648
Bioconjugate Chem. 2009, 20, 648–655
ARTICLES Folate Receptor Targeted Bimodal Liposomes for Tumor Magnetic Resonance Imaging Nazila Kamaly,†,‡ Tammy Kalber,‡ Maya Thanou,† Jimmy D. Bell,‡ and Andrew D. Miller*,†,§ Imperial College Genetic Therapies Centre, Department of Chemistry, Flowers Building, Armstrong Road, Imperial College London, London, SW7 2AZ, Metabolic and Molecular Imaging Group, Imaging Sciences Department, MRC Clinical Sciences Centre, Imperial College London, Hammersmith Hospital, London, W12 0HS, and ImuThes Ltd, Flowers Building, Armstrong Road, London SW7 2AZ, United Kingdom. Received June 1, 2008; Revised Manuscript Received February 8, 2009
Folate-targeted bimodal paramagnetic and fluorescent liposomes were developed and showed enhanced accumulation in a folate receptor expressing tumor model. These bimodal liposomes were composed of both a paramagnetic and a fluorescent lipid, and utilized a PEG-lipid amphiphile for prolonged in vivo circulation. The particles were formulated to ensure a size distribution of approximately 100 nm with a low polydispersity index. IGROV-1 cells were used to induce tumors in nude Balb/c mice, and the folate-targeted liposomes were injected intravenously. Rapid accumulation of the folate-targeted liposomes within the tumor tissue compared to nontargeted liposomes was observed. Furthermore, folate-labeled liposomes showed a 4-fold increase in tumor T1 signal intensity at just 2 h postinjection with similar results being obtained for the nontargeted liposomes only 24 h postinjection. In addition, the folate-targeted liposomes were injected at half the nontargeted liposome dose, further demonstrating their effectiveness. Histological analysis of sectioned tumor slices revealed distinct fluorescence patterns between the targeted and nontargeted systems, with a more localized and hyperintense fluorescence signal observed from tumor sections post-folate-targeted liposome injections. These results demonstrate the effectiveness of folate targeting for dynamic real-time solid tumor MRI and provide insight into kinetics of targeted and nontargeted nanoparticles to solid tumors.
INTRODUCTION Folate based targeting systems present an effective means of selectively delivering therapeutic or imaging agents to tumors (1-11). It has been shown that aggressive or undifferentiated tumors at an advanced stage have an increased folate receptor (FR) density, indicating that cancer therapy could benefit from the broad approach that FR mediated drug delivery offers (12, 13). The FR is overexpressed in several cancer types, such as brain, kidney, lung, and breast cancers, and, in particular, in epithelial carcinomas such as ovarian cancers (14). The FR ligand, folate (or folic acid), is a vitamin that is used for the biosynthesis of nucleotides and is utilized in high levels to meet the needs of proliferating cancer cells (15). In addition to numerous drug delivery efforts, folate-targeted technology has been successfully applied to radio-imaging of therapeutic agents (16), fluorescence imaging of cancer cells (17), MRI contrast agents (18), and gadolinium liposomes (19). Choi et al. have demonstrated the use of folate-targeted iron oxide nanoparticles for the imaging of induced KB tumors and showed these particles to have a 38% signal intensity increase compared to controls (20). Successful tumor MRI with a nontargeted bimodal liposomal contrast agent was shown recently, whereby bimodal paramagnetic and fluorescent liposomes of ∼100 nm in size were seen to accumulate in a mouse * Author e-mail address:
[email protected]. † Imperial College Genetic Therapies Centre. ‡ MRC Clinical Sciences Centre. § ImuThes Ltd.
xenograft model of ovarian cancer (21). Liposomes are able to accumulate within tumor tissue due to the widely reported enhanced permeation and retention effect (EPR) which relies on the passive accumulation of colloidal macromolecules of ∼40 kDa and above in tumors (22). The EPR effect arises due to aberrant tumor endothelium, which as a result of its “leakiness” allows the penetration of nanoparticles into tumor tissue. Liposome accumulation in tumor tissue could be improved through the use of receptor targeting moieties that are either postconjugated to the surface of liposomes or attached to lipids that become incorporated within the liposomal bilayer. Since FR binding affinity (Kd ) 1 × 1-10 M) does not appear to be affected when its ligand, folate, is conjugated to an imaging agent or therapeutic moiety via its γ-carboxyl (10, 23), a folate ligand tethered onto the distal end of a lipidic PEG amphiphile allows for the development of a FR targeted liposomal system. The human nasopharyngeal KB carcinoma cell line is considered to have the highest level of FR expression, yet the number of cases for this cancer are low in comparison to ovarian cancer, which has the highest frequency (>90% of cases) (6). In particular, the R-FR isoform which is a glycosyl phosphatidylinositol (GPI)-anchored membrane protein is highly expressed in ovarian carcinoma (24). Additionally, the R-FR isoform has also been shown to have a specific biomarker value, aiding in the identification of metastatic tumor site origin (25). Therefore, we were interested in using this receptor in order to test the efficacy of folate-targeted bimodal liposomes for the imaging of ovarian tumors using MRI. Folate-based liposomal drug delivery has been studied extensively (7, 24, 26); however,
10.1021/bc8002259 CCC: $40.75 2009 American Chemical Society Published on Web 03/10/2009
Bimodal Folate Receptor Targeted Liposomes for Tumor Imaging
the rate-enhancing effect of liposome accumulation in tumors due to folate targeting has not been studied dynamically in real time to a great extent. Effective tumor signal enhancement was anticipated, since the FR is expressed in significantly lower amounts in normal tissue, limited mainly to kidney tubuli, lung epithelium, and placenta tissue (27). To asses the value of the addition of a targeting ligand on the rate and extent of accumulation of liposomes in solid tumors, FR targeted bimodal fluorescent and paramagnetic liposomes were formulated and compared to nontargeted liposomes by both MRI and fluorescence microscopy.
EXPERIMENTAL RESULTS Materials. Phosphatidylethaolamine-lissamine rhodamine B (DOPE-Rhodamine), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-Nmethoxy(poly(ethylene glycol))-2000 (DSPE-PEG2000), and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[folate(poly(ethylene glycol))-2000] (DSPE-PEG(2000)folate, ammonium salt) were purchased from Avanti Polar Lipids Inc. (Alabaster, AL, USA). All other chemicals were of analytical grade or the best grade available and purchased from Sigma-Aldrich (UK). Gadolinium (III) 2-{4,7-bis-carboxymethyl-10-[(N,N-distearylamidomethyl-N′-amido-methyl]-1,4,7,10-tetra azacyclododec1-yl}-acetic acid (Gd.DOTA.DSA) was synthesized in-house as shown previously (21). General Procedures. Liposome zeta potentials were recorded on a Nanoseries Nano-ZS zetasizer (Malvern instruments). Microscopy experiments were conducted on a Nikon Eclipse E600 microscope. FACS analysis was conducted on a Becton Dickinson FACSCalibur machine. ICP-AES experiments were run on a Varian VISTA PRO instrument. All MRI experiments were conducted on a 4.7 T Magnex magnet (Oxford, UK) with a Varian Unity Inova console (Palo Alto, CA, USA). All procedures on animals were carried out in accordance with UK Home Office regulations and the Guidance for the Operation of Animals (Scientific Procedures) Act (1986). Liposome Preparations. All lipids were stored as stock solutions in anhydrous organic solvents (CHCl3, MeOH, or a mixture of both), at -20 °C under argon. Liposomes were made with defined molar ratios of individual lipids to give a predetermined total lipid concentration of 15 (folate-targeted) or 30 (nontargeted) mg mL-1 in HEPES (20 mM, NaCl 135 mM, pH 6.5) for in vivo experiments. Targeted liposomes consisted of Gd.DOTA.DSA/DOPC/Cholesterol/DSPE-PEG2000/ DSPE-PEG(2000)Folate/DOPE-Rhodamine 30/32/30/4/3/1 mol %, and nontargeted liposomes were made up of Gd.DOTA.DSA/ DOPC/Cholesterol/DSPE-PEG2000/DOPE-Rhodamine 30/32/ 30/7/1 mol %. Appropriate volumes of each lipid stock were placed in a round-bottom flask (typically 5 mL) containing distilled CHCl3 (500 µL) and stirred to ensure thorough mixing of the lipids. The solvent was slowly removed in vacuo to ensure production of an even lipid film. The film was rehydrated with buffer at a defined volume. The resulting solution was sonicated for 30 min (at 30 °C) for in vivo formulations in order to form gadolinium liposomes of appropriate size. For each preparation, the size of liposomes was measured by photon correlation spectroscopy (PCS) to ensure a size distribution of, on average, below 100 nm. Folate-targeted liposomes were prepared containing a total of ∼4.94 µmol of Gd and control nontargeted liposomes contained ∼9.87 µmol of Gd. Zeta Potential Measurements of Folate-Targeted and Nontargeted Liposomes. Targeted or nontargeted liposomes were added to plastic electrode cuvettes (1 mL, total liposome concentration: 1.2 mg mL-1) and their electrophoretic mobility measured. Measurements were carried out at 25 °C, on a Zetasizer Nanoseries, Nano-ZS.
Bioconjugate Chem., Vol. 20, No. 4, 2009 649
Relaxivity Measurement of Folate-Targeted and Nontargeted Liposomes. Liposomes were prepared in HEPES (20 mM, NaCl, 135 mM, pH 6.5) buffer. These liposomes were prepared and diluted in order to obtain five different Gd atom concentrations ranging from 1.973 to 0.246 mM. The liposome solutions were placed in small Eppendorf tubes (200 µL), imaged at 4.7 T in the coronal plane, T1 values obtained, and the molar relaxivity r1 (mM-1 s-1) determined. T1 values were acquired using saturation recovery experiments performed with a standard spin-echo sequence and a 2 mm single slice acquisition (TR ) 50, 100, 200, 300, 500, 700, 1200, 3000, 5000, 7000 ms, TE ) 15 ms), number of signal averages; 2, FOV; 70 × 70 mm2, collected into a matrix of 256 × 128. Immunohistochemical Analysis of Folate Receptor Expression. At 48 h prior to FACS experiment, adherent IGROV1, OVCAR-3, HeLa, and SKBR-3 cells were grown in RPMI media(folicacidfree)with10%FCSand1%penicillin-streptomycin (Sigma-Aldrich). Cells were grown in 25 cm3 culture flasks (5 × 105 cells per well, 5 mL of growth medium) and in a wet (37 °C) 10% CO2/90% air atmosphere. The cells were grown until 80% confluence. The media was then removed and the cells washed with PBS. The cells were then treated with trypsin-EDTA (500 µL trypsin, 1 mM EDTA) for 2 min at 37 °C, then transferred to 25 mL Falcon tubes and centrifuged (1500 rpm for 5 min). The PBS was removed, and the cells were washed further with PBS and centrifuged (1500 rpm for 5 min). The cells were then incubated in 10% FCS, 1% NaN3 solution (1 mL) for 1 h at 4 °C, then spun down (1500 rpm for 5 min), washed with PBS (1 mL), and then this washing step was repeated twice more. The primary antibody (mAb Mov18/ ZEL, Axxora, UK) selective for the R-FR was then added to each Falcon (5 µg/mL in 3% BSA/PBS (total volume 1 mL)) and the cells incubated at 4 °C for 1 h (for each cell line, one tube contained no antibody). The cells were centrifuged (1500 rpm for 5 min) and washed with PBS (3 × 1 mL, followed by centrifugation), and the secondary antibody (goat antibody IgG, FITC conjugated, Santa Cruz Biotechnology, Inc., USA) was then added to the cells (1:1000 dilution, in 3% BSA/PBS (total volume 1 mL)) and the cells were incubated at 4 °C for a further hour. The cells were then washed with PBS (3 × 1 mL) and centrifuged (400 rpm for 5 min), diluted in 10% FCS/PBS 1% NaN3 (1 mL), and analyzed for their FITC fluorescence on a flow cytometer. Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES) Analysis of Cell Lysates Postincubation with Liposomes. At 48 h prior to gadolinium liposome uptake experiments, adherent IGROV-1 cells were grown in RPMI media (folic acid free) with 10% FCS and 1% penicillin-streptomycin (Sigma-Aldrich) in 48 welled plates. The cells were grown until 80% confluence. The media was then removed and replaced with fresh media (same volume as previous). Folatetargeted and nontargeted liposomes were added in a dose of 12 µg to each well, swirled to ensure even dispersion, and then incubated in a wet (37 °C) 10% CO2/90% air atmosphere for 6 h. After liposome uptake, the cells were washed with PBS (2 × 200 µL) and then treated with lysis buffer (200 µL per well). Samples containing the cell lysate (200 µL) were submitted for ICP-AES measurements of 157Gd content. Briefly, 150 µL of each sample was pipetted and diluted to 3.75 mL with 2% nitric acid to give a total dilution of 1 in 25 and then analyzed on a Varian VIST PRO instrument. All data were corrected for these dilutions, and a blank 2% nitric acid sample was also analyzed. Mouse Tumor Model. IGROV-1 cells (5 × 106/1 mL PBS) were implanted into the flanks of 6-8 week old Balb/c nude mice for generation of subcutaneous tumors. After ∼2 weeks (estimated tumor weights 40-50 mg), the mice were anaesthetized with an isoflurane/O2 mix and placed into a quadrature
650 Bioconjugate Chem., Vol. 20, No. 4, 2009 1
H volume coil and positioned into the magnet. Baseline scans were obtained, and then the mice were injected intravenously via lateral tail vein with a 200 µL liposome solution (HEPES (20 mM, NaCl 135 mM, pH 6.5)) and imaged at 4.7 T (spin-echo sequence: TR ) 400-2800 ms; TE ) 10 ms; FOV ) 45 × 45 cm2; averages, 1; matrix size, 256 × 128; thickness, 2.0 mm; and 20 slices). The mice were imaged prior to injection to obtain baseline images and subsequently placed in the scanner at 2, 16, and 24 h postinjection intervals and scanned further. Histology Experiments. Following MRI, the animals were sacrificed; and the tumors, livers, and kidneys were excised, frozen on liquid nitrogen, embedded in OCT (VWR) embedding fluid, and 10 µm thick sections cut, mounted on slides, and studied by fluorescence microscopy.
RESULTS Folate Receptor Expression. In order to establish whether the IGROV-1 cell line, a human ovarian carcinoma cell line, expresses a sufficient level of the folate receptor, FACS analysis of four different cell lines was carried out. For this purposes, the R-folate receptor (R-FR) isoform, which is a folate transporter with restricted expression levels in normal tissues, was chosen. To measure the R-FR expression levels of the human ovarian cell lines IGROV-1, OVCAR-3, and HeLa (cervical cancer) cells, flow cytometry experiments were carried out. In addition to these cell lines, a breast cancer cell line (SKBR-3) was also analyzed as a negative control cell line with no R-FR expression. Cells were grown in folic acid free media and incubated with serum to block any nonspecific interactions. Immunostaining was carried out with a monoclonal antibody (mAb Mov18/ZEL) specific for the R-FR, and post incubation with this antibody, a secondary FITC labeled antibody (Goat antibody IgG, FITC conjugated), was allowed with the cells. Poststaining, the cells were fixed and analyzed by fluorescence microscopy. From the FACS R-FR expression analysis (Figure 1), where all cell lines were cultured under the same standardized conditions using folate free cell culture medium, it was shown that the IGROV-1 cell line exhibited a distinctly higher level of R-FR expression. From these typical FACS data, the R-FR expression was measured to be in the order IGROV-1 . OVCAR-3 > HeLa > SKBR-3 (three days postseeding). Folate-Targeted Gadolinium Liposomes for Tumor Imaging. Having established the overexpression of the R-FR on the IGROV-1 cell line, folate-targeted gadolinium liposomes were prepared for specific cell receptor binding and uptake into IGROV-1 tumor cancer cells. These folate-targeted bimodal liposomes consisted of Gd.DOTA.DSA/DOPC/Cholesterol/ DSPE-PEG2000/DSPE-PEG(2000)Folate/DOPE-Rhodamine 30/ 32/30/4/3/1 mol % (see Figure 2). The synthesis of Gd.DOTA.DSA, a lipidic T1 MRI contrast agent, has been shown elsewhere (21). Briefly, Gd.DOTA.DSA is made from the metalfree ligand 2-{4,7-bis-carboxymethyl-10-[(N,N-distearylamidomethyl-N′-amidomethyl]-1,4,7,10-tetra-azacyclododec-1-yl}acetic acid (DOTA.DSA) by the addition of a gadolinium source such as gadolinium chloride. DOTA.DSA is formed by amide coupling of N,N-distearylamidomethylamine with 2-(4,7,10tris(2-tert-butyloxycarbonylmethyl)-1,4,7,10-tetra-azacyclododecan-1yl)acetic acid and the subsequent removal of the tert-butyl groups in acidic conditions (21). Nontargeted bimodal control liposomes contained no targeting moiety and consisted of Gd.DOTA.DSA/DOPC/Cholesterol/ DSPE-PEG2000/DOPE-Rhodamine 30/32/30/7/1 mol %. The serum aggregation of these liposomes was also assessed, and no significant change in liposome size was observed over a 24 h incubation period of folate-targeted and nontargeted liposomes in a 1:1 FCS/HEPES mixture at 37 °C (data not shown).
Kamaly et al.
Relaxivity of Folate-Targeted Liposomes. The relaxivities of both the folate-targeted and nontargeted liposomes were measured by formulating liposomes with varying concentrations of the Gd.DOTA.DSA lipid to obtain five formulations with atomic Gd concentrations within the range 1.972 to 0.2466 mM. The r1 and r2 relaxivities measured for the folate-targeted liposomes were 1.30 mM-1 s-1 and 5.38 mM-1 s-1, respectively. For the nontargeted liposomes, r1 and r2 were measured as 0.926 mM-1 s-1 and 5.096 mM-1 s-1, respectively, and were comparable to the targeted liposomes relaxivities. Zeta Potential Measurements. The zeta potentials of both the folate-targeted and nontargeted PEGylated liposomes were also obtained, and these values were found to be close to zero, indicating an overall neutral charge for these liposomes (zeta potential for folate-targeted liposomes was measured to be 3.21 ( 0.284 mV and that for the control nontargeted liposomes was 5.76 ( 0.0673 mV). These liposomes were also formulated using the thin-film technique and sonicated to obtain particles of approximately 100 nm in size with a polydispersity index of
