Studies on liposome formulations for intra-articular delivery of clodronate

journal of

controlled release

ELSEVIER

Journal of Controlled Release 35 (1995) 145-154

Studies on liposome formulations for intra-articular delivery of clodronate Jukka M 6 n k k 6 n e n a

a,* Jaana

L i u k k o n e n a, Markku Taskinen Arto Urtti a

a, Timothy D. Heath b,

Department of Pharmaceutical Technology, University of Kuopio, P.O. Box 1627, FIN- 70211 Kuopio, Finland b School of Pharmacy, University of Wisconsin, Madison, W153706, USA

Received 3 October 1994; accepted 17 February 1995

Abstract

Liposomes have been proposed as a means to target intra-articularly injected anti-inflammatory agents to phagocytic cells in inflamed synovial joints. Clodronate (dichloromethylene bisphosphonate) is a new candidate for this kind of liposomal drug therapy owing to its macrophage suppressive effects and considerably increased potency through liposome-encapsulation. We undertook an investigation to assess, in vitro, liposome formulations ofclodronate for intra-articular drug delivery. Brief exposure of macrophages to treatment increases the potency of liposome formulations relative to free drug; with 1-24 h exposure, the potency of liposomal clodronate was over 100-fold greater than that of free drug compared only a 58-fold difference with 48 h exposure, indicating that liposomes more rapidly affect the target cells than free drug. Fast and extensive release of calcein from highly negatively charged liposomes (25-100 mol% DSPG) with high internal osmotic pressure was observed in human plasma and synovial fluid, while lower surface charge density and iso-osmotic pressure of liposomes resulted in negligible leakage. Liposomes with neutral surface charge ( 100 mol% DSPC) were unable to deliver clodronate to macrophages, but inclusion of 25 mol% of DSPG in liposomes provided effective delivery of the drug regardless of internal osmotic pressure. The results indicate that the balance between liposome stability in biological fluids and effective drug delivery in cells is provided by using liposomes containing 10-25 mol% of DSPG and having an aqueous phase iso-osmotic with surrounding medium. Keywords: Liposome; Stability; Synovial fluid; Clodronate; Intra-articular therapy

1. Introduction

Liposomes have proved to be effective carriers of clodronate (dichloromethylene bisphosphonate) and other bisphosphonates to macrophages, which do not readily internalize these compounds in the free form [ 1-3]. High concentrations of free bisphosphonates are required to inhibit the growth of R A W 264 mac* Corresponding author. Tel.: + 358-71-162489; Fax: + 358-71162456; e-mail [email protected]. 0168-3659/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved S S D I 0 1 6 8 - 3 6 5 9 ( 9 5 ) 0003 1-3

rophages in vitro, but the encapsulation of the drugs in negatively charged unilamellar liposomes enhances their potency by a factor of 20-200. In vitro, the effects of liposomal bisphosphonates are specific for highly endocytotic cells (macrophages, C V I - P ) , when large ( 1 5 0 - 2 0 0 nm in diameter) unilamellar vesicles are used [3], while smaller (extruded) liposomes are also effective for clodronate delivery to other cell types, for example fibroblasts [2]. In vivo, clodronate encapsulated in multilamellar vesicles ( M L V ) eliminates macrophages from spleen, liver, lungs, and lymph nodes of

146

J. MiinkkOnen et al. / Journal of Controlled Release 35 (1995) 14~154

mice, when administered via the appropriate routes [4]. The macrophage suppressive effects of liposomal bisphosphonates suggest that they might be useful in the treatment of inflammatory diseases, like rheumatoid arthritis [ 1,5,6], and autoimmune uveitis [7], where macrophages are involved in pathological processes [ 8-10]. This hypothesis is strongly supported by the findings that bisphosphonates, especially liposomal clodronate, inhibits effectively the secretion of proinflammatory cytokines from activated macrophages in vitro [11, Pennanen et al., manuscript submitted]. Moreover, the potential use of clodronate for the treatment of rheumatoid arthritis has been confirmed by the demonstration that clodronate possesses antiarthritic activity in animal models of arthritis [ 12]. Both intra-articular injection of clodronate in MLV [6] and intravenous administration of clodronate in small unilamellar vesicles (SUV) [ 13 ] or MLV [ 5 ] have been suggested for the treatment of rheumatoid arthritis. Although intravenously injected SUV accumulate at inflamed joints both in animal models of arthritis [14] and patients with rheumatoid arthritis [ 15], and enhance somewhat the uptake of clodronate in the inflamed rat paw [ 13], they most probably still face two major problems limiting the usefulness of liposomes in systemic administration: rapid clearance of i.v. injected liposomes by the reticuloendothelial system, and poor accessibility of the target cells owing the natural anatomic barriers [ 16]. Especially the reticuloendothelial uptake of liposomes would be expected to lead to unwanted side-effects due to depletion of splenic and hepatic macrophages by liposomal clodronate [4]. Administration of liposomes intra-articularly should avoid the above mentioned problems, and is of interest in the treatment of rheumatoid arthritis. Moreover, liposomes have been proposed to prolong the retention time of drugs in synovialjoint cavity [ 17 ]. In order to optimize the liposomal formulation for local use, liposomes with appropriate size, surface properties, and composition should be selected with respect to the site of administration, the disease, and the drug used [ 17], and a balance is needed between the stability of the liposomes and their ability to deliver drug. In order to control better the intra-articular delivery of liposomal bisphosphonates, we evaluated the time-course of the effects of clodronate on the macrophages, the leakage of aqueous content from var-

ious types of liposomes in cell culture medium, plasma, and synovial fluid, and the delivery properties of 'nonleaky' liposomes in vitro.

2. Experimental 2.1. Materials

Clodronate was obtained from Leiras Pharmaceutical Co. (Tampere, Finland). Calcein, MTT (3,(4,5)dimethylthiazol-2-yl) 2,5-diphenyltetrazolium bromide), LPS (E. coli, serotype O 127:B 8), and cholesterol were purchased from Sigma Chemical Co. (St. Louis, MO), SDS (sodium dodecyl sulphate) from Merck (Darmstadt, Germany), and DMF (N,N-dimethylformamide) from Fluka Chemie AG (Buchs, Switzerland). Distearoylphosphatidylglycerol (DSPG) and distearoylphosphatidylcholine (DSPC) were obtained from Orion Farmos Co. (Turku, Finland). Dulbecco's modified Eagle's medium (DMEM), 10 000 units/ml penicillin and streptomycin, and fetal bovine serum were from Gibco (Grand Island, NY). Sephadex G50 was purchased from Pharmacia LKB (Uppsala, Sweden). All other reagents were from various suppliers and were reagent grade or better. 2.2. Preparation o f liposomes

Stock solutions of clodronate at the concentrations of 200 mM (hyper-osmotic solution, osmotic pressure ~ 600 mosm/kg) and 110 mM (iso-osmotic solution, ~ 290 mosm/kg) were prepared as described earlier [ 1 ]. A 60 mM calcein solution was prepared either in deionized water (iso-osmotic solution) or in 75 mM NaCI, 50 mM morpholinoethanesulfonic acid, 50 mM hydroxyethylpiperazine-N'-2-ethanesulfonate (MES/ HEPES, pH 7.2) (hyperosmotic solution, ~600 mmosm/kg). The pH of the solutions was adjusted to 7.02 with sodium bicarbonate powder. Liposomes containing clodronate or calcein were prepared by reversephase evaporation (REV) [18] from phospholipid/ cholesterol (67:33) as described [1] and will subsequently be referred to by phospholipid content only. Clodronate encapsulated in liposomes was analyzed spectrophotometrically [3], and phospholipid content by phosphorus assay [ 19]. The size of calcein containing liposomes was measured by Nicomp Submicron

J. MOnkkrnen et aL / Journal of Controlled Release 35 (1995) 145-154

Particle Sizer utilizing laser light scattering (model 370, Nicomp, Santa Barbara, CA, USA), and they had a mean diameter of 200-300 nm. A mean diameter of various clodronate-liposomes is 160-180 nin, as reported elsewhere [ 2 ].

2.3. Growth inhibition experiments The growth inhibition properties of various clodronate liposome preparations were studied on a murine macrophage cell line, RAW 264. The cells were grown in DMEM supplemented with 10% fetal bovine serum and 100 U / m l penicillin and streptomycin in a 7% CO2 atmosphere at 37°C. Cells were plated at a density of 4X l03 cells per well in 96-well plates (Nunc, Roskilde, Denmark) and allowed to grow overnight. Quadruplicate wells were exposed to various treatments, and the cell growth was assayed 48 h later by an MqT assay [ 20,21 ]. The rationale of 'wash-out' experiments was adapted from Heath et al. [22] and Straubinger et al. [ 23 ], and were performed in the same manner as above described growth inhibition studies except that the medium was changed to drug free medium after defined periods of exposure (0.25-48 h) and were returned to culture for the remainder of the growth period.

2.4. Leakage experiments The leakage of water-soluble fluorescent dye, calcein, from liposomes of high or isotonic osmotic pressure in various biological fluids was assessed as follows. Unencapsulated dye was first removed by passage through a 1 x 30-cm Sephadex column equilibrated with MES/HEPES. The calcein release was detected in MES/HEPES buffer, cell culture medium, human plasma, or synovial fluid aspirated from the knee joints of patients suffering active rheumatoid arthritis. The synovial fluid was obtained from the Department of Rheumatology, University Central Hospital, Kuopio, centrifuged (4500 rpm, 30 min) and stored at - 2 0 ° C . All the fluids were diluted 1:4 with MES/HEPES and temperature was kept at 37°C during the experiments. Release of calcein was monitored with Perkin Elmer LS 50B Luminescence Spectrometer (England) at 494 nm excitation and 515 nm emission wavelenghts. 5/~1 of liposomes were mixed with 2 ml of appropriate fluid, leading to 5-27.5/zM of final phospholipid concentra-

147

tion, and the intensity of fluorescence was monitored continuously for 200 seconds. At the end of the experiment, 0.1 ml of 2.5% Triton X- 100 solution was added to the mixture to release the dye remaining in liposomes. The percentage of calcein leakage from liposomes was calculated with the following equation: 100 X ( F / F T - F o / F o . T )

where Fo is the fluorescence intensity in MES/HEPES buffer at time T (typically 0-5% from iso-osmotic liposomes, 5-20% from hypertonic liposomes); F is the fluorescence intensity in test fluid at time T; and Fo.T and F T are the fluorescence intensities of the respective mixtures after the addition of Triton-X.

2.5. Cytokine experiments The cytokine inhibitory action of clodronate encapsulated in liposome formulations was studied with RAW 264 cells as described [ 11 ]. Briefly, RAW 264 cells were plated at 2 x 105 cells per well in 96-well plate (Nunc), allowed to adhere, and treated with the formulations of interest. After 20 h of incubation, the cells were washed free of drugs, and induced to produce cytokines with 10/zg/ml of LPS for 24 h. Then the cell free supernatants were collected and assayed for IL-6 by a dissociation enhanced fluoroimmunoassay (DELFIA) [11].

3. Results and discussion

3.1. Effects of length of exposure on the potency of free and encapsulated clodronate One of the major problems in intraarticular drug therapy is a rapid clearance of therapeutic agents from synovial joint cavity, and liposomes have been proposed as a way to prolong the retention time of drugs [16,17,24]. However, even the use of liposomes is unlikely to engender a continuous exposure of the synovial joint cavity to drugs. Previous studies have shown that reducing the exposure time of cells to free or liposome-encapsulated drugs modulates their growthinhibitory potencies [ 22,23,25 ]. Thus, we studied the effect of shortened treatment periods on macrophage growth inhibitory properties of free and liposomeencapsulated clodronate.

148

J. M6nkkgnen et al./ Journal of Controlled Release 35 (1995) 145-154

10000 ~ : %

A

g=o|

o

10'

1

0

6

lZ 18 Z4 30 36 42 Exposure Time (h)

48

continuous, but greater than 100-fold at shorter ( 1-24 h) exposure times (Fig. 1B ), indicating that the major benefit of liposomal clodronate, compared to free drug, is established with brief exposure to macrophages. Thus, in addition to the fact that liposomes may prolong the retention time of the drug in synovial joint cavity, they may also provide a more rapid effect on the target cells. This most probably arises, because of the rapid binding and endocytosis of liposomes by macrophages resulting in the effective accumulation of intracellular clodronate. 3.2. L i p o s o m e l e a k a g e

,. 120]

,oo1_

B m

~a._ 801 II.

40

0

Exposure Time (h)

Fig. 1. The effect of exposure length on the growthinhibitorypotency of free and liposome-encapsulatedclodronate for RAW 264 cells in vitro. The cells were exposed to the drug for the specified time, washed free of drug, and returned to culture in fresh medium for the remainder of the 48 h incubation. The experiments were performed in quadruplicate and repeated twice. Liposomeswere prepared from DSPG/cholesterol (67:33) in a 200 mM clodronate stock solution; the drug/lipid ratio was 1.72 mol/mol. (C)) free clodronate; (0) liposome-encapsulatedclodronate; ( • ) the ratio of the ICs0for free clodronate relative to encapsulated clodronate. Fig. 1A shows that in exposure lengths shorter than 6 h, free clodronate did not affect the growth of R A W 264 cells (ICso was greater than the highest drug concentration used, 3 0 0 0 / z M ) . With 6 h exposure, free drug had an ICso of 2039 /xM, decreasing gradually with increasing length of exposure down to 757 /~M with continuous (48 h) treatment. In contrast, clodronate encapsulated in DSPG liposomes affected the cell growth already with 1 h exposure, and reached almost the maximum potency with 6 h treatment (Fig. 1A). The difference in potencies between free and encapsulated clodronate was 58-fold, when the exposure was

A balance between effective cellular uptake and adequate stability of liposomes is required for successful delivery of therapeutic agents to desired targets. Liposome-dependent drugs rely on the adsorptive endocytosis of the liposome for their efficient delivery to cells [ 26,27 ]. This mechanism of delivery requires that liposomes retain their contents until endocytosis has occurred. Consequently, the leakage of liposomes at the site of administration, for example in a joint cavity, would be expected to reduce drug delivery, and should be minimized in order to optimize targeting of therapeutic agents to phagocytic cells. However, liposome formulations for intra-articular therapy have not received much attention, although the targeting of antiinflammatory agents to local phagocytic cells in the treatment of osteoarthritis and rheumatoid arthritis has been suggested to be of great clinical interest [6,17,24,38]. Liposome surface charge density is one important factor contributing the stability and cellular uptake of liposomes. A negative liposome surface charge induces the leakage of encapsulated material in biological fluids, for example serum [28,29], and makes vesicles susceptible to calcium-induced fusion and release of encapsulated material [30]. However, a negative liposome surface charge is also a prerequisite for effective uptake of the vesicles by cells [2,25,27,31,32]. Consequently, in order to optimize drug delivery, it is necessary to establish the negative surface charge density that induce minimal leakage, while still permitting effective drug delivery. Another factor that we have also addressed is the osmotic activity of the internal aqueous contents. The use of hyperosmotic internal aqueous contents can render liposomes less stable

149

J. Mrnkkrnen et al. / Journal of Controlled Release 35 (1995) 145-154

[30,33]. However, we have previously encapsulated clodronate solutions whose osmotic activity is greater than 290 m o s m / k g (about 600 m o s m / k g ) , because it allows us to increase the drug:lipid ratio to more favourable level [ 1,3 ]. Therefore, we examined the effects of liposomal surface charge density, and the osmotic pressure of the internal aqueous contents on the leakage of a water soluble fluorescent dye, calcein, from liposomes in cell culture medium, human plasma, and synovial fluid from the knee joints of rheumatic patients. Fig. 2 shows the calcein release from liposomes prepared from various mixtures of DSPG and DSPC and containing a hyperosmotic solution ( ~ 600 m o s m / k g ) in the internal aqueous space. Calcein was released rapidly from the liposomes prepared from 100 mol/ 100 mol DSPG in all fluids tested, being the most pronounced in synovial fluid (Fig. 2C), where about 80% of calcein was released within first 2 min. Calcein leakage reached a plateau during the next 200 s, except from 25-100 mol/100 mol DSPG liposomes in plasma (Fig. 2B). The reduction of DSPG content to 0-10 mol/100 mol in the liposomes decreased the leakage to about 20% in all fluids, while liposomes with 25 mol/100 mol DSPG still leaked considerably more in plasma and synovial fluid than those with lower DSPG content. To evaluate the role of calcium-dependent processes in the leakage of calcein, we added 4 mM EDTA in the test fluids before the leakage experiments, which decreased the leakage from liposomes containing 2 5 100 mol/100 mol DSPG the to same level as seen with liposomes with a lower negative charge (Fig. 3 A - B ) , indicating that most of the leakage from negatively charged liposomes was caused by calcium or calcium dependent processes, such as the complement pathway. E D T A is known to chelate divalent cations such as C a 2 + and Mg 2+ . These cations are able to induce the liposome fusion and consequently, the leakage of lipo-

some contents [ 30,34,35 ]. Moreover, C a 2 + and Mg 2 + are needed for complement activity in plasma, which also detabilizes liposomal vesicles [36]. The time course of calcein release observed in this study is not, however, consistent with the time course of liposome lOO

A

=. 80 II e

60. C

i

O

40-

.~

:

=

O

z0-

:_. ...............

0

T

=

•

•

•

0

I

"

"

"

50

"

I

"

•

"

"

100 Time (s)

l

"

=

"

•

150

|

200

I O0" B •.

80.

-~ 6 0 -

"~ 40" ll~ 2 0 -

0

T

"

•

"

"

0

I

•

"

50

"

"

I

•

•

•

•

100 Time (s)

I

•

"

"

I

200

1 O0" •

•

150

C

_.____--4

80=

e

"~60" Fig. 2. Calcein release from liposomes prepared with hyper-osmotic (600 mosm/kg) internal aqueous contents, prepared from various mixtures of negatively charged DSPG and neutral DSPC in (A) cell culture medium, (B) human plasma, and (C) synovial fluid. (0) 100:0 DSPG/DSPC; ((3) 25:75 DSPG/DSPC; ( • ) 10:90DSPG/ DSPC; (11) 0:100 DSPG/DSPC. Liposomes were prepared from phospholipid/cholesterol (67:33) by reverse-phase evaporation. Data points are the mean of four separate experiments, the deviation from average values is less than 20%.

e-

"~

40 =

ll~ 20=

0

T

o

=

=

•

"

i

50

"

"

"

"

i

•

•

I oo Time (s)

"

"

II

I 5o

•

=

•

"

II

2oo

150

J. M6nkk&Ten et al. / Journal of Controlled Release 35 (1995) 145-154 100

Q II

80

II qll

60 = C

40=

0 U

20 =

0

'IF

"

"

"

"

o

I

•

•

•

=

I

•

"

•

•

v i"i m ° ° e' ,~

so

II

•

•

=

I

200

100 •

•

1so

B

80

= U

-~ 6 0 c "~ 40"

u

T

.

.

.

.

I

0

.

.

.

.

50

I

.

.

.

.

I

100 T i m e (ll)

.

.

.

.

vesicle content occurring during the fusion and subsequent collapse of fused liposomes [ 34,35 ]. Further, the present results are in agreement with reports that calcium induced fusion and leakage of negatively charged vesicles is strongly inhibited by phosphatidylcholine [34]. Thus, it seems obvious that the massive leakage of aqueous marker from highly negatively charged liposomes with high osmotic pressure is due to calcium (and magnesium) induced fusion phenomenon. The remaining 20-30% leakage (Fig. 3A-B) was caused by high internal osmotic pressure, since the reduction of osmotic pressure inside the liposomes to the iso-osmotic level ( ~ 2 9 0 mosm/kg) decreased considerably the leakage of liposomes in cell culture medium, only liposomes with 100 mol/100 mol DSPG showing any notable release of calcein (Fig. 4A). In plasma (Fig 4B), liposomes with 25-100 mol/100 mol DSPG, and in synovial fluid (Fig. 4C), liposomes with 100 mol/100 mol DSPG still showed a substantial leakage. This was, however, abolished by 4 mM EDTA (data not shown), indicating that the high calcium concentration present in plasma and synovial fluid induces the instability of vesicles with high negative surface charge even in the case of liposomes with an isoosmotic internal aqueous content.

I

150

200

3.3. G r o w t h inhibition s t u d i e s

1O0"

C

80III qP

-~ 6o: e,,

"i _o 4 0 -

A

g

20-

0

T

0

"

"

"

"

l

50

"

"

•

"

I

"

•

T i J O0 (ll)

•

"

I

•

150

"

"

•

I

200

Fig. 3. Calcein release from liposomesprepared with hyperosmotic internal aqueous contents in the presence of 4 mM EDTA. Details in all other respects identical to those in Fig. 2. degradation through the complement pathway, which characteristically exhibit 1O- 15 min lag time [ 36,37 ], but is consistent with the time-course of the leakage of

In our earlier in vitro studies, we used liposomes containing 100 mol/100 mol DSPG with hyperosmotic internal aqueous contents for the delivery of clodronate to RAW 264 macrophages in vitro [ 1-3,11 ]. Although we showed that 25 mol/100 mol DSPG in DSPG/ DSPC liposomes provides comparable delivery of clodronate in RAW 264 cells in vitro than those with 100 mol/100 mol DSPG, the latter ones were chosen for further studies because of the simplicity of liposome preparation procedure [2]. Based on the present results, however, 2-and 1.5-fold higher leakage of aqueous marker in synovial fluid and plasma compared to cell culture medium, respectively, would result in a less effective delivery of clodronate in vivo than in vitro. Since liposomes containing 0-25 mol/100 mol DSPG and iso-osmotic internal aqueous contents leaked less, and had comparable leakage properties in various fluids, we tested their capability to deliver clodronate in RAW 264 cells by growth inhibition assay.

J. M6nkk6nen et al. / Journal of Controlled Release 35 (1995) 145-154 100"

•

A

80-

el I

~-, 8 0 t ~

-i i.

151

60 °

e-

o"i 4 0 II (,I

20"

o

OT

. . . . .

•

. . . . .

so

•

. . . .

Tit°°',

l

"

"

"

"

is0

;

z0o

100B

eo-

• 80R IJ Q

"~ 40" m

20,

0

SO

100 . Time ( | )

1 SO

200

100 I

80-

60-

...-.e

J

0

25 50 75 1O0 14ol DSPG/100 Mol Phoephollpid Fig. 5. Effect of surface charge density on the IC5ovalue of liposome encapsulated clodronate. ( 0 ) liposomes containing hyperosmotic internal aqueous contents (the data partially adapted from Ref. [2 ] ) ; (©) liposomes containing iso-osmotic internal aqueous contents. Data points are the mean of four separate experiments. ( * ) indicates that the IC5ois greater than the highest drug concentration used ( 100 /zM). iso-osmotic internal aqueous contents. The results indicate that 25 m o l / 1 0 0 mol DSPG is sufficient for optimal delivery of clodronate in both types of liposomes, and the intrinsic osmotic pressure does not affect the delivery and growth inhibitory potency of clodronate. They further indicate that the leakage observed with 100 m o l / 1 0 0 mol DSPG in cell culture medium (Fig. 2A) does not increase the IC5o of clodronate compared to the liposomes that leak less (Figs. 2A and 4 A ) . However, the liposomes containing 100 m o l / 1 0 0 mol DSPG and iso-osmotic internal contents are still leaking extensively in synovial fluid (Fig. 4C), and one would expect consequently much less effective delivery of clodronate to synovial phagocytic cells by these liposomes.

40"

3.4. Cytokine studies

ill 20-

o. : 0

. - : - : . - ; ....... SO

1 O0 Time (ll)

150

; 200

Fig. 4. Calcein release from ]iposomes prepared with iso-osmotic

(290 mosm/kg) internal aqueous contents. Details in all other respects identical to those in Fig. 2. In Fig. 5, the IC5o values of clodronate are plotted as a function of molar percentage of D S P G of total liposomal phospholipid in liposomes containing hyper- or

The cytokine inhibitory action of liposomal clodronate was originally established by hyperosmotic liposomes prepared from 100 m o l / 1 0 0 mol DSPG ( [ 11 ], and Pennanen et al., manuscript submitted). To establish the suitability of 25-100 m o l / 1 0 0 mol D S P G liposomes with isotonic internal aqueous contents for cytokine inhibition by clodronate, the IL-6 secretion from R A W 264 cells was studied. As shown in Fig. 6A, clodronate encapsulated in either liposome formulation effectively inhibited the IL-6 secretion from

152

J. MOnkk6nen et al. / Journal of Controlled Release 35 (1995) 145-154

1 60

A

l

Clod/25 mol% DSPG

• []

Clod/1 O0 mol% DSPG 25 tool% DSPG 100 rnol% DSPG

140 120 100 80 60

40

1

10 Clodronate (OI4)

100

tration used for drug delivery. Only the highest lipid concentration slightly inhibited the IL-6 production (Fig. 6A). The cytokine inhibition was also not caused by the cytotoxic effects of clodronate liposomes, because the MTT-staining of the cells was not decreased at the concentrations, which strongly inhibited the IL-6 secretion (Fig. 6B ). However, at the highest drug concentration ( 100/.LM), the viability of the cells was considerably decreased by clodronate in 25 mol/100 mol DSPG liposomes, while the corresponding non-loaded liposomes were non-toxic. Clodronate in 100 mol/100 tool DSPG was as toxic as non-loaded 100 tool/100 mol DSPG liposomes, and far less toxic than clodronate in 25 tool/100 mol DSPG liposomes, further indicating the lower leakage and better delivery of clodronate by the latter liposomes.

1 O0 "

4. Conclusion

0¢-

ou 6 0 w~ 4 0 "

200

II

1

10 Clodronate

I

(/JM)

|

100

Fig. 6. Effect of clodronate encapsulated in 25:75 DSPG/DSPC (C1od/25 tool% DSPG), or 100:0 DSPG/DSPC (Clod/100 tool% DSPG) liposomeswith iso-osmoticinternal aqueous contents, or the corresponding non-loaded liposomes (25 mol% DSPG, 100 tool% DSPG ) on 1L-6production (A) and MTT staining ( B ) in RAW 264 cells in vitro. In the absence of drugs (control), the IL-6 production was 31.8+ I1.0 /Lg/ml. Data points are mean_+SD of one (nonloaded liposomes) or two (clodronate liposomes) separate experiments. LPS-induced cells, the inhibition being very similar to that with DSPG liposomes that had hyperosmotic internal aqueous contents [ l 1 ]. Clodronate in 25 mol/100 mol DSPG liposomes was slightly more potent than clodronate in 100 mol/100 mol DSPG liposomes, which is consistent with the lower leakage in cell culture medium (Fig. 4A). As shown earlier [ 11], the cytokine inhibition is not due to the phospholipid used for drug delivery, since non-loaded liposomes, instead of inhibiting IL-6 secretion, induced it at the concen-

High liposome negative surface charge and hyperosmotic internal aqueous contents can cause a marked instability of vesicles in plasma and synovial fluid, impairing their usefulness as drug carriers. This instability can be prevented by inclusion of 75-100 mol/ 100 mol of neutral DSPC in the liposome membrane, and by using internal aqueous contents that are isoosmotic with the surrounding medium. However, liposomes prepared from pure DSPC are unable to deliver clodronate to macrophages. Thus, liposomes containing 25 mol / 100 tool DSPG and 75 tool / 100 mol DSPC, and cholesterol (67:33), and iso-osmotic internal aqueous contents, would be expected to provide optimal delivery of clodronate in phagocytic cells in the synovial joint cavity, and also to give in in vitro cell culture models results that parallel the in vivo situation. Further, the present data may be helpful in optimizing the liposome formulations for intra-articular use of other anti-rheumatic drugs, such as methotrexate [38].

Acknowledgements

The authors wish to thank Mr. Juhani Toimela for skilful technical assistance. The study was financially supported by the Finnish Ministry of Education, the Technology Development Centre of Finland

J. M6nkk6nen et al. / Journal of Controlled Release 35 (1995) 145-154

(TEKES), the Academy of Finland, and Rheumatism Research Foundation, Finland.

References [1] J. M6nkk6nen and T.D. Heath, The effects of liposomeencapsulated and free clodronate on the growth of macrophage like cells in vitro: The role of calcium and iron, Calcif. Tissue Int. 53 (1993) 139-146. [2] J. M6nkk6nen, R. Valjakka, M. Hakasalo and A. Urtti, The effects of liposome surface charge and size on the intracellular delivery of clodronate and gallium in vitro, Int. J. Pharm. 107 (1994) 189-197. [3] J. M6nkk6nen, M. Taskinen, S.O.K. Auriola and A. Urtti, Growth inhibition of macrophage-like and other cell types by liposome-encapsulated, calcium-bound, and free bisphosphonates in vitro, J. Drug Target. 2 (1994) 299-308. [4] N. van Rooijen, The liposome-mediated macrophage 'suicide' technique, J. Immunol. Methods 124 (1989) 1-6. [ 5 ] R. Kinne, R. Hoppe, C. Schmidt, E. Buchner, F. N/Jrnberg and F, Emmrich, Established adjuvant arthritis can be treated by systemic application of clodronate containing multilamellar liposomes (CLO-LIPO), Clin. Rheumatol. 12 (1993) 35. [6] P,L.E.M. van Lent, L.A.M. van den Bersselaar, A.E.M. Holthuyzen, N. van Rooijen, L.B.A. van de Putte and W.B. van den Berg, Phagocytic synovial lining cells in experimentally induced chronic arthritis: down-regulation of synovitis by CLzMDP-liposomes, Rheumatol. Int. 13 (1994) 221-228. [7] M.R. Niesman, M. Ni and J.N. Bloom, Monocyte depletion reduces the severity of experimental autoimmune uveitis, Invest. Ophthal. Vis. Sci. 43 (1993) 1477. [8] G.S. Firestein and N.J. Zvaifler, How important are T-cells in chronic rheumatoid synovitis, Arthritis Rheum. 33 (1990) 768-773. [ 9 ] W.P. Arend and J.-M. Dayer, Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis, Arthritis Rheum. 33 (1990) 305-315. [ 10] B. Bresnihan, The synovial lining cells in chronic arthritis, J. Rheum. 31 (1992) 433-435. [11] J. M6nkk6nen, N. Pennanen, S. Lapinjoki and A. Urtti, Clodronate (dichloromethylene bisphosphonate) inhibits LPSstimulated IL-6 and TNF production by RAW 264 cells, Life Sci. 54 (1994) PL229-234. [ 12] C.J. Dunn, L.A. Galinet, H. Wu, R.A. Nugent, S.T. Schlachter, N.D. Staite, D.G. Aspar, G.A. Elliot, N.A. Essani, N.A. Rohloff and R.J. Smith, Demonstration of novel anti-arthritic and antiinflammatory effects of diphosphonates, J. Pharmacol. Exp. Ther. 266 (1993) 1691-11698. [ 13] J,P. Camilleri, A.S. Williams, N. Amos and B.D. Williams, Uptake of liposome clodronate compared to free clodronate in inflamed rat paws, Arthritis Rheum. 36, Suppl. 9 (1993) $213. [ 14] W.G. Love, N. Amos, I.W. Kellaway and B.D. Williams, Specific accumulation of cholesterol-rich liposomes in the inflammatory tissue of rats with adjuvant arthritis, Ann. Rheum. Dis. 49 (1990) 611-614.

153

[ 15] B.D. Williams, M.M. O'Sullivan, G.S. Saggu, K.E. Williams, L.A. Williams and J.R. Morgan, Synovial accumulation of 99mTechnetium labelled liposomes in rheumatoid arthritis, Ann. Rheum. Dis. 46 (1987) 314-318. [16] G. Poste, Liposome targeting in vivo: problems and opportunities, Biol. Cell 47 (1983) 19-38. [17] D.D. Lasic, Liposomes and inflammations, in: D.D. Lasic (Ed.), Liposomes: From Physics to Applications, Elsevier, Amsterdam, 1993, pp. 399-404. [ 18 ] F.C. Szoka and D. Papahadjopoulos, Procedure for preparation of liposomes with large internal aqueous space and high capture by reverse-phase evaporation, Proc. Natl. Acad. Sci. USA 75 (1978) 4194-4198. [ 19] G.R. Bartlett, Phosphorus assay in column chromatography, J. Biol. Chem. 234 (1959) 466-468. [20] M.B. Hansen, S.E. Nielsen and K. Berg, Re-examination and further development of a precise and rapid dye method for measurement of cell growth/cell kill, J. Immunol. Methods 119 (1989) 203-210. [21] F.E. Nargi and T.J. Yang, Optimization of the L-M cell bioassay for quantitating tumor necrosis factor a in serum and plasma, J. Immunol. Methods 159 (1993) 81-91. [22] T.D. Heath, N.G. Lopez, J.R. Piper, J.A. Montgomery, W.H. Stern and D. Papahadjopoulos, Liposome-mediated delivery of pteridine antifolates to cells in vitro: potency of methotrexate, and its a and y substituents, Biochim. Biophys. Acta 862 (1986) 72-80. [23] R.M. Stranbinger, N.G. Lopez, R.J. Debs, K. Hong and D. Papahadjopoulos, Liposome-based therapy of human ovarian cancer: parameters determining potency of negatively charged and antibody-targeted liposomes, Cancer Res. 48 (1988) 5237-5245. [24] B. Axelsson, Liposomes as carriers for anti-inflammatory agents, Adv. Drug Deliv. Rev. 3 (1989) 391-404. [25] A. Sharma, N.L. Stranbinger and R.M. Straubinger, Modulation of human ovarian tumor cell sensitivity to N(phosphonacetyl)-L-aspartate (PALA) by liposome drug carriers, Pharm. Res. 10 (1993) 1434-1441. [26] T.D. Heath and C.S. Brown, Liposome dependent delivery of N-(phosphonacetyl)-L-aspartic acid to cells in vitro, J. Liposome Res. 1 (1989-90) 303-317. [27] T.D. Heath, N.G. Lopez and D. Papahadjopoulos, The effects of liposome size and surface charge on liposome-mediated delivery of methotrexate-y-aspartate to cells in vitro, Biochim. Biophys. Acta 820 (1985) 74-84. [28] S.J. Comiskey and T.D. Heath, Leakage and delivery of liposome-encapsulated methotrexate-y-aspartate in chemically defined medium. Biochim. Biophys. Acta 1024 (1990) 307317. [29] S.J. Comiskey and T.D. Heath, Serum-induced leakage of negatively charged liposomes at nanomolar lipid concentrations. Biochemistry 29 (1990) 3626-3631. [30] J. Wilschut, Membrane fusion in lipid vesicle systems, in: J. Wilschut and D. Hoekstra (Eds.), Membrane Fusion, Marcel Dekker Inc., New York, NY, 1991, pp. 89-126.

154

J. Mfnkk6nen et al. / Journal of Controlled Release 35 (l 995) 145-154

[ 31 ] M. Stevenson, A.J. Baillie and R.M.E. Richards, Quantification of uptake of liposomal carboxyfluorescein by professional phagocytes in vitro. A flow microfluorometric study on the J774 murine macrophage cell line. J. Pharm. Pharmacol. 36 (1984) 824-830. [32] K.-D. Lee, K. Hong and D. Papahadjopoulos, Recognition of liposomes by cells: in vitro binding and endocytosis mediated by specific lipid head groups and surface charge density, Biochim. Biophys. Acta 1103 (1992) 185-197. [33] N. Oku, R. Naruse, K. Doi and S. Okada, Potential usage of thermosensitive liposomes for macromolecule delivery, Biochim. Biophys. Acta 1191 (1994) 389-391. [34] N. Dtizg~ines, J. Wilschut, R. Fraley and D. Papahadjopoulos, Studies on the mechanism of membrane fusion. Role of headgroup composition in calcium- and magnesium-induced fusion

[351

[36]

[37]

[38]

of mixed phospholipid vesicles, Biochim. Biophys. Acta 642 (1981) 182-195. H. Ellens, J. Bentz and F.C. Szoka, H ÷- and Ca2+-induced fusion and destabilization of liposomes, Biochemistry 24 (1985) 3099-3106. K. Funato, R. Yoda and H. Kiwada, Contribution of complement system on destabilization of liposomes of hydrogenated egg phosphatidylcholine in rat fresh plasma, Biochim. Biophys. Acta 1103 (1992) 198-204. H. Harashima, Y. Ochi and H. Kiwada, Kinetic modelling of liposome degratation in serum: effect of size and concentration of liposomes in vitro, Biopharm. Drug Dispos. 15 (1994) 217225. W.C. Foong and K.L. Green, Treatment of antigen-induced arthritis in rabbits with liposome-entrapped methotrexate injected intra-articularly, J. Pharm. Pharmacol. 45 ( 1993 ) 204209.