Delivery of zoledronic acid encapsulated in folate-targeted liposome results in potent in vitro cytotoxic activity on tumor cells

Journal of Controlled Release 146 (2010) 76–83

Contents lists available at ScienceDirect

Journal of Controlled Release j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j c o n r e l

Delivery of zoledronic acid encapsulated in folate-targeted liposome results in potent in vitro cytotoxic activity on tumor cells Hilary Shmeeda a, Yasmine Amitay a, Jenny Gorin a, Dina Tzemach a, Lidia Mak a, Joerge Ogorka b, Saran Kumar b, J. Allen Zhang b, Alberto Gabizon a,⁎ a b

Experimental Oncology Laboratory, Shaare Zedek Medical Center, Jerusalem, Israel Novartis Pharmaceuticals Corporation, East Hannover, NJ, USA

a r t i c l e

i n f o

Article history: Received 22 January 2010 Accepted 26 April 2010 Available online 10 May 2010 Keywords: Liposomes Folate receptor Targeting Bisphosphonates In vitro cytotoxicity

a b s t r a c t Introduction: Zoledronic acid (ZOL), a nitrogen-containing bisphosphonate, is a potent inhibitor of farnesylpyrophosphate synthase with poor in vitro cytotoxic activity as a result of its limited diffusion into tumor cells. The purpose of this study was to investigate whether liposomes targeted to the folate receptor (FR) can effectively deliver ZOL to tumor cells and enhance its in vitro cytotoxicity. Methods: ZOL was entrapped in the water phase of liposomes of various compositions with or without a lipophilic folate ligand. Stability and blood levels after i.v. injection were checked. The in vitro cytotoxic activity and cell uptake of liposomal ZOL (L-ZOL) were examined on various human and mouse cell lines. Results: All formulations were highly stable and resulted in high blood levels in contrast to free ZOL which was rapidly cleared from plasma. Non-targeted L-ZOL was devoid of any in vitro activity at concentrations up to 200 µM. In contrast, potent cytotoxic activity of folate-targeted L-ZOL (FTL-ZOL) was observed, with optimal activity, reaching the sub-micromolar range, for dipalmitoyl-phosphatidylglycerol (DPPG)-containing liposomes and relatively lower activity for pegylated (PEG) formulations. IC50 values of FTL-ZOL on FRexpressing tumor cells were N 100-fold lower than those of free ZOL. Compared to doxorubicin, the cytotoxicity of DPPG-FTL-ZOL was equivalent in drug-sensitive cell lines, and greatly superior in drugresistant cell lines. When tested on the non-FR upregulated cell lines, the cytotoxicity of FTL-ZOL was lower but still superior to that of L-ZOL. The uptake of ZOL by FR-expressing tumor cells was enhanced ∼ 25-fold with DPPG-FTL-ZOL, and only ∼ 4-fold with PEG-FTL-ZOL. Conclusions: FR targeting of ZOL using liposomes is an effective means to exploit the tumor cell growth inhibitory properties of ZOL. DPPG-FTL-ZOL is significantly more efficient at intracellular delivery of ZOL than PEG-FTL-ZOL in FR-expressing tumor cells. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Zoledronic acid (ZOL) and other nitrogen-containing bisphosphonates inhibit osteoclast activity and are extensively used in the treatment of bone metastases and osteoporosis. Anti-angiogenic and anti-tumor effects of nitrogen-containing bisphosphonates have also been demonstrated [1]. Most recently, clinical studies have demon-

Abbreviations: Chol, cholesterol; DPPG, dipalmitoyl-phosphatidylglycerol; FTL, folate-targeted liposomes; FTL-Zol, Folate-targeted liposome-encapsulated zoledronic acid; FR, Folate receptor; HSPC, Hydrogenated soybean phosphatidyl-choline; L-ZOL, liposome-encapsulated zoledronic acid; PEG, polyethylene,glycol; mPEG-2000-DSPE, methoxy-polyethylene,glycol-2000-distearoylphosphatidylethanolamine; PK, pharmacokinetics; PL, phospholipid; PHPC, partially hydrogenated phosphatidyl-choline; ZOL, zoledronic acid. ⁎ Corresponding author. Shaare Zedek MC, Oncology Institute, POB 3235, Jerusalem 91031, Israel. Fax: +972 2 6555080. E-mail address: [email protected] (A. Gabizon). 0168-3659/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2010.04.028

strated effective prevention of disease recurrence in premenopausal breast cancer when ZOL is given together with adjuvant hormonal therapy after resection of the primary tumor [2]. After i.v. administration, ZOL is rapidly cleared by the kidneys with ∼ 50% of the administered dose retained in bone mineral matrix (hydroxy-apatite) where the osteoclast-inhibitory effect takes place [3,4]. The doselimiting toxicity of ZOL is nephrotoxicity resulting in a recommended dose of 4 mg (∼50 µg/kg for an 80-kg patient) every 4 weeks in patients with bone metastases. In cell-free experiments, ZOL is a potent inhibitor of farnesyl-pyrophosphate synthase at nanomolar concentrations, interfering with the mevalonate pathway and related critical processes in cell signaling and growth [1]. Coxon et al. [5] have shown that bisphosphonates may be taken up into osteoclasts and J774 macrophages by fluid-phase endocytosis. However, bisphosphonates and specifically ZOL diffuse very poorly into other cell types and therefore have a very weak cytotoxic effect against tumor cells. Recognizing the potential value of ZOL as an anti-tumor agent, we designed a study to examine whether liposome encapsulation of ZOL

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

77

and targeting to an internalizing receptor frequently over-expressed in tumor cells, the folate receptor (FR) [6], can improve cell uptake and cytotoxicity of ZOL. Liposome formulation of ZOL could be advantageous for several reasons. Liposome encapsulation of ZOL should reduce renal clearance and, when formulated for long circulation, may increase accumulation of ZOL in the tumor [7] by a passive targeting process known as the EPR (enhanced permeability and retention) effect [8]. This creates an in situ depot of liposomal drug in the tumor interstitial fluid, exposing tumor cells, endothelial cells, and other tumor-infiltrating cells to a high and sustained drug concentration [9,10]. Based on the folate receptor-mediated endocytosis pathway [11,12], targeted liposome formulations such as folatetargeted liposomes could conceivably enable intracellular delivery of ZOL to FR-expressing tumor cells and further enhance its anti-tumor activity. This pathway has been shown to internalize folate-bound macromolecules and liposomes [13,14]. In this study, several formulations of ZOL were prepared for the purpose of comparing folate-targeted to non-targeted liposomal ZOL and to free ZOL. The stability of these formulations in the presence and absence of plasma as well as their in vivo blood clearance were tested, and their in vitro cytotoxicity and cell uptake were determined on a number of FR-expressing and non-expressing tumor cell lines.

phosphorus; however, ZOL-associated phosphorus remains in the upper (aqueous) phase and the phosphorus of the lipid components in the lower (organic) phase of the Folch extraction [15]. Using the Folch separation method, ZOL (2 phosphorus atoms per molecule) and phospholipids (1 phosphorus atom per molecule) could be quantified by the Bartlett method without interfering with each other. Liposome size and Zeta potential were determined using a NanoZ (Malvern Instruments, Malvern, UK). The final preparation was tested for residual free ZOL on a BioGel A-15 M (BioRad) or Sepharose 4B/6B column (Sigma, Israel). A liposome volume of ~ 0.25 ml was loaded onto a 10 ml-column rinsed with dialysis buffer and 20 one-ml fractions were collected. No residual free zoledronic acid was detectable after formulation preparation. This method was also used for shelf and plasma stability testing (see below). The relevant fractions containing liposomal ZOL and/or free ZOL collected after chromatographic separation were extracted by the Folch method (8:4:3 chloroform:methanol:DDW (with sample)) and subsequently quantified by the Bartlett phosphorus assay.

2. Materials and methods

The folic acid lipophilic conjugate (Folate-PEG-DSPE, MW = 4501) was added to a fraction of the liposomal preparations at 0.5% molar ratio relative to total phospholipid. The folic conjugate was weighed, suspended in the liposome-containing buffer, and incubated at 45 °C for 2 h with shaking for incorporation in the lipid bilayer, as previously reported [16]. The liposome suspensions were then cooled and centrifuged in a bench top centrifuge (10 min, 3000 rpm) to remove any conjugate unassociated with liposomes which precipitates due to its insolubility in water. Final folate concentration after loading was determined spectroscopically at 284 nm after disruption of the liposomes by dilution 1:10 in 3% sodium dodecyl sulfate (SDS). About 80–90% of conjugate was inserted in liposomes with no significant difference between the various formulations. No ZOL leaked out from liposomes during the conjugate insertion procedure.

2.1. Lipids and other chemicals Hydrogenated soybean phosphatidyl-choline (HSPC) and partially hydrogenated phosphatidyl-choline (PHPC) were obtained from Lipoid (Germany); cholesterol (Chol) was purchased from Sigma (St. Louis, MO), dipalmitoyl-phosphatidylglycerol (DPPG) and mPEG (2000)-DSPE were purchased from Bio-lab (Jerusalem, Israel) and from Avanti Lipids (Birmingham, AL). Folate-derivatized PEG (3350)DSPE was synthesized as previously described [14]. Zoledronic acid and C14-Zoledronic acid were provided by Novartis (East Hannover, NJ, and Basel, Switzerland). Tissue culture reagents were purchased from Beit HaEmek, Israel. HPLC reagents were purchased from BioLab, Jerusalem, Israel. 2.2. Formulations We prepared four formulations containing the following lipid components at a molar ratio of 55/40/5 respectively: (i) PHPC/Chol/ DPPG, (ii) PHPC/Chol/PEG-DSPE, (iii) HSPC/Chol/DPPG, (iv) HSPC/ Chol/ PEG-DSPE. Since the Chol component is constant we will refer to the formulations henceforth without it. The lipid components were weighed and dissolved in tertiary-butanol, frozen in liquid nitrogen and lyophilized overnight. The dried lipids were rehydrated in buffer containing 100 mM zoledronic acid and 15 mM histidine at pH 7.0. The lipid concentration during hydration was 100 µmoles/ml. In some preparations a spike of radioactive C14-ZOL was added to facilitate detection of ZOL. The lipid vesicles formed during hydration were downsized by serial extrusion through polycarbonate filters with pore size from 1000 nm to 50 nm. PHPC-based formulations, which are fluid at 37 °C, were hydrated and extruded at 35–40 °C, while HSPC formulations, which are solid at 37 °C, were hydrated and extruded at 60–65 °C. Following extrusion, the liposome suspension was dialyzed against a solution of 5% dextrose: 0.9% saline (9:1 volume ratio) containing 15 mM histidine, pH 7.0. The liposome suspension was then passed through a small column with Dowex anion exchange resin 1 × 2–400 beads (Sigma) to remove residual free zoledronic acid. The liposomes were sterilized by filtration through 0.22 μM filters and stored in Vacutainer™ tubes at 4 °C. Phospholipid (PL) and ZOL content were determined by Bartlett phosphorus assay of Folch extracted samples (8:4:3 chloroform: methanol:DDW) [15]. Both ZOL and liposomal phospholipids contain

2.3. Folate conjugate insertion in liposomal preparations

2.4. HPLC determination of liposomal zoledronic acid A 100 µl liposome sample was extracted by the Folch method. The Folch upper phase contains N99% of zoledronic acid as confirmed by testing liposomes spiked with C14-ZOL. Liposome-extracted samples and standard samples of ZOL were diluted further in methanol:DDW and run on a Luna C18(2), 3 µm, 150 × 4.6 mm HPLC column using a mobile phase composed of 50/50 methanol/phosphate buffer (40 mM, pH 7.3, 10 mM tetrahexylammonium, and 7.3 µg/ml EDTA), at a flow rate of 1 ml/min, at 40 °C, with UV detection at 215 nm. The retention time of ZOL was 10.3 min.

2.5. Plasma stability Fractionation by gel chromatography was utilized to differentiate between free zoledronic acid and liposome-associated zoledronic acid before and after exposure to plasma. For this purpose, C14-ZOL spiked formulations were prepared. A 100 µl sample of L-ZOL or FTL-ZOL was diluted with 400 µl plasma or buffer to 500 µl, incubated for 2 h at 37 °C, and loaded onto a BioGel A-15 M (BioRad) column (exchangeable with Sepharose 4B/6B, Sigma) prepared in a 10-ml disposable pipette. Upon elution with dialysis buffer, one-ml fractions were collected and the radioactivity, phosphorus content, and protein concentration were assessed. Fractions were tested for C-14 radioactivity in a β scintillation counter, for protein content by Bradford assay, and extracted by Folch for phosphorus determination of the upper and lower phase by the Bartlett method as described above.

78

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

2.6. Shelf stability

3. Results

Triplicates of 1.0 ml of two ZOL formulations (DPPG-PHPC and DPPG-HSPC) prepared with a spike of C14-ZOL, were incubated in sterile vacuum tubes at 4, 25 and 40 °C for 1, 3, and 6 months. Samples were tested for pH, osmolarity, and particle size. In addition, drug retention in liposomes was evaluated using Dowex anion exchange resin beads to remove any free ZOL. The radioactivity of each sample was counted before and after Dowex treatment. Samples were also assessed by HPLC, as described above, to determine the chemical stability and verify the concentration of zoledronic acid.

3.1. Characterization of formulations

2.7. Pharmacokinetics of radiolabeled ZOL formulations The various ZOL formulations labeled with C14-ZOL were injected i.v. into BALB/c female mice (Harlan, Israel), aged 10 weeks. Plasma radioactivity was measured at 1, 4, 24, and 48 h after injection. Each group and time point consisted of 3 mice receiving a dose of 30 µg/ mouse. An additional group of mice received free ZOL prepared in saline and administered i.v. at a dose of 30 µg/mouse, containing a spike of 40,000 cpm of C14-ZOL. Animals were anesthetized, bled through the eye, and sacrificed. C14 radioactivity was measured by scintillation counting.

2.8. In vitro cytotoxicity testing Cytotoxicity was assessed in a variety of cell types listed in Table 2: Normal human fibroblasts; J774 mouse macrophages; M109 mouse lung carcinoma and its multidrug resistant subline, M109R, both cell lines in their FR upregulated (HiFR) and non-upregulated forms; KB human head and neck carcinoma in HiFR form and non-upregulated; NCI/ADR doxorubicin-resistant human ovarian [17] carcinoma in HiFR form and non-upregulated; and, FR-expressing IGROV-1 human ovarian carcinoma. Cells (1500–3000 per well) were plated in 96well plates. Free and liposomal ZOL was added at concentrations ranging from 0.01 to 200 μM. Treated cells were incubated for 72 h at 37 °C. For FR-expressing tumor cells, a folate-free RPMI medium (Beyt-Haemek), while for other cells as well as for normal cells the standard RPMI-1640 medium was used. Cell survival was measured colorimetrically in a plate OD reader after staining plastic-adherent cells with methylene blue.

2.9. In vitro cell uptake of zoledronic acid The in vitro uptake of ZOL in FR-expressing cells exposed to free ZOL, L-ZOL, and FTL-ZOL was examined using formulations prepared with a spike of C14–ZOL. 1 × 106 cells from various FR-expressing cell lines were plated in 30-mm Petri dishes and incubated with 30 µM ZOL of the various formulations containing tracer C14-ZOL for 3 h at 37 °C, then washed 3× with PBS. For C14 counting, cell samples were digested by 0.5 N NaOH, followed by neutralization with HCl. Quick Safe scintillation fluid was added and samples were counted in a Beta counter. In additional experiments to further investigate the role of FR in liposomal ZOL uptake, KB-FR cells were incubated for 24 h with FTL-ZOL and L-ZOL of either DPPG-PHPC or PEG-PHPC composition at concentration of 25 µM and 50 µM of phospholipid, the cells were washed with PBS, and then folic acid cell binding was measured by incubation with H3-folic acid (Amersham, Buckinghamshire, UK) for 1 h at 37 °C in folate-free and serumfree RPMI 1640 medium. The cells were then washed 3× with PBS and treated as described above for measurement of C14-ZOL radioactivity.

Four formulations of liposomes were prepared initially for screening: PHPC-DPPG (type 1), PHPC-PEG (type 2), HSPC-DPPG (type 3), and HSPC-PEG (type 4) (Table 1). All formulations contained Chol. Phospholipid recovery was ∼50%. ZOL recovery in encapsulated form was generally between 4 and 5%, with final concentrations of ∼1 mg/ml (∼3.5 µmoles/ml). The final drug/PL molar ratio ranged between 15 and 20%. Lipid composition had no apparent impact on ZOL encapsulation efficiency. Mean vesicle size was in the range of 100–110 nm for all formulations. DPPG formulations showed slightly negative values in zeta potential measurements while pegylated formulations were neutral (Table 1). Shelf stability analysis after 6 months at 40 °C demonstrated highly stable drug retention (∼100%) for all formulations. The ZOL concentration remained unchanged. Other parameters (pH, osmolarity and vesicle size) also remained stable during at least 6 months. 3.2. Stability in plasma A plasma stability study of all 4 formulations labeled with C14-ZOL was done to assess drug release in the presence of 80% human plasma for 2 h at 37 °C (Fig. 1). Elution patterns of liposomes (as liposomal phospholipid — Pi-), plasma proteins, and free ZOL (based on C14-ZOL cpm) are shown in Fig. 1A as controls. All formulations appear to be highly stable (Fig. 1B–E) as evidenced by the fact that nearly all the C14-ZOL cpm eluted with the liposomal phospholipid peak (fractions 5–6), away for the protein peak (fractions 9–10) and from the free ZOL peak (fractions 10–11). A small amount of leakage (∼10–15% of the drug load) was detected only in the DPPG-PHPC formulation (Fig. 1E). Two formulations of FTL-ZOL (DPPG-PHPC and PEG-PHPC) were also tested for drug stability by exposure to plasma for 2 h at 37 °C. Both were highly stable with no detectable drug leakage. There was no release of free ZOL after the folate ligand insertion procedure and no leakage of ZOL after incubation in plasma (Fig. 1F). 3.3. Pharmacokinetics of L-ZOL formulations Free ZOL was cleared very quickly: at 1 h after injection, only 0.1– 0.2% of the injected dose is present in plasma (data not shown), in contrast to nearly 100% of the injected dose for liposomal ZOL (Fig. 2A). PEG preparations are cleared more slowly than DPPG preparations as expected. In any case, the 24-hour L-ZOL plasma levels were N10% of the injected dose for all four formulations indicating that they all displayed relatively long circulation times. Given the extremely fast clearance of free ZOL, this also suggests that leakage of

Table 1 Formulations of liposome-encapsulated ZOLa. Component

Type 1

Type 2

Type 3

Type 4

PCb Cholesterol Other lipid DSPE-PEG-folate PLc (μmoles/ml) ZOL (mg/ml) ZOL/PL molar ratio Size (nm) Zeta potential (mV) Osmolarity (mOsm/l) pH 37 °C state

PHPC + DPPG +/− 23.75 1.01 0.15 100 ± 13 −35.5 ± 0.5 271 6.6 Fluid

PHPC + DSPE-PEG +/− 22.35 1.09 0.17 106 ± 20 0 265 6.55 Fluid

HSPC + DPPG +/− 23.40 1.34 0.21 102 ± 21 −36.4 ± 0.2 263 6.55 Solid

HSPC + DSPE-PEG +/− 25.85 1.08 0.15 110 ± 10 0 260 6.6 Solid

a b c

Results shown are the mean of two representative batches. PC = phophatidyl-choline. PL = phospholipid including PC and DPPG or DSPE-PEG components.

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

79

Fig. 1. Plasma stability of L-ZOL formulations (radiolabeled with C14-ZOL). Chromatographic elution after 2 h 37 °C incubation in 80% human plasma or buffer, as described in Materials and methods. A. Control elution profile of free ZOL spiked with C14-ZOL, liposomal phospholipid based on phosphorus (Pi) assay, and plasma proteins; B. Elution profile of DPPG-PHPC L-ZOL; C. Elution profile of DPPG-HSPC L-ZOL; D. Elution profile of PEG-PHPC L-ZOL; E. Elution profile of PEG-HSPC L-ZOL; and F. Elution profile of DPPG-PHPC FTL-ZOL.

ZOL from circulating liposomes is negligible. DPPG-HSPC L-ZOL was cleared significantly faster than DPPG-PHPC L-ZOL. This observation was reproduced in a subsequent experiment (Fig. 2B) in which we also tested the corresponding FTL-ZOL formulations of DPPG-HSPC and DPPG-PHPC. FTL-ZOL was found to be cleared faster than L-ZOL, a finding that is consistent with our previous experience with other FTL formulations [18]. 3.4. In vitro cytotoxicity in high FR-expressing and low FR-expressing cell lines A number of cell lines with intrinsic folate receptor (FR) overexpression or upregulated FR by selective adaptation in folate-deleted medium were studied. Control cell lines of low FR-expressing cells were tested, including a macrophage cell line (J774), and normal human skin fibroblasts. Free ZOL was cytotoxic at concentrations usually greater than 30 µM with a maximal range of 10–200 µM. Although these concentrations may be achievable in bone, they are not pharmaco-

logically attainable in other body tissues, including tumor tissue [4], thus the need for an effective drug delivery vector. Non-targeted formulations were consistently not cytotoxic indicating no drug leakage, and probably no liposome uptake by cells in vitro (Table 2, Fig. 3). Targeted and non-targeted formulations displayed minimal cytotoxicity against low folate receptor-expressing normal fibroblasts and J774 macrophages (Table 2). In addition, KB cells expressing low amounts of FR were much less sensitive to FTLZOL than those expressing high levels indicating that the FR plays a critical role in the cytotoxic effect of these targeted formulations (Table 2). Drug-free liposomes, whether targeted or non-targeted, were non-cytotoxic (data not shown). Among all formulations tested, DPPG-PHPC FTL-ZOL was the most active displaying high cytotoxicity against KB-HiFR, NCI/ADR-HiFR, and M109-HiFR cells with major growth inhibition at sub-micromolar concentrations (Table 2, Fig. 3). DPPG formulations were more effective than the free zoledronic acid by N100 fold. Consistently, the folate-targeted DPPG formulations were found to be more cytotoxic than the folate-targeted PEG formulations (Fig. 3), an

80

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

Fig. 2. A. Plasma clearance of L-ZOL (DPPG-PHPC, DPPG-HSPC, PEG-PHPC, and PEGHSPC formulations) in BALB/c mice after i.v. injection of 30 µg L-ZOL (radiolabeled with C14-ZOL). An equal dose of i.v. radiolabeled free ZOL was rapidly cleared (0.2% of injected dose in plasma at 1 h after injection, not shown). B. Plasma levels of L-ZOL and FTL-ZOL (DPPG-PHPC and DPPG-HSPC formulations) at 24 h after i.v. injection of 30 µg (radiolabeled with C14-ZOL) in BALB/c mice.

observation that appears to be related to the differential cell uptake of these formulations (see below). The growth inhibition curve appears to reach a plateau at FTL-ZOL concentration of ∼ 5 µM without

Fig. 3. Comparison of the in vitro cytotoxicity of Free ZOL, FTL-ZOL and non-targeted LZOL against human KB-HiFR carcinoma cells. A. DPPG-PHPC formulation. B. PEG-PHPC formulation. Results based on the standard 72-hour cytotoxicity assay as described in Materials and methods.

reaching 100% inhibition. This observation was consistently reproduced and may be related to interference with liposome internalization caused by a high concentration of liposomes bound to the cell membrane. Of note, the in vitro cytotoxic activity of FTL-ZOL was comparable or even greater than that of doxorubicin (Table 2). In human ovarian

Table 2 IC50 values of Free ZOL, L-ZOL, and FTL-ZOL in various cell lines (μM).a Cell line (FR Lo/Hi) Fibroblastb (Lo) J774c (Lo) M109d (Hi) M109Re (Hi) KBf (Hi) KBf (Lo) NCI/ADRg (Hi) NCI/ADRg (Lo) IGROV-1h (Hi) a b c d e f g h i

Free ZOL

DPPG/PHPC L-ZOL

DPPG/PHPC FTL-ZOL

DPPG/HSPC L-ZOL

DPPG/HSPC FTL–ZOL

PEG/PHPC L-ZOL

PEG/PHPC FTL-ZOL

PEG/HSPC L-ZOL

PEG/HSPC FTL-ZOL

200

147

N 200

150

N200

160

N200

N 200

3.0 174 200 117 61 98

83 N 200 N 200 N 200 N 200 N 200

45 0.6 2.2 0.4 N200 7.7

N 200 NDi ND N 200 97 N 200

153 ND ND 1.5 88 45

90 N200 N200 N200 N200 N200

153 100 N 200 7.4 81 196

137 ND ND N200 N200 N200

153 ND ND 29 N 200 N 200

136

155

61

200

68

ND

ND

ND

ND

35

50

0.15

ND

ND

ND

ND

ND

ND

5.5

Mean of ≥2 experiments. Free Doxorubicin IC50 for selected cell lines: M109 = 0.7 μM; M109R = 10 μM; KB = 0.1 μM; NCI/ADR = 156 μM; and IGROV-1 = 0.4 μM. Human skin fibroblasts. J774 mouse macrophage cell line. M109 mouse carcinoma. M109R multidrug resistant (MDR1+) mouse carcinoma. KB human carcinoma with high FR expression or low (culture-downregulated) FR expression. NCI/ADR multidrug resistant (MDR1+) human carcinoma with high (culture-upregulated) or low FR expression. iGROV-1 human carcinoma. ND = not done.

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

81

receptors irrespective of their sensitivity to common chemotherapeutic agents. 3.5. In vitro ZOL uptake by FR-expressing tumor cells

Fig. 4. In vitro cytotoxicity of Free ZOL, free doxorubicin (Dox), FTL-ZOL and nontargeted L-ZOL (DPPG-PHPC formulation) against human IGROV-1 ovarian carcinoma cells. Results based on the standard 72-hour cytotoxicity assay as described in Materials and methods.

IGROV-1 (naturally overexpressing FR), the IC50 of FTL-ZOL was approximately 0.1 μM, which was significantly lower than that of nontargeted L-ZOL, free ZOL, and free doxorubicin (∼0.4 μM) (Fig. 4). In the doxorubicin-resistant HiFR cell lines (NCI/ADR and M109R), the cytotoxicity of the FTL-ZOL (DPPG-PHPC) formulation was greatly superior to that of the free doxorubicin (Table 2) indicating that FTLZOL is not affected by the MDR phenotype. Altogether, these data demonstrate that folate-targeted liposomal ZOL is highly cytotoxic against tumors that overexpress folate

Fig. 5. Uptake of Free ZOL, FTL-ZOL, and L-ZOL by FR-expressing tumor cells. C14-ZOL radiolabeled formulations (DPPG-PHPC, DPPG-HSPC, PEG-PHPC, and PEG-HSPC) and free ZOL tested at a concentration of 30 µM ZOL for 3 h at 37 °C. A. KB (HiFR) cells and B. NCI/ADR cells.

Data on cell uptake of drug-free, radiolabeled FTL with and without PEG coating point to a substantial increase in liposome uptake when compared to radiolabeled non-targeted liposomes [19]. Here, we examined drug uptake of FTL-ZOL and L-ZOL using radioactive zoledronic acid tracer (C14-ZOL) in three cell lines: KB-HiFR (Fig. 5A), NCI/ADR-HiFR (Fig. 5B), and M109R (data not shown) cells. Results were similar for all cell models. Drug uptake with FTLZOL averaged 50-fold more than with non-targeted L-ZOL and ∼25fold more than with free ZOL. Consistent with cytotoxic data, the DPPG-containing formulations displayed greater uptake (∼ 5-fold more) than the PEG formulations. However there was no significant difference in uptake between the PHPC (fluid phase) and HSPC (solid phase)-based formulations (Fig. 5). To determine whether the availability of FR was affected by prior cell exposure to the FTL, we examined KB-HiFR cell binding of radiolabeled folic acid after exposure to FTL-ZOL of either DPPG-PHPC or PEG-PHPC composition. A 3-hour exposure to FTL-ZOL, and

Fig. 6. Effect of FTL-ZOL pretreatment on uptake of radiolabeled folic acid by FR-expressing KB cells. The formulations studied were: DPPG-PHPC and PEG-PHPC. A. Inhibition of H3folic acid uptake after 3 or 24-hour exposure to FTL-ZOL (25–50 µM) or L-ZOL (25-50 µM). The 100% uptake value was set as H3-folic acid uptake in non-liposome exposed cells. Note that the inhibitory effect is greater with longer time of exposure and with higher phospholipid dose (25 vs. 50 µM). Non-targeted L-ZOL did not have any inhibitory effect on folate uptake. B. Recovery of H3-folic acid uptake after prior exposure to FTL-ZOL (25 µM), L-ZOL (25 µM), or cold folic acid (1 µg/ml) followed by 6-hour incubation in liposome-free medium. To correct for cell recovery which varied among the different treatment groups, values were adjusted to a fixed number of recovered cells. SD were b 20% of the mean.

82

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

subsequent H3-folic acid binding for 1 h indicated that DPPG-PHPC reduced the FR available by 4.4 fold whereas PEG-PHPC reduced the FR available by 2.9 fold. After a 24 hour exposure, both FTL-ZOL formulations decreased H3-folic acid binding by more than 95% (N20fold reduction of available FR). Non-targeted liposomes had no effect on folic acid binding (Fig. 6A). In further experiments to follow the recovery of the FR after exposure to FTL-ZOL, H3-folic acid cell uptake was determined immediately and after 6 h incubation in fresh (liposome-free) medium. Despite a 6-hour recovery period, the uptake of H3-folic acid was still significantly lower in the FTL-exposed cells than in the non-exposed or L-ZOL exposed cells. Lower uptake of H3-folic acid was noted for cells exposed to DPPG-PHPC than for those exposed to PEG-PHPC FTL-ZOL. There was evidence of partial recovery for both formulations, similar to that observed after the cold folate treatment, indicating that FTL-ZOL pretreatment does not interfere with the recycling of the receptor (Fig. 6B). Another factor that may affect liposome-cell uptake could be differences in the bilayer distribution and/or level of insertion of postloaded conjugate in the various formulations. Analysis of the formulations after micellar insertion of the ligand did not reveal any significant differences in the total amount of folate conjugate in the DPPG and PEG-based liposome preparations. To confirm that the differences in activity between the DPPG and PEG formulations are not due to a difference in bilayer distribution of the post-loaded ligand, we prepared two FTL formulations (DPPG-PHPC and PEGPHPC liposomes) incorporating the folate conjugate together with the initial lipid components of the formulation prior to hydration. When we tested the uptake of these pre-loaded, C14-ZOL labeled FTL formulations into KB-HiFR cells, we observed 6- to 7-fold greater ZOL cell levels with DPPG-PHPC as compared to PEG-PHPC (data not shown), confirming the results obtained with formulations prepared with the post-inserted ligand. These data indicate that there is a correlation between folate targeting, liposomal drug uptake and cytotoxicity and that DPPGcontaining FTL are more efficient than the PEG-based FTL for targeting ZOL. The latter show reduced interaction with the folate receptor despite an optimal level of ligand present in the liposome. 4. Discussion Recent clinical studies have demonstrated potential benefits to breast cancer patients using ZOL in metastases prevention [20]. In addition, ZOL treatment reduced the level of serum VEGF in cancer patients suggesting an anti-angiogenic effect [21]. Preclinical studies also reveal non-skeletal anti-tumor effects of ZOL in various tumor models (breast cancer, myeloma, and others) which suggest direct tumor-inhibitory activity of ZOL [22–24]. The anti-tumor effect appears to be mediated by a molecular mechanism similar to that observed in osteoclasts [25]. However the full impact of this drug's efficacy is severely limited due to its low cellular permeability. An intracellular drug delivery mechanism should significantly facilitate the cytotoxic potential of ZOL. This was the rationale behind this study by using folate-targeted liposomes to deliver encapsulated ZOL. As demonstrated here, folate-targeted liposomes can greatly enhance the delivery and subsequent cytotoxicity of zoledronic acid in a variety of tumor cell types. The use of liposomal bisphosphonates, specifically liposomal clodronate, a non-nitrogen-containing bisphosphonate, has been known for many years as a highly specific tool to suppress and even deplete macrophages, as reviewed by Van Rooijen [26]. In more recent years, depletion of blood monocytes has been demonstrated with liposomal alendronate, a nitrogen-containing bisphosphonate [27]. However, no studies have been reported yet with liposomeencapsulated zoledronic acid. All four liposome preparations of ZOL studied here were extraordinarily robust in terms of physicochemical stability with

projected shelf times N1 year. Although the efficiency of ZOL encapsulation was low as expected from a passively encapsulated water-soluble compound, leakage in storage was nil and stability in plasma, a good predictor of in vivo stability in circulation [28], was excellent. PK studies confirm the prediction of plasma stability studies and add important information on the relative performance of the various formulations. The presence of ZOL did not seem to cause any important change from the PK patterns of these liposome formulations. Particularly, the better performance of PEG-containing liposomes over the DPPG ones is clearly apparent, and was an expected finding. An unexpected observation with no clear explanation is the shorter circulation time of DPPG-HSPC as compared to DPPG-PHPC. As previously reported for drug-free liposomes [9], the insertion of the folate ligand resulted in faster clearance of ZOL liposomes from plasma. Clearly, the PK profile of these various formulations is an important factor for in vivo application targeted to cancer. Formulation composition affected the uptake and cytotoxicity of folate-targeted L- ZOL. The optimal formulation of those studied here in terms of ZOL uptake and cytotoxicity was folate-targeted DPPGPHPC. Based on this formulation, the gain in cytotoxicity is ∼ 200-fold when compared to free ZOL and N400-fold when compared to the non-targeted formulation. This enormous gain was consistent across several FR-expressing cell lines and makes the cytotoxic potency of ZOL equivalent to that of other powerful DNA damaging cytotoxic agents. This stresses the potential value of ZOL as an anti-tumor agent once the drug efficiently penetrates tumor cells. In addition, the fact that FTL-ZOL was as active against MDR-positive cells indicates that FTL-ZOL could be a valuable complement to therapy acting through a non-cross-resistant mechanism and killing as well cells otherwise resistant to common cytotoxic drugs. ZOL can be considered a liposome-dependent drug in line with drugs such as methotrexateγ- aspartate, 5-fluoro-orotate, and PALA (N-(phosphonacetyl)-Laspartic acid) for which liposome delivery enables in vitro cytotoxicity while free drug is hardly cytotoxic [29]. This is in contrast to doxorubicin which freely diffuses into cells and has comparable in vitro cytotoxicity to folate-targeted liposomal doxorubicin [30,31]. PEG formulations delivered less ZOL than DPPG formulations to FR-expressing cells (Fig. 5). Past studies have shown that PEG coating of liposomes interferes with binding of the folate ligand to the receptor [19]. Although this problem can be reduced by conjugating the folate ligand to a longer PEG than the PEG used for liposome coating, as done here, pegylated liposomes are still less effective in cell uptake experiments than PEG-free liposomes such as DPPG-based formulations. This was not the result of differential ligand uptake into the post-loaded liposomes of different compositions, and the difference in uptake is not accounted for by differences in turnover of the folate receptor between the DPPG and PEG-based formulations. PEG hindrance of liposome interaction with FR appears to be the prevailing mechanism. Our liposome-cell binding measurements indicate the total amount of cell-associated liposomal ZOL, but we do not know the fraction of liposomal drug that is internalized and the impact of liposome composition on this process. Regarding folate conjugates, Paulos et al. [32] have reported that only up to 25% of the receptorbound material is internalized in a number of tumor cell lines. Clearly, an important fraction of the liposomal ZOL bound to the surface FR is internalized. Otherwise, no cytotoxic activity would have been observed. Following receptor-mediated liposome endocytosis, the rate limiting factor for the cytotoxic effect is the release rate of ZOL from liposomes. The greater cytotoxic activity of PHPC-DPPG over HSPC-DPPG is not surprising since fluid-phase liposomes (PHPC is fluid at 37 °C) will be broken down in cell lysosomes and leak their contents faster than solid-phase liposomes (HSPC is solid at 37 °C) in cells and biological fluids. When considering in vivo drug delivery, it has yet to be determined whether the better uptake and cytotoxicity of the DPPG-PHPC

H. Shmeeda et al. / Journal of Controlled Release 146 (2010) 76–83

formulation will be offset by the longer circulation time (Fig. 2) and predictably higher tumor accumulation of the classical Stealth, PEGbased, formulations [33]. Another factor to take into account is the possibility of a change in liposome clearance and biodistribution by bisphosphonate-induced macrophage blockade as shown in mice pretreated with liposomal clodronate [34]. In conclusion, intracellular delivery of ZOL in liposome-encapsulated form via folate receptor-mediated endocytosis results in major potentiation of in vitro tumor cell cytotoxicity. Acknowledgments We thank the Zometa Global Project Team for their helpful comments on the manuscript. Support by Novartis Pharmaceuticals Corporation, and by the Israel Cancer Research Fund is gratefully acknowledged. References [1] J.R. Green, Bisphosphonates: preclinical review, Oncologist 9 (Suppl 4) (2004) 3–13. [2] M. Gnant, B. Mlineritsch, W. Schippinger, G. Luschin-Ebengreuth, S. Postlberger, C. Menzel, R. Jakesz, M. Seifert, M. Hubalek, V. Bjelic-Radisic, H. Samonigg, C. Tausch, H. Eidtmann, G. Steger, W. Kwasny, P. Dubsky, M. Fridrik, F. Fitzal, M. Stierer, E. Rucklinger, R. Greil, C. Marth, Endocrine therapy plus zoledronic acid in premenopausal breast cancer, N. Engl. J. Med. 360 (7) (2009) 679–691. [3] A. Skerjanec, J. Berenson, C. Hsu, P. Major, W.H. Miller Jr., C. Ravera, H. Schran, J. Seaman, F. Waldmeier, The pharmacokinetics and pharmacodynamics of zoledronic acid in cancer patients with varying degrees of renal function, J. Clin. Pharmacol. 43 (2) (2003) 154–162. [4] H.M. Weiss, U. Pfaar, A. Schweitzer, H. Wiegand, A. Skerjanec, H. Schran, Biodistribution and plasma protein binding of zoledronic acid, Drug Metab. Dispos. 36 (10) (2008) 2043–2049. [5] F.P. Coxon, K. Thompson, A.J. Roelofs, F.H. Ebetino, M.J. Rogers, Visualizing mineral binding and uptake of bisphosphonate by osteoclasts and non-resorbing cells, Bone 42 (5) (2008) 848–860. [6] N. Parker, M.J. Turk, E. Westrick, J.D. Lewis, P.S. Low, C.P. Leamon, Folate receptor expression in carcinomas and normal tissues determined by a quantitative radioligand binding assay, Anal. Biochem. 338 (2) (2005) 284–293. [7] A. Gabizon, D. Papahadjopoulos, Liposome formulations with prolonged circulation time in blood and enhanced uptake by tumors, Proc. Natl. Acad. Sci. U. S. A. 85 (18) (1988) 6949–6953. [8] H. Maeda, The enhanced permeability and retention (EPR) effect in tumor vasculature: the key role of tumor-selective macromolecular drug targeting, Adv. Enzyme Regul. 41 (2001) 189–207. [9] A. Gabizon, A.T. Horowitz, D. Goren, D. Tzemach, H. Shmeeda, S. Zalipsky, In vivo fate of folate-targeted polyethylene-glycol liposomes in tumor-bearing mice, Clin. Cancer Res. 9 (17) (2003) 6551–6559. [10] D.S. Alberts, F.M. Muggia, J. Carmichael, E.P. Winer, M. Jahanzeb, A.P. Venook, K.M. Skubitz, E. Rivera, J.A. Sparano, N.J. DiBella, S.J. Stewart, J.J. Kavanagh, A.A. Gabizon, Efficacy and safety of liposomal anthracyclines in phase I/II clinical trials, Semin. Oncol. 31 (6 Suppl 13) (2004) 53–90. [11] S. Wang, P.S. Low, Folate-mediated targeting of antineoplastic drugs, imaging agents, and nucleic acids to cancer cells, J. Control. Release 53 (1–3) (1998) 39–48. [12] J. Sudimack, R.J. Lee, Targeted drug delivery via the folate receptor, Adv. Drug Deliv. Rev. 41 (2) (2000) 147–162. [13] C.P. Leamon, P.S. Low, Delivery of macromolecules into living cells: a method that exploits folate receptor endocytosis, Proc. Natl. Acad. Sci. U. S. A. 88 (13) (1991) 5572–5576. [14] R.J. Lee, P.S. Low, Delivery of liposomes into cultured KB cells via folate receptormediated endocytosis, J. Biol. Chem. 269 (5) (1994) 3198–3204.

83

[15] H. Shmeeda, S. Even-Chen, R. Honen, R. Cohen, C. Weintraub, Y. Barenholz, Enzymatic assays for quality control and pharmacokinetics of liposome formulations: comparison with nonenzymatic conventional methodologies, Methods Enzymol. 367 (2003) 272–292. [16] H. Shmeeda, L. Mak, D. Tzemach, P. Astrahan, M. Tarshish, A. Gabizon, Intracellular uptake and intracavitary targeting of folate-conjugated liposomes in a mouse lymphoma model with up-regulated folate receptors, Mol. Cancer Ther. 5 (4) (2006) 818–824. [17] M. Liscovitch, D. Ravid, A case study in misidentification of cancer cell lines: MCF7/AdrR cells (re-designated NCI/ADR-RES) are derived from OVCAR-8 human ovarian carcinoma cells, Cancer Lett. 245 (1–2) (2007) 350–352. [18] A. Gabizon, H. Shmeeda, A.T. Horowitz, S. Zalipsky, Tumor cell targeting of liposome-entrapped drugs with phospholipid-anchored folic acid-PEG conjugates, Adv. Drug Deliv. Rev. 56 (8) (2004) 1177–1192. [19] A. Gabizon, A.T. Horowitz, D. Goren, D. Tzemach, F. Mandelbaum-Shavit, M.M. Qazen, S. Zalipsky, Targeting folate receptor with folate linked to extremities of poly(ethylene glycol)-grafted liposomes: in vitro studies, Bioconjug. Chem. 10 (2) (1999) 289–298. [20] R. Coleman, M. Gnant, New results from the use of bisphosphonates in cancer patients, Curr. Opin. Support. Palliat. Care 3 (3) (2009) 213–218. [21] D. Santini, B. Vincenzi, S. Galluzzo, F. Battistoni, L. Rocci, O. Venditti, G. Schiavon, S. Angeletti, F. Uzzalli, M. Caraglia, G. Dicuonzo, G. Tonini, Repeated intermittent low-dose therapy with zoledronic acid induces an early, sustained, and longlasting decrease of peripheral vascular endothelial growth factor levels in cancer patients, Clin. Cancer Res. 13 (15 Pt 1) (2007) 4482–4486. [22] A. Guenther, S. Gordon, M. Tiemann, R. Burger, F. Bakker, J.R. Green, W. Baum, A.J. Roelofs, M.J. Rogers, M. Gramatzki, The bisphosphonate zoledronic acid has antimyeloma activity in vivo by inhibition of protein prenylation, Int. J. Cancer 126 (1) (2010) 239–246. [23] P.D. Ottewell, D.V. Lefley, S.S. Cross, C.A. Evans, R.E. Coleman, I. Holen, Sustained inhibition of tumour growth and prolonged survival following sequential administration of doxorubicin and zoledronic acid in a breast cancer model, Int. J. Cancer 126 (2) (2010) 522–532. [24] A.C. Hirbe, A.J. Roelofs, D.H. Floyd, H. Deng, S.N. Becker, L.G. Lanigan, A.J. Apicelli, Z. Xu, J.L. Prior, M.C. Eagleton, D. Piwnica-Worms, M.J. Rogers, K. Weilbaecher, The bisphosphonate zoledronic acid decreases tumor growth in bone in mice with defective osteoclasts, Bone 44 (5) (2009) 908–916. [25] L.M. Mitrofan, J. Pelkonen, J. Monkkonen, The level of ATP analog and isopentenyl pyrophosphate correlates with zoledronic acid-induced apoptosis in cancer cells in vitro, Bone 45 (6) (2009) 1153–1160. [26] N. van Rooijen, Liposomes for targeting of antigens and drugs: immunoadjuvant activity and liposome-mediated depletion of macrophages, J. Drug Target. 16 (7) (2008) 529–534. [27] H. Epstein, D. Gutman, E. Cohen-Sela, E. Haber, O. Elmalak, N. Koroukhov, H.D. Danenberg, G. Golomb, Preparation of alendronate liposomes for enhanced stability and bioactivity: in vitro and in vivo characterization, AAPS J. 10 (4) (2008) 505–515. [28] J. Senior, G. Gregoriadis, Stability of small unilamellar liposomes in serum and clearance from the circulation: the effect of the phospholipid and cholesterol components, Life Sci. 30 (24) (1982) 2123–2136. [29] J.S. Kim, T.D. Heath, In vitro cytotoxic effect of N-(phosphonacetyl)-L-aspartic acid in liposome against C-26 murine colon carcinoma, Arch. Pharm. Res. 23 (2) (2000) 167–171. [30] D. Goren, A.T. Horowitz, D. Tzemach, M. Tarshish, S. Zalipsky, A. Gabizon, Nuclear delivery of doxorubicin via folate-targeted liposomes with bypass of multidrugresistance efflux pump, Clin. Cancer Res. 6 (5) (2000) 1949–1957. [31] R.J. Lee, P.S. Low, Folate-mediated tumor cell targeting of liposome-entrapped doxorubicin in vitro, Biochim. Biophys. Acta 1233 (2) (1995) 134–144. [32] C.M. Paulos, J.A. Reddy, C.P. Leamon, M.J. Turk, P.S. Low, Ligand binding and kinetics of folate receptor recycling in vivo: impact on receptor-mediated drug delivery, Mol. Pharmacol. 66 (6) (2004) 1406–1414. [33] M.C. Woodle, D.D. Lasic, Sterically stabilized liposomes, Biochim. Biophys. Acta 1113 (2) (1992) 171–199. [34] E. Claassen, N. Van Rooijen, The effect of elimination of macrophages on the tissue distribution of liposomes containing [3H]methotrexate, Biochim. Biophys. Acta 802 (3) (1984) 428–434.