractions m each experiment as 100%.

GdCI3t~atment Blood Spleen Liver PC NPC PC* NPC*

- 2.1±0.7 10.3±0.3 50.4±8.2 33.9±8.7 20.7±0.9 60.7±5.2 39.3±5.2

+ 2.1±0.7 6.7±1.9 50.0±4.6 54.0±7.9 4.8±0.2 91.6±0.9 8.4±0.9

TABLE 11

Effects of GdCi ~ on the in vivo fate of muitilamellar liposomes

[14C]Cholesterol ether-containing multilamellar liposomes composed of egg phosphatidylcholine, dicetyl phosphate and cholesterol (4:5 : 1) were injected intravenously and the radioactivity in blood, spleen, liver and parenchymal (PC) and non-parenchymal (NPC) cell fractions was determined after I h in the control animals and after 2 h in the gadolinium chloride-pretreated animals, as described in the Materials and Methods. For details, see also Table l. The values are percent of injected dose±S.D, for three experiments. * These values for parenchymal and non-parenchymal cell fractions give the relative intrahepatic distributions, taking the sum of the two fractions in each experiment as 100%.

GdCl3t~atment Bi~d Spleen Liver PC NPC PC* NPC*

- 16.7~9.0 18.8±4.5 33.6±8.9 1.6±0.5 24.4±5.4 6.5±3.4 93.5±3.4

+ 17.1±6.8 28.5±2.4 11.4±2.5 i.0±0.04 11.5±1.0 8.1±0.4 92.0±0.4

ated liposomal radioactivity was recovered in the parenchymal cells, as compared to more than 90% in the gadolinium chloride-pretreated rats (Table I). Simi- lar results were obtained with small unilamellar lipo- somes composed of egg phosphatidylcholine, cholesterol and phosphatidylserine (4:5 : 1) (not shown).

In agreement with previous reports, multilamellar vesicles composed of egg phosphatidylcholine, choleste- rol and dicetyl phosphate (4 :5 :1) were cleared from the blood stream much faster than small unilamellar vesicles, while the elimination rate was also significantly depressed by gadolinium chloride (Fig. 1). The organ uptake and intrahepatic distribution of liposomes were determined at time points at which the blood concentra- tions in the control and in the gadolinium chloride-pre- treated rats were approx, the same, i.e., after 1 and 2 h, respectively. Gadolinium chloride strongly depressed total hepatic uptake of liposomes in favor of uptake by the spleen (Table If). In contrast to our observations on the small unilamellar vesicles in Table l, the lanthanide failed to cause a shift in the iiposomal uptake of multi- lamellar liposomes from the non-parenchymal to the parenchymal cell fraction, in support of our earlier observations [3].

The effects of gadolinium chloride on the in vivo uptake of multilamellar liposomes by sub-fractions of Kupffer cells are demonstrated in Fig. 2. In the control animals we observe a gradual increase in the specific uptake of liposomal lipid, from approx. 1 nmol per 106 cells in the small cells, elutriating at 20.5 ml/min, to

40 nmd t~id perlO~ceUs

30-

20-

20.5 25 30 35 40 46.5 flow n~te (ml/min)

Fig. 2. Effects of gadolinium chloride on in vivo liposome uptake by Kupffer cells of different size. Z4C-labeled multilamellar liposomes composed of cholesterol, egg phosphatidylcholine and dicetyl phos- phate (5:4:1) were injected intravenously (50 ~amoi of iipid/kg body weight) and uptake of liposomes by Kupffer cell sub-fractions, ob- tained by counterflow centrifugation, was determined 1 or 4 h after liposome administration. Gadolinium chloride was given 24 h before liposome administration. Open bars, control animals, 1 h after lipo- somes administration; hatched bars, GdCl.~-treated animals, I h after liposome administration: solid bars. GdCI3-treated animals, 4 after iiposome administration. Flow rate numbers refer to the counterflow rate values used to elutriate the various fractions out of the rotor: small cells are washed out at a low flow rate, large cells at a high

flow rate.

100

about 30 nmol per 10 6 cells in the large cells, elutriating at 46.5 ml/min. In the GdCI 3 animals there is no longer a gradual increase but uptake shows an optimum in the cells of intermediate size; uptake by the smaller cells has increased or remained the same, while that by the larger cells is greatly diminished. This phenomenon is seen both at 1 and 4 h after liposome injection. The 1-h results (Fig. 2, hatched bars) serve to enable a comparison at the same time point as that applied in the control experiment; apparently, the gadolinium caused an approx. 3-fold decrease in total Kupffer cell uptake at that time. At 4 h after liposome injection, liposome uI~take in the total Kupffer cell fraction is about the same as that in the control, allowing a com- parison at similar total uptake values, which is probably more informative. Similar results were obtained with multilamellar liposomes composed of cholesterol, egg phosphatidylcholine and phosphatidylserine (5 : 4: 1) (not shown).

Discussion

Several attempts have been undertaken to target liposomes to specific tissues or cells. One of the possibil- ities for the targeting of liposomes is to modify the liposomal surface with iigands recognizing specific cells. As indicated in the introduction, multilamellar liposomes, because of their size, do not seem to be the most suitable for targeting towards liver pare.nchymal cells. Nonetheless, targeting of multilamellar vesicle pre- parations to hepatic parenchymal cells in vivo has been reported, in studies using surface-exposed galactose re- sidues hydrophobically anchored in the liposomal hi- layer [18], ~o allow specific uptake via the galactose receptor at the surface of the hepatocyte [19]. Also, small vesicles have reportedly been ta:geted to hepato- cytes by galactose-exposing liposomd lipid, both by ourselves [9] and others [20,21]. In the latter two re- ports, however, no cell separations were carried out nor was any morphological evidence of actual hepatocyte uptake provided which leaves the possibility that the observed increase in hepatic uptake reflected increased Kupffer cell uptake mediated by the galactose-specific lectin on these cells [22], in spite of the observed inhibi- tion by asialoglycoprotein [21], similar to our own ob- servations on sphingomyelin-based vesicles [23]. On the other hand, hepatocytic contributions to total liver up' take may depend on the bulk lipid composition of the liposomes, as we observed when comparing the results on the sphingomyelin-based liposomes with those ob- tained earlier [9] on dimyristoylphosphatidylcholine- based vesicles.

Because of the lack of a continuous basement mem- brane and because of the presence of a fenestrated endothelial lining, liver parenchymal cells are directly accessible from the blood compartment [7,8]. Liposomes

with a diameter smaller than 0.1 ~m, i.e., the average diameter of the endothelial fenestrations, can reach the hepatocytes directly. Nonetheless, a considerable pro- portion of an intravenously injected dose of small un- ilamellar vesicle may still be taken up by the Kupffer cells, depending on the liposomal lipid composition [11]. We reported ealrier [23] that the incorporation of the galactose-exposing glycolipid lactosylceramide into such small vesicles, although resulting in a 3-fold decreased half-life in blood and a concomitant increase in hepatic uptake, failed to produce the increase in hepatocyte uptake expected on account of the hepatocyte galactose receptor [19]. Rather, we observed a considerable incre- ment in the liposome uptake by the Kupffer cells. This was attributed to the galactose-binding lectin on the Kupffer cell surface [22].

In the experiments described in the present paper, we achieved a substantial shift in the uptake of small unilamellar liposomes from the Kupffer cells to parenchymai cells of the liver without modifying the liposomal surface. Gadolinium chloride, which effec- tively blocks Kupffer cell activity [5], caused only a slight decrease in the blood elimination and did not at all modify the total hepatic uptake of the intravenously injected small unilamellar liposomes, but it almost dou- bled uptake by the hepatocytes and decreased that by the Kupffer cells several-fold thus greatly altering the intrahepatic distribution of the vesicles. Our earlier studies with 51Cr-labeled foreign red blood cells [6] and with bacterial endotoxin [24] demonstrated that, as a consequence of the reduction in the hepatic component of the clearance, gadolinium chloride leads to an in- crease in the involvement of extrahepatic phagocytic sites in the vascular clearance of particulate matter. Since macrophages localized in different tissues display specific functions, the alteration in che distribution of particulate matter (for example, particulate antigens or bacterial endotoxins) may have functional consequences [25,26].

The present experiments provide evidence that by depressing Kupffer cell function, gadolinium salts may also result in increased uptake of intravenously adminis- tered particulate matter, i.e., liposomes, by cells not belonging to the MPS, in this case the liver parenchymal cells. When large liposomes were used, which cannot reach the hepatocytes, gadolinium chloride caused a shift in liposome uptake within the Kupffer cell popula- tion from the phagocytical!y most active cell fraction, containing predominantly large cells, to the cells of a much smaller size, which are normally only slightly active. These observations support the concept that specific impairment of the function of the normally active macrophage populations may evoke the involve- ment of entirely different cell populations of either macrophage- or non-macrophage origin, which, in turn, may lead to substantial changes in physiological or

pharmacological responses to externally administered stimuli. Further investigations will be required to de- termine whether the gadolinium-induced effects ~re merely the result of a passive over-flow mechanism compensating for impaire~ phagocytic capacity or whether specific stimulatory influences on normally non-active or weakly active cells are involved. Since the inhibitory effect of GdC! 3 requires almost a day to develop in vivo [5], it even remains possible that the agent acts to influence the composition of the liver macrophage population, comparable to the type of ef- fect seen with biological response modifiers such as muramyldipeptide or zymosan [27,28]. We are currently investigating these possibilities.

Acknowledgements

G.L. was financially supported by a foreign visitors grant from the Netherlands Organization for Scientific Research (NWO). We wish to thank Bert Dontje and Jan Wijbenga for excellent biotechnical assistance and Rinske Kuperus and Lineke Klap for help with the preparation of the manuscript.

References

1 Poste, G., Kirsh, R. and Koestler, T. (1984) in Liposome Technol- ogy (Gregoriadis, G., ed.), Vol. II!, pp. 1-28, CRC Press, Boca Raton.

2 Hwang, KJ. (1987) in Li~,osomes, from Biophysics to Ther- apeutics (Ostro, M.J., ed.), pp. 109-156, Marcel Dekker, New York.

3 Roerdink, F., Dijkstra, J., Hartman, G., Bolscher, B. and Scherphof, G. (1981) Biochim. Biophys. Acta 677, 79-89.

4 Rahman, Y.E., Cerny, E.A., Patel, K.R., Lau, E.H. and Wright, B..I. (1982) Life Sci. 31, 2061-2071.

5 l,fizar, G. (1973) J. Reticuloendothel. Soc. 13, 231-237. 6 Husztik, E., Lfizfir, G. and Parducz, A. (1980) Br. J. Exp. Pathol.

61,624-630. 7 Scherphof, G., Roerdink, F., Dijkstra, J., Eilens, H., De Zanger, R.

and Wisse, E. (1983) Biol. Cell 47, 47-58.

101

8 Wisse, E., De Zanger, R. and Jacobs, R. (1982) in Smusoidal Liver Cells (Knook, D.L. and Wisse, E., eds.), pp. 61-67, Elsevier Biomedical Press, Amsterdam.

9 Spanjer, tl. and Scherphof, G. (1983) Bioehim, Biophys. Acta 734, 40-47.

10 Roerdink, F.H., Regts, J., Van Leeuwen, B. and Scherphof, G.L. (]984) Bloc,him. Biophys. Act:J 770, 195-202.

11 Spanjer, H.H., Van Galen, M., Roerdink, F.H., Regts, J. and Scherphof, G.L. (1986) Biochim. Biophys. Acta 863, 224-230,

12 Poste, G., Bucana, C., Raz, A., Bugelski, P., Kir:,h, R. and Fidler, IJ. (1982) Cancer Res. 42, 1412-1416.

13 Daemen, T., Veninga, A., Roerdink, F.H. and Scherphof, G.L. (1988) in Liposomes as Drug Carriers: Trends and Progress (Gregoriadis, ed.), pp. 431-4450 John Wiley & Sons Ltd., New York.

14 Derksen, J.T.P., Morselt, H.W.M. and Scherph~f, G.L. (1988) Biochim. Biophys. Acta 971,127-136.

15 Olson, F., Hunt, C.A., Szoka, F.C., Vail, W.J. and Papahadjopou- los, D. (1979) Biochim. Biophys. Acta 557, 9-23.

16 Kooistra, T., Duursma, A.M., Bouma, J.W.M. and Grt~ber, M. (1979) Biochim. Biophys. Acta 587, 282-298.

17 Dijkstra, J., Van Galen, W.J.M., Hulstaert, C.E., Kalicharan, D., Roerdink, F.H. and Scherphof, G.L. (198a) Exp. Cell Res. 150, 161-176.

18 Ghosh, P., Das, P.K. and Bachhawat, B.K. (1982) Arch. Biochem. Biophys. 213, 266-270.

19 Ashwell, G. and Morell, A.G. (1974) Adv. Enzymol. 41, 99-128. 20 Szoka, F.C. and Mayhew, E. (1983) Biochem. Biophys. Res. Com-

mun. 110, 140-146. 21 Gregoriadis, G. and Senior, J. (1984) Biochem. Soc. Trans. 12,

337-339. 22 Kolb-Bachofen, V., Schlepper-Schiifer, J., Vogell, W. and Kolb, H.

(1982) Cell 29, 859-866. 23 Spanjer, H., Morselt, I-L and Scherphof, G. (1984) Biochim. Bi-

phys. Acta 774, 49-55. 24 Lfizar, G., Husztik, E. ai~d Ribarszki, S. (1984) in Tissue Culture

and RES (RiJchlich, P. and B:icsy, E., eds.), pp. 193-197, Akad6miai Kiad6, Budapest.

25 LAz6r, G. and Husztik, E (1985) Experienfia 41,516. 26 Lazar. G., Husztik, E. and Ribhrszki, S. (1985) in Macrophage

Biology (Reichard, S. and Kojima, M., eds.), pp. 571-582, Alan R. Liss inc., New York.

27 Daemen, T. (1987) Ph. D. Thesis, State University Groningen, The Netherlands.

28 Bouwens, L. and Wiss¢, E. (1988) Hepatology 8, 46-52.