Interaction of liposomes with human leukocytes in whole blood

Biochimica et Biophysica Acta, 762 (1983) 119-127

119

Elsevier Biomedical Press

BBA 11113

INTERACTION OF L I P O S O M E S WITH HUMAN LEUKOCYTES IN W H O L E BLOOD STEPHANUS H. K U H N , BEATRICE GEMPERLI, E N I D G. SHEPHARD, JAMES R. JOUBERT, PIETER A.C. W E I D E M A N N , G E R A L D WEISSMANN a and MORRIS C. F I N K E L S T E I N b,.

Faculty of Medicine, University of Stellenbosch, P.O. Box 63, Tygerberg 7505 (Republic of South Africa), h Department of Pathology, New York Medical College, Basic Sciences Building, Valhalla, N Y 10595 and ~ Department of Medicine, New York University Medical Center, New York, N Y 10016 (U.S.A.) (Received June 2nd, 1982)

Key words: Liposome," Cell-liposome interaction; Leukocyte; (Blood)

The uptake of multilamellar liposomes into human leukocytes in whole blood in vitro was evaluated on the basis of the cellular association of liposomal markers (3H-labelled cholesterol, lipid phase; [14C]inulin, aqueous phase). The entry of liposomes into human blood leukocytes was linear for 60 min and was mediated by a saturable mechanism displaying affinity constants of 0.28 + 0.17 and 0.16 + 0.05 mM liposomal lipid (~ + S.E.) for liposomal lipid and aqueous phase markers, respectively. Amicon filtration analysis of incubation mixtures containing blood and liposomes (phosphatidylcholine: dicetyl phosphate: cholesterol, 70: 20: 10) showed that 34% of [14Clinulin was lost (neither liposome-associated nor cell-associated) after 60 min. By preincorporating sphingomyelin (35 mol%) into multilamellar liposomes, the leakage of the model aqueous phase marker inulin was reduced to 8% after 60 min, thus enhancing the drug carrier potential of liposomes in blood. As a consequence of their interaction with liposomes, the polymorphonuclear leukocytes in whole blood decreased in apparent buoyant density, while maintaining their viability. These results indicate that blood leukocytes in their natural milieu of whole blood are capable of interacting with, and taking up multilamellar liposomes.

Introduction

Liposomes may serve to carry drugs and other physiologically active molecules to cellular targets. The liposome-entrapped materials, which are present within either the aqueous compartment or the lipid membranes, may be internalized by appropriate cells following endocytosis of the liposome by the target cell or liposome-cell fusion [1,2]. Previous studies have shown that the phagocytic leukocytes isolated from the blood of dogfish [3], human patients with Tay-Sachs disease [4] and normal human donors [5] are capable of internaliz-

* To whom correspondence should be addressed. 0167-4889/83/0000-0000/$03.00 © 1983 Elsevier Biomedical Press

ing multilamellar liposomes via phagocytosis. In our most recent report [5], dual-radiolabelled liposomes (3H-labelled phosphatidylcholine, lipid phase, [14C]inulin, aqueous phase) were used to demonstrate that liposome uptake by blood leukocytes in vitro was linear for 15 rain, and was mediated by an active process that appeared to be both energy-dependent and surface-dependent. The uptake of multilamellar liposomes by leukocytes was saturable and displayed an affinity constant of 1.1-1.7 mM liposomal lipid. Ultrastructural analyses of leukocytes after incubation with liposomes (which contained the cytochemical marker, horseradish peroxidase) revealed that monocytes internalized several-fold more liposomes than did neutrophils, on a per cell basis; no

120

uptake by lymphocytes could be detected. Moreover, the liposomes were always observed within the leukocytes rather than adherent to the outer plasma membrane. The precoating of liposomes with a high molecular weight fraction of heat-aggregated human immunoglobulin G slightly enhanced their subsequent uptake by human mononuclear leukocytes, under in vitro conditions in the absence of serum. These aggregates are a surrogate for the immunoglobulin configuration in immune complexes; upon interaction of the aggregates with the exposed liposomal bilayers, they may act as ligands for the Fc receptors of leukocytes and thus provoke endocytosis of the coated liposomes. The uptake by leukocytes of both coated and uncoated liposomes could also be substantially enhanced by adding serum (10%, v / v ) to the incubation medium. We have previously determined the extent of liposome association with isolated blood leukocytes [3,5]. Since liposomes are often systemically introduced via the intravenous route in vivo [6,7], the leukocytes in peripheral blood are exposed to liposomes before they reach the reticuloendothelial system or other potential target tissues. In this paper, we therefore report studies of the uptake of liposomes by blood leukocytes in their natural milieu of whole blood in vitro, so as to more closely mimic the environment of liposomes immediately following intravenous administration. Materials and Methods

Preparation of liposomes. Multilamellar liposomes (egg phosphatidylcholine: dicetyl phosphate : cholesterol, 7 : 2 : 1) were prepared as previously described [5], with [7(n)-3H]cholesterol (New England Nuclear) as the lipid phase marker and [14C]inulin ( M r 5000) as the aqueous phase marker. The hand-shaken liposomes (15 ~mol lipid/ml) were chromatographically fractionated and the liposome peak fractions were pooled (9.3_+0.8 /~mol lipid/ml; .~ + S.D., n = 14). These liposomes entrapped 57.4-+ 7.2 #g inulin//.tmol lipid ( X -+ S.D., n = 26) within their aqueous spaces; this entrapment of soluble inulin also corresponds to 8.6 + 1.0 #1 per ~mol lipid (2 -+ S.D., n = 26). Coating of liposomes with immunoglobulins. Human immunoglobulin G was heat-aggregated and

chromatographically fractionated; a high molecular weight fraction (24-280 S sedimentation coefficient) was isolated and was then associated with the outer monolayer of preformed liposomes, as detailed previously [5].

Incubation

of liposomes

in whole blood.

Heparinized whole blood (25-ml aliquots containing 10 units heparin/ml) obtained from human volunteer donors was incubated in the presence of various concentrations of dual-radiolabelled liposomes: from 2.00 to 1000 nmol liposomal lipid, and from 115 to 57400 ng liposomal inulin per ml of blood. After 60 min of incubation at 37°C in a shaking water bath, the blood was layered onto Hypaque/Ficoll gradients and separated into polymorphonuclear and mononuclear cell fractions by the procedure of Boyum [8]. Contaminating erythrocytes were hypotonically lysed. Cells were washed once by centrifugation and then resuspended in Dulbecco's phosphate-buffered saline (pH 7.4) at a final concentration of (10-20). 106 cells/ml. Cellular viability was 95-99% as determined by exclusion of Trypan blue. Differential cell counts were made on Leishman's eosin methylene blue-stained smears. The isolated leukocytes (triplicate 200-/~1 aliquots) were washed free of unassociated liposomes for a second time by rapid centrifugation (at 8000 × g for 2 min in an Eppendorf Model 3200 microfuge) through a silicone layer (G.E. Versilube F-50; specific gravity, 1.05) and into 0.6 M sucrose (specific gravity, 1.10); the leukocytes pass through the silicone layer whereas the unassociated liposomes remain in the supernatant [5]. The extent of liposome association with these leukocytes was evaluated on the basis of the recovery of both liposomal 3H-labelled cholesterol and liposome-encapsulated [14C]inulin in the cellular fractions. Liposomal uptake was expressed in terms of ng liposomal inulin/106 cells and nmol liposomal lipid/106 cells. In the kinetic analysis, the concentration-dependent uptake rates were computer fit to the Michaelis-Menten equation using the Marquardt algorithm, a general non-linear curve-fitting procedure [9]. Analysis of liposomal latency in whole blood. Two types of multilamellar liposomes of differing composition were compared with respect to their stability in whole blood: phosphatidylcholine:di-

121

cetylphosphate : cholesterol, 70 : 20 : 10; and sphingomyelin : phosphatidylcholine : dicetylphosphate : cholesterol, 35 : 35 : 20 : 10. The method of preparation of the former has been described above, and that of the latter has been previously described [10]. The two types of liposomes were incubated at 37°C in whole blood (heparinized) at a concentration of 1.25 /~mol lipid/ml blood; the pooled liposome concentration was approx. 10 /~mol lipid/ml phosphate-buffered saline. After various time intervals (0, 15, 30, 45 and 60 min), the duplicate incubation mixtures (1-ml aliquots) were diluted with ice-cold Dulbecco's phosphatebuffered saline (5-ml aliquots), transferred to Centriflo Membrane Cones (type CF50A; Am±con Corp., Lexington, MA) and centrifuged at 300 × g for 10 min, as previously described [10]; this membrane retains greater than 95% of molecules having molecular weights in excess of 50 000. Aliquots of the liposomes added to the incubation mixtures and of the filtrates following centrifugation were analyzed for radioactivity. Results

Incubation of liposomes in whole blood." the effects of liposome concentration on leukocyte distribution on Hypaque / Ficoll gradients The leukocytes present in venous blood obtained from normal human donors were separated into a mononuclear cell fraction (96 _+ 4% mononuclear leukocytes) and a polymorphonuclear fraction (96 _+ 4% polymorphonuclear leukocytes), as presented in Table I. Furthermore, the total num-

ber of cells recovered ((4.4 + 2.3). l06 leukocytes per ml blood), and the differential count results obtained on whole blood smears prior to fractionation (80 -+ 10% p o l y m o r p h o n u c l e a r leukocytes; 20 + 10% mononuclear leukocytes), are in line with previously reported normal values [11 ]. In parallel experiments, blood (25-ml) was incubated together with liposomes at 37°C for 60 min, and was then separated into mononuclear and polymorphonuclear cellular fractions. The liposomal concentration in the blood was 1.25 /xmol lipid/ml blood (and 65.1 jag liposome-entrapped inulin/ml). These blood samples could not be separated into homogeneous polymorphonuclear and mononuclear cell fractions by the conventional Hypaque/Ficoll gradient sedimentation procedure. Instead, all of the leukocytes were recovered in the less-dense mononuclear cell fraction, which now was comprised of 24_+ 10% mononuclear cells and 76_+ 10% polymorphonuclear cells (Table I). The total number of leukocytes recovered following exposure of blood to iiposomes ((3.3 _+ 1.2). 10 6 cells/ml) was not significantly different from the recovery of unexposed blood ((4.4_+ 2.3). 10 6 cells/ml). In both instances cellular viability was greater than 96%, as determined by exclusion of Trypan blue dye. Next, we sought to determine whether the effects of liposomes on the density distribution of polymorphonuclear leukocytes on Hypaque/Ficoll gradients was dependent on the dose of liposomes. Whole blood was incubated with liposomes at concentrations ranging from 1.4 to 1250 nmol lipid/ml blood (and also corresponding to 104 to

TABLE 1 ISOLATION O F P O L Y M O R P H O N U C L E A R (PMN) A N D M O N O N U C L E A R (MN) L E U K O C Y T E S F R O M B L O O D BY HYPAQUE/FICOLL GRADIENT SEDIMENTATION At zero-time leukocytes were isolated from whole blood by Hypaque/Ficoll gradient sedimentation (Y ± S.D., n = 15). The leukocytes were later isolated from whole blood by Hypaque/Ficoll gradient sedimentation after whole blood was incubated with liposomes (1.25/~mol lipid per ml) (phosphatidylcholine: dicetylphosphate: cholesterol, 70:20: 10) for 60 min at 37°C (~ ± S.D., n = 15). Preincubation of blood with liposomes (min)

0 60

MN-fraction

PMN-fraction

MN

PMN

MN

PMN

(9~)

(%)

(%)

(%)

96± 4 24±10

4± 4 76±10

4±4 0±0

96±4 0±0

Total cells recovered (cells/ml blood)( × l0 6)

4.4±2.3 3.3±1.2

122 T A B L E II T H E EFFECTS OF T H E C O N C E N T R A T I O N OF LIPOSOMES ON T H E DENSITY D I S T R I B U T I O N OF POLYM O R P H O N U C L E A R L E U K O C Y T E S F O L L O W I N G LIPOSOME-WHOLE BLOOD INCUBATION, AS DET E R M I N E D BY H Y P A Q U E / F I C O L L G R A D I E N T SEDIMENTATION Aliquots (25 ml) of blood were incubated with dual-radiolabelled liposomes (phosphatidylcholine : dicetylphosphate : cholesterol, 70 : 20 : 10) at various concentrations. After 60 min of incubation at 37°C, the blood was layered onto Hypaque/Ficoll gradients, separated into polymorphonuclear and mononuclear cell fractions, and the number of polymorphonuclear cells contaminating the mononuclear fractions was determined on the basis of differential cell counts. The maxim u m number of polymorphonuclear cells expected to contaminate the mononuclear-cell fraction was 80%, since the whole blood initially contained 80+ 10% polymorphonuclear cells and 20 4- 10% mononuclear cells. Liposome concentration Lipid phase (nmol lipid/ml blood

Aqueous phase (ng i n u l i n / m l blood

Polymorphonuclear cells recovered in the mononuclear cell fraction (% of cells)

1.4__+ 0.2 2.9_+ 0.4 4.4__+ 0.6 6.4__+ 0.9 8.5__+ 1.2 10.3+_ 1.4 1250 __+200

104__+ 9 207__+ 18 320__+ 29 351-+ 30 432-+ 39 525_+ 47 65070__+130

5__+ 4 ( 3 ) 15__+ 5 (3) 52__+25 (3) 55_+15(3) 76_+ 4(3) 80_+ 6(3) 80__+ 6(8)

Time course of inulin release from multilamellar liposomes in the presence of whole blood Multilamellar liposomes consisting of phosphatidylcholine : d i c e t y l p h o s p h a t e : cholesterol (70 : 20 : 10) displayed a time-dependent release of previously entrapped inulin. Inulin release from liposomes incubated in whole blood increased from 0% at time zero, to 28% after 15 min and finally to 34% after 60 min (Fig. 1). By contrast, multilamellar liposomes comprised of sphingomyelin : phosphatidylcholine : dicetylphosphate : cholesterol (35 : 35 : 20 : 10) released only 5% of entrapped inulin within the first 15 min of incubation in whole blood at 37°C. This value reached a plateau of 9% by 30-60 min (Fig. 1). Time course of liposome uptake and free inulin uptake by leukocytes in whole blood The time courses for the cellular association of the individual radiolabels (3H-labelled cholesterol, lipid phase; [14C]inulin, aqueous phase) of dual-

P

I

I

c

I

2 $ ¥

65 070 ng liposome-encapsulated inulin/ml blood). As seen in Table II, exposure of blood to 1.4 nmol liposomal lipid/ml blood had no effect, whereas concentrations of liposomes of 2.9 nmol lipid/ml or greater caused a progressive shift of polymorphonuclear leukocytes from the high-density fraction to the lower density mononuclear cell fraction. Liposomes at a concentration of 10.3 nmol l i p i d / m l in the blood caused a total shift of the polymorphonuclear cells to the mononuclear cell fraction in the subsequent fractionation procedure; with such a total shift, no leukocytes were recoverable in the conventional polymorphonuclear cell fraction. Therefore, all subsequent uptake data at liposome concentrations exceeding 10.3 n m o l / m l refers to uptake by unfractionated leukocytes.

1

4O

T

30

E

o~ "6

10

J 0

15

J

J

L

30

45

6O

Time (minutes) Fig. 1. Time courses of [14C]inulin release from muhilamellar liposomes comprised of phosphatidylcholine : dicetylphosphate : cholesterol (70 : 20 : 10) (O) and of sphingomyelin : phosphatidylcholine: dicetylphosphate: cholesterol (35 : 35 : 20 : 10) (O). Each point represents the m e a n + S . E , of four experiments, each performed in duplicate.

123 TABLE IIl UPTAKE OF LIPOSOMES AND FREE INULIN BY LEUKOCYTES IN WHOLE BLOOD AS A FUNCTION OF TIME Aliquots (25 ml) of blood were incubated with either dual-radiolabelled liposomes or unentrapped inulin. The liposomes (phosphatidylcholine: dicetylphosphate: cholesterol,. 70:20: 10) were either uncoated or coated with high molecular weight aggregates of human immunoglobulin G at 100/~g/ml. The total amounts of liposomal lipid and liposome-entrapped inulin incubated per ml of whole blood equalled 1250-+200 nmol lipid and 65100+ 1400 ng inulin (.~ -+S.D., n = 4), respectively. In parallel incubations, blood was exposed to free (unentrapped) inulin at concentrations equalling that entrapped within liposomes. The cell association of liposomal markers and of free inuhn was determined after 15, 30 and 60 min of incubation at 37°C and was expressed as nmol lipid/106 leukocytes and ng inulin/106 leukocytes (~ + S.D., n = 4). Substrate

Uncoated liposomes Aggregated IgG-coated liposomes Free inulin

15 rain

30 min

60 min

Lipid

Inulin

Lipid

Inulin

Lipid

Inulin

1.3+0.2

50+ 8

2.5 + 0.3

106 + 10

4.1 + 0.3

200 -4-_15

1.2 +_0.1 0 +_0

50 _+10 5+_ 3

2.3_+0.2 0 _+0

105_+15 6-+ 4

3.6_+0.4 0 _+0

180_+20 15+_ 4

radiolabelled liposomes were parallel (Table III). T h e uptake of both liposomal markers was linear for 60 min. As a control for uptake of liposome-encapsulated inulin via a pinocytotic mechanism independent of liposomes as a vehicle, parallel incubations were performed with free, unentrapped inulin in whole blood. The free inulin was present at the same concentration as when carried within liposomes. As can be seen in Table III, free inulin entered the leukocytes in whole blood, but much less efficiently than when encapsulated in liposomes. The extent of inulin uptake by leukocytes over the first 15 min was enhanced by 10-fold over the control value when the inulin was encapsulated within liposomes; likewise, uptake of encapsulated inulin was 18- and 13-fold greater than of free inulin after 30 min and 60 min, respectively. The free unentrapped inulin presumably entered the leukocytes in whole blood via pinocytosis. The coating of dual-radiolabelled liposomes with high molecular weight aggregates of h u m a n immunoglobulin G prior to incubation with whole blood did not influence the extent of liposomal marker uptake into leukocytes (Table III). The ratio of cell-associated liposomal markers (ng i n u l i n / n m o l liposomal phospholipid) was 39 after 15 rain of incubation, 42 after 30 min and 49 after 60 rain of incubation of liposomes with whole blood, as c o m p a r e d to 52 (2, n = 4) for the origi-

nal liposomal suspension employed (Table III). The similarity of the 'cell-association' ratios to the original ratio in the liposome preparation suggests that the liposomes were intact at the time of their association and internalization by the cells. These observations are compatible with liposome uptake by leukocytes in whole blood via a phagocytic mechanism.

Kinetics of liposomal entry into leukocytes in whole blood U p t a k e of multilamellar liposome after 60 min was determined at multiple concentrations of liposomes ranging from 2.0 to 1000 nmol liposomal p h o s p h o l i p i d / m l whole blood; the a m o u n t of inulin encapsulated within liposomes corresponded to 115-57 400 ng liposomal i n u l i n / m l whole blood. W h e n the cellular uptake of both liposomal [ 14C]inulin and 3H-labelled cholesterol was plotted as a function of the concentration of liposomes (Fig. 2), the curves indicated that saturation was being approached at the highest concentrations (more than 200 nmol l i p i d / m l blood) with respect to both liposomal markers. This was confirmed by c o m p u t e r fitting the data to the Michaelis-Menten equation for a single c o m p o n e n t [9]. By such an analysis, kinetic parameters were obtained for liposomal lipid uptake in which K m equalled 280 + 170 nmol l i p i d / m i and V,~axequalled 3.83 _ 0.91 nmol lipid/106 cells per h (2 + S.E.). Similarly,

124 I

1

I

I H

80 =

I 0

0

6O ~20 E

40

20 o=i= I

l

I

J

250 500 750 1,000 Liposome ConcenfrGfion (nmol Lipid/ml) Fig. 2. Concentration dependence of liposomal uptake by human leukocytes in whole heparinized blood. Aliquots of blood were incubated for 60 rain with 9 concentrations of uncoated, doubly radioactively labelled liposomes (phosphatidylcholine : dicetylphosphate: cholesterol, 70 : 20 : 10) ranging from 2.0 to 1000 nmol lipid/ml blood, e, rate of uptake of liposomal aqueous phase marker [14C]inulin (ng/106 cells per h); O, rate of uptake of liposomal lipid phase marker 3H-labelled cholesterol (nmol liposomal lipid/106 cells per h). Each plotted value represents the mean of three experiments. The curves were drawn on the basis of a computer fit of the data to the Michaelis-Menten equation using the Marquardt alogrithm, a general non-linear curve fitting procedure [9].

which was 7 in the original preparation), decreased gradually from 8 to 6 as the liposome concentration during incubation decreased from 1000 to 300 nmol lipid/ml, respectively; the dose curve then dropped precipitously in the range of 50-150 n m o l / m l blood so that at 10 nmol/ml, the ratio of phosphatidylcholine to cholesterol was 2.7. This data (not shown) was successfully fit to the power function y = 1.591x °2315 with a correlation coefficient r = 0.94, where y equals the molar ratio of liposomal phosphatidylcholine to cholesterol becoming associated with cells and x equals the liposome concentration (nmol lipid/ml blood). With the use of this equation and the actual 3Hlabelled cholesterol uptake values (Fig. 2), the presumptive phosphatidylcholine uptake values were calculated and were also computer fit to the Michaelis-Menten equation (as in Fig. 2). The presumptive phosphatidylcholine uptake data yielded a K m of 650 _+ 500 nmol total lipid/ml and a V,,,ax of 5.44 _+ 2.16 nmol lipid/106 cells per h ( .~ + S.E.). While these projected values are 132 and 42% greater than those experimentally determined using 3H-labelled cholesterol as the lipid phase marker, they are not significantly different. Discussion

the data for liposomal inulin uptake yielded a K m of 164 + 50 nmol lipid/ml and a Vmax of 103 + 11 ng inulin/106 cells per h ( £ + S.E.). These data indicate that multilamellar liposomes associate with leukocytes in blood via a saturable mechanism. The percentage of the total liposomal inulin taken up by leukocytes in blood during the 1 h incubation decreased significantly from 25.0 to 0.48% as the liposome concentration was raised from 2.0 to 1000 nmol lipid/ml blood, respectively (Fig. 2). In a separate set of experiments (n = 4), we sought to determine the ratio of liposomal phosphatidylcholine to cholesterol that becomes cell-associated. Dual-radiolabelled multilamellar liposomes ([3H]phosphatidylcholine : dicetylphosphate : [14C]cholesterol, 70 : 20 : 10) were incubated in whole blood at four liposome concentrations: 10, 330, 660 and 1000 nmol liposomal lipid/ml. The molar ratio of phosphatidylcholine to cholesterol which became cell-associated (and

The present study demonstrates that multilamellar liposomes may be taken up by leukocytes in whole blood; the presence of plasma proteins and formed elements besides leukocytes did not interfere with the internalization of liposomes, which proceeded linearly with time up to 60 min. The time course and extent of uptake of liposomal markers by leukocytes in whole blood (Table II D, on a per cell basis, was similar to previous observations (Ref. 5, Fig. 1) on liposomal uptake by isolated leukocytes in the presence of serum (10%, v/v). Liposomes, utilized as a vehicle for delivery of model solute inulin, enhanced inulin uptake by the leukocytes in blood by greater than 10-fold as compared to uptake of free inulin. From the dose-dependence experiment (Fig. 2), we have determined that the fraction of total liposomes taken up by leukocytes was inversely related to the concentration of liposomes added to the blood: incubations in which the liposome concentration was low (2.0 nmol lipid/ml blood) re-

125 suited in 12.0% of the total liposomal lipid present being taken up by the leukocytes; at 1000 nmol liposomal lipid/ml, liposomal lipid uptake corresponded to 0.93% of the total present (Fig. 2). The kinetics revealed that the leukocytes approached saturation with respect to liposome uptake at liposome concentrations of greater than 200 nmol lipid/ml blood. One might, therefore, expect that the most efficient protocol for liposome delivery to leukocytes via an intravenous route would include infusion of low concentrations over a prolonged period of time, so as not to saturate the blood leukocytes at any one time. There are several limitations in quantifying liposome uptake using a single marker for either the aqueous phase or the lipid phase [1,2,5]. Firstly the cellular association of a liposomal aqueous phase marker may represent the uptake by pinocytosis of leaked marker, as well as true internalization of intact liposomes. Secondly, the cellular association of a liposomal lipid phase marker may represent liposome-cell membrane lipid exchange, which may occur at different rates for each type of constituent lipid present, as well as true internalization of intact liposomes by either endocytosis or fusion. Because of these limitations in the interpretation of uptake data for single markers, we have used doubly-radiolabelled liposomes in this study. Indeed, we observed that both liposomal markers ([14C]inulin and 3H-labelled cholesterol) (i) associated with the blood leukocytes with parallel time courses (Table III), and the ratio of inulin to lipid becoming cell-associated was similar to the ratio in the original liposome suspension employed, and (ii) displayed similar affinity (Km) constants (Fig. 2). These observations are compatible with the uptake of intact liposomes by leukocytes in whole blood. In an earlier study [5], it was shown that the coating of liposomes with high molecular weight IgG aggregates enhanced their uptake in the absence but not in the presence of serum. The present findings indicate that the uptake of liposomes by leukocytes in whole blood is not enhanced by prior coating with immunoglobulins (Table III). It is possible that serum factors associate with liposomes (either uncoated or pre-coated), saturate their surface and thereby facilitate their entry into leukocytes in whole blood; the enhancement effect

of immunoglobulin coating observed previously in the total absence of serum [5] is overshadowed by the greater enhancement of liposome uptake mediated via plasma proteins in whole blood. The polymorphonuclear leukocytes in whole blood were seen to decrease in their apparent buoyant density as a result of their interaction with multilamellar liposomes (Table I). The fraction of polymorphonuclear leukocytes affected was directly related to the concentration of liposomes present (Table II). Possible physical changes in the polymorphonuclear leukocytes which may account for the observed shift in the density distribution on Hypaque/Ficoll gradients include alterations in cell density, as well as changes in cell volume, cell shape, intercellular aggregation and degree of membrane rigidity or deformability [12,13]. The Hypaque/Ficoll gradient method employed did not permit conclusions as to whether or not mononuclear leukocytes were also altered in their apparent buoyant density. The isolated neutrophils remained viable (more than 95%) despite these alterations as were the mononuclear leukocytes isolated after incubation with liposomes. A major concern in the use of liposomes as a therapeutic vehicle in vivo is their stability in the bloodstream. It is now well established that upon intravenous administration, the liposomal membranes are disrupted to some extent, and the liposomal contents are released. Through experiments in vitro in which liposomes of various compositions were incubated with blood, plasma, serum or isolated plasma proteins, it has been determined that the components of plasma which contribute to the destablizing effect include high-density lipoproteins [14-16] and complement system activation; other lipoproteins and unidentified non-lipoproteins [14,15] may be involved secondarily in the mode of action of high-density lipoproteins. We [10] and others [15-20] have shown that the destabilization of liposomes in serum and plasma can be minimized by incorporating a high proportion of cholesterol into the bilayers. Cholesterol presumably renders the liposomal membranes less vulnerable to attack by plasma proteins and subsequent loss of their phospholipids to high-density lipoproteins. Similarly, we [10] and others [15,19] have observed that the substitution of sphingomyelin for phosphatidylcholine also stabi-

126 lizes the liposomes in serum and plasma. Indeed, Gregoriadis and coworkers have demonstrated in vivo that small unilamellar liposomes either composed of sphingomyelin [19] or having a high cholesterol content [17,20] display a prolonged solute survival in circulation in animals, thus influencing the tissue distribution of the entrapped solute. The literature in this area of liposomology has been expertly reviewed recently by Scherphof et al. [21]. The integrity of liposomes in the blood is a major prerequisite to their use as drug carriers as they must provide for the direct delivery of their contents to target cells. We have previously shown that liposomes are susceptible to lysis in the presence of h u m a n serum or plasma; the extent of serum-induced damage could be minimized by preincorporating either sphingomyelin (35 mol%) or cholesterol (25 mol%) into the liposomal bilayers [10]. In 50% ( v / v ) h u m a n serum (the approximate serum concentration in whole blood), the rate of inulin leakage from multilamellar liposomes comprised of phosp h a t i d y l c h o l i n e : d i c e t y l p h o s p h a t e : cholesterol ( 7 0 : 2 0 : 10) was 57% per h, and 13% per h for those comprised of s p h i n g o m y e l i n : p h o s p h a t i d y l choline : dicetylphosphate : cholesterol (35 : 35 : 20: 10) [10]. In the present study, a significant fraction of the liposomal cargo was lost (34% of entrapped i n u l i n / h ) from multilamellar liposomes ( p h o s p h a t i d y l c h o l i n e : d i c e t y l p h o s p h a t e : cholesterol, 7 0 : 2 0 : 10) incubated in h u m a n whole blood (Fig. 1). It is interesting that the extent of blood-induced leakage (34% per h) (Fig. 1) is significantly less than the serum-induced leakage, previously determined by us to be 57% per h [10]. O u r results are compatible with the finding in other recent studies in which liposome stability was greater in the blood than in the serum of rats [20] and mice [17]. Gregoriadis and coworkers [17,20] have suggested that the presence of erythrocytes reduces liposomal deterioration either by (i) donating cholesterol to the liposomes and thus stabilizing them, or (ii) being involved in interactions with lipoproteins which take precedence over potentially disruptive lipoprotein-liposome interactions. In this study, the blood-induced d a m a g e to liposomes was reduced significantly (to 9% leakage of entrapped i n u l i n / h ) when lipos o m e s o f a m o d i f i e d lipid c o m p o s i t i o n

(sphingomyelin : phosphatidylcholine : dicetylphosphate : cholesterol, 35 : 35 : 20 : 10) were used (Fig. 1). Sphingomyelin-containing liposomes are therefore more dependable in h u m a n whole blood in vitro and perhaps offer a means of improving the efficiency of liposomal delivery to desirable tissue targets in vivo, thus minimizing the possibility of inappropriate immunological or physiological responses to the entrapped solutes.

Acknowledgements This study was aided by a grant from the American Lung Association to M.C.F., grants (AM-11949, HL-19721) from the National Institutes of Health and the Arthritis F o u n d a t i o n to G.W. and by the South African Medical Research Council.

References 1 Finkelstein, M.C. and Weissmann, G. (1981) in Liposomes: from Physical Structure to Therapeutic Applications (Knight, C.G., ed.), pp. 443-464, Elsevier/North-Holland, New York 2 Pagano, R.E., Schroit, A.J. and Struck, D.K. (1981) in Liposomes: from Physical Structure to Therapeutic Applications (Knight, C.G., ed.), pp. 323-348, Elsevier/NorthHolland, New York 3 Weissmann, G., Bloomgarden, D., Kaplan, R., Cohen, C., Hoffstein, S., Collins, T., Gottlieb, A. and Nagle, D. (1975) Proc. Natl. Acad. Sci, U.S.A. 72, 88-92 4 Cohen, C.M., Weissmann, G., Hoffstein, S., Awasthi, Y.C. and Srivastava, S.K. (1976) Biochemistry 15, 452-460 5 Finkelstein, M.C., Kuhn, S.H., Schieren, H., Weissmann, G. and Hoffstein, S. (1981) Biochim. Biophys. Acta 673, 286-302 6 Patel, H.M. and Ryman, B.E. (1981) in Liposomes: from Physical Structure to Therapeutic Applications (Knight, C.G., ed.), pp. 409-441, Elsevier/North-Holland, New York 7 Juliano, R.L. (1981)in Liposomes: from Physical Structure to Therapeutic Applications (Knight, C.G., ed.), pp. 391-407, Elsevier/North-Holland, New York 8 Boyum, A. (1968) Scand. J. Clin. Lab. Invest. 21 (Suppl. 97), 77-89 9 Marquardt, D.W. (1963) J. Soc. Ind. Appl. Math. 11, 431-441 10 Finkelstein, M.C. and Weissmann, G. (1979) Biochim. Biophys. Acta 587, 202-216 11 Orfanakis, N.G., Ostlund, R.E., Bishop, C.R. and Athens, J.W. (1970) Am. J. Clin. Pathol. 53, 647-651 12 Mel, H.C. (1963) Nature 200, 423-425 13 Miller, R.G. and Phillips, R.A. (1969) J. Cell. Physiol. 73, 191- 202

127 14 Damen, J., Dijkstra, J., Regts, J. and Scherphof, G. (1980) Biochim. Biophys. Acta 620, 90-99 15 Damen, J., Regts, J. and Scherphof, G. (1981) Biochim. Biophys. Acta 665, 538-545 16 Kirby, C., Clarke, J. and Gregoriadis, G. (1980) FEBS Lett. 111,324-328 17 Kirby, C., Clarke, J. and Gregoriadis, G. (1980) Biochem. J. 186, 591-598 18 Allen, T.M. and Cleland, L.G. (1980) Biochim. Biophys. Acta 597, 418-426

19 Gregoriadis, G. and Senior, J. (1980) FEBS Lett. 119, 43-46 20 Gregoriadis, G. and Davis, C. (1979) Biochem. Biophys. Res. Commun. 89, 1287-1293. 21 Scherphof, G., Damen, J. and Hoekstra, D. (1981) in Liposomes: from Physical Structure to Therapeutic Applications (Knight, C.G., ed.), pp. 299-322, Elsevier/North-Holland, New York