Antibody-Mediated versus Nontargeted Delivery in a Human Small

and determination of tumor cell fraction uptake was performed in LX-1 tumor xenografts. In vivo study showed that MPEGs-PL-DTPA progressively accumula...
0 downloads 0 Views 114KB Size
Download PDF
184

Bioconjugate Chem. 1998, 9, 184−191

Antibody-Mediated versus Nontargeted Delivery in a Human Small Cell Lung Carcinoma Model Edgardo Marecos, Ralph Weissleder, and Alexei Bogdanov, Jr.* Center for Molecular Imaging Research, Department of Radiology, Massachusetts General Hospital, Building 149, 13 Street, Boston, Massachusetts 02129. Received August 1, 1997; Revised Manuscript Received December 5, 1997

The uptake of macromolecular agents in tumor cells (LX-1, human small cell lung carcinoma) and in corresponding tumor xenografts was compared in a parallel study utilizing a long-circulating biocompatible graft copolymer, MPEGs-PL-DTPA [Bogdanov, A., Jr., et al. (1995) Adv. Drug Delivery Rev. 16, 335-348; Bogdanov, A., Jr., et al. (1996) Bioconjugate Chem. 7, 144-149] and a tumorspecific chimeric monoclonal antibody, BR96 [Hellstrom, I., et al. (1990) Cancer Res. 50, 2183-2190; Garrigues, J., et al. (1993) Am. J. Pathol. 142, 607-622]. Covalent grafted conjugates of methoxy(polyethylene glycol)succinate and polylysine and BR96 were modified with DTPA, biotinyl, or rhodamine-X-residues. Using radionuclide and fluorescent labeled derivatives of the copolymer and the antibody, we established that (1) the copolymer does not associate with the plasma membrane in N-ethylmaleimide-treated cells and is slowly internalized by live cells at 37 °C; (2) the antibody binds rapidly to the surface of LX-1 cells and shows active internalization in vesicles with a subsequent slow decrease in the cell-associated antibody concentration; (3) LX-1 cells bear more than 1 million BR96 binding sites/cell (with an apparent Kd of 4.5 × 10-7 M); and (4) intravesicular fluorescence intensity in LX-1 cells was linearly dependent on copolymer concentration, suggesting fluid phase endocytosis. Tumor localization by nuclear imaging, biodistribution, microdistribution by histology, and determination of tumor cell fraction uptake was performed in LX-1 tumor xenografts. In vivo study showed that MPEGs-PL-DTPA progressively accumulates in the tumor, yielding from 2.8 ( 1.5% injected dose per gram of tissue (ID/g) at 24 h to 5.2 ( 1.7% ID/g of tissue at 48 h. The antibody accumulation peaked at 24 h (6.0 ( 3.2% ID/g) and decreased thereafter. We determined that at 24 h 43.9 ( 11.29% of the polymer accumulated in the tumor was associated with tumor cell fraction with the remainder of the accumulated dose localized in the interstitium. Accumulation of the biotinylated graft copolymer and the antibody in LX-1 xenografts and their uptake in cells were confirmed by histology using avidin-peroxidase staining. Our study demonstrates that, although BR96 is highly specific in vitro, tumoral drug delivery in vivo can be equally high with long-circulating graft copolymers because of slow extravasation at the tumor site.

INTRODUCTION

A variety of tumor targets have been tested experimentally to deliver therapeutic or diagnostic payloads to solid tumors (5, 6). The most widely investigated approach of tumor targeting includes the conjugation of a tumor-specific antibody or antibody fragments to a given drug carrier so that the resultant conjugate specifically associates with tumor cells in vivo. Successful tumor antibody-mediated delivery to several classes of tumorassociated antigens has been achieved: (1) oncogeneencoded proteins [e.g., c-erb proto-oncogene product, an EGF-receptor like protein kinase (7-9)], (2) tissueassociated differentiation antigens, onco-fetoproteins [e.g., carcinoembrionic antigen (10-13)], (3) tumor-associated mucins (14) or glycoproteins [TAG-72 (11, 15)], (4) receptor subtypes [e.g., the folate receptor in adenocarcinomas (16-18)], or (5) antibodies against histones (TNT-1) to detect nuclei in the regions of massive cell necrosis in tumors (19, 20). Several difficulties are * Author to whom correspondence should be addressed at Center for Molecular Imaging Research, Department of Radiology, Massachusetts General Hospital, Bldg 149, Rm 5420, 13th St., Boston, MA 02129. Telephone: (617) 726-5788. Fax: (617) 726-5708. E-mail: [email protected].

associated with this targeting strategy because of three major obstacles: (1) rare occurrence of tumor-specific antigens (expression of antigen in normal tissues), (2) antigen expression heterogeneity of tumors, and (3) generation of anti-ideotype antibodies to the targeting ligand requiring the use of chimeric antibodies or small antibody fragments. Most of the alternative approaches utilize the intravascular delivery to shuttle a given drug to the tumor and exploit elevated permeability of tumor vessels to achieve drug accumulation within the tumor. It has been demonstrated that macromolecular agents, such as linear polymer-drug conjugates (21, 22), accumulate in experimental tumors, presumably at sites of increased capillary permeability (23, 24). The permeability itself can be pharmacologically modulated, thus leading to more efficient intratumoral delivery (reviewed in ref 25). We have previously presented the evidence that MPEGgrafted long-circulating polymers and their conjugates with cisplatin efficiently accumulate in experimental mammary adenocarcinomas (2, 26). The property of tumor accumulation is shared also by other long-circulating drug carriers, such as liposomes bearing MPEGconjugated lipids (27-32). Furthermore, it has been suggested that long-circulating polymeric agents are

S1043-1802(97)00146-8 CCC: $15.00 © 1998 American Chemical Society Published on Web 02/11/1998

Antibody-Mediated versus Nontargeted Delivery

taken up by tumor cells by fluid phase endocytosis (pinocytosis) (26, 33, 34), a process which is typically upregulated in rapidly proliferating cells and is cell cycledependent (35, 36). In this study, we intended to compare directly the “specific” targeting of tumor-specific antibody and the “non-specific” accumulation of a long-circulating polymer in a model of human cancer in vitro and in vivo, i.e., to address the issue of “active” versus “passive” delivery of macromolecules into tumors. MATERIALS AND METHODS

Graft Copolymer. Synthesis of a protected graft copolymer [O-methyl-poly(ethylene glycol)-O′-succinyl-NO′-poly(L-lysine) (MPEGs-PL)1 ] and its analogue modified with DTPA has been accomplished as described in refs 2 and 37. The product was purified by passing the copolymer through Sephadex A-25 and subsequent ultrafiltration (YM 50 membrane, Amicon, Beverly, MA). The duration of ultrafiltration was determined by subjecting an aliquot of purified copolymer to size-exclusion HPLC analysis (Hydropore-5-SEC column, 4.5 mm inside diameter × 25 cm long, Rainin Instrument Co., Woburn, MA) eluted with 50 mM sodium phosphate (pH 6.8) at 0.5 mL/min. The composition of the copolymer was confirmed by elemental analysis (Galbraith Labs, Knoxville, TN). The molecular mass of MPEG-PL was calculated as 410 kDa. To label MPEGs-PL with rhodamineX-isothiocyanate (X-RITC, Molecular Probes, Eugene, OR), a solution of 20 µM MPEGs-PL in 10 mM sodium bicarbonate and 0.15 M NaCl (pH 8.7) was treated with a 5-fold molar excess of X-RITC over MPEGs-PL. The rest of the free amino groups in MPEGs-PL were blocked with DTPA cyclic anhydride. Three or four X-RITC groups/mol of MPEGsPL were incorporated, assuming  ) 92 000 at 572 nm for X-RITC. Biotinylation of MPEGs-PL was performed by reacting 100 mg of O-biotinyl-poly(ethylene glycol) 3500 succinyl N-hydroxysuccinimide (Shearwater Polymers Inc., Birmingham, AL) with 200 mg of poly(L-lysine) [MW of 35 500, at 25 mM lysine monomer in 0.05 M sodium bicarbonate (pH 8.7)] for 3 h with subsequent modification of 30% of N--amino groups of poly(L-lysine) with MPEG 5000 succinate as described previously (2, 37). The biotinylated polymer was purified using ultrafiltration as described before. Indium (111In) labeling of MPEGs-PL-DTPA was performed by incubating MPEGs-PL-DTPA in the presence of 111InCl3 (specific activity of 1.7 × 107 Ci/mol) at a ratio of 50 mCi/mg of copolymer (1 mol of In/600 mol of MPEGs-PL-DTPA) dissolved in 10 mM sodium citrate and 0.15 M NaCl at pH 6.0 (CBS), for 30 min. The yield of isotope chelation to DTPA was 85% from added 111In. The labeled product was purified by gel filtration on Sephadex G-25m spin columns equilibrated with CBS. The properties of the labeled agent have been described in more detail previously (26). Antibody Conjugates. Chimeric monoclonal antibody [Chi BR96 (38)] was kindly provided by P. Senter (Bristol-Myers Squibb Pharmaceutical Research Institute, Seattle, WA). Antibody was dialyzed overnight against 10 mM borate and 0.15 M NaCl (pH 8.7) and 1 Abbreviations: MPEGs-PL-DTPA, O-methyl poly(ethylene glycol)-O′-(succinyl)-N--poly(L-lysyl)diethylenetriaminepentaacetamide; NEM, N-ethylmaleimide; ID, injected dose; hIgG, human immunoglobulin; PBS, 10 mM sodium phosphate and 0.15 M NaCl (pH 7.4).

Bioconjugate Chem., Vol. 9, No. 2, 1998 185

centrifuged at 15000g for 15 min. BR96 modified with DTPA was obtained by reacting a 6-fold molar excess of DTPA cyclic anhydride per mole of antibody (9.4 mg of BR96/mL, final concentration) for 4 h with subsequent overnight dialysis against 10 mM sodium citrate and 0.15 M NaCl (pH 6) and Sephadex G-25m chromatography. Labeling of the BR96-DTPA with 111In has been performed as described above. According to TNBS titration, three DTPA residues were conjugated with mAb on the average. Labeling with X-RITC has been performed using a ratio of 6 mol of XRITC/mol of mAb at pH 8.7. Fluorescent antibody was purified from aggregates by centrifugation at 15000g for 30 min and passed through a column of Sephadex G-25m. The resultant preparation contained 1.2 mol of X-RITC/mol of antibody. The biotinylation of BR96 was performed with a 3-fold molar excess of biotinylation reagent per mole of antibody by reacting 65 µM N-hydroxysulfosuccinimidyl[6-(biotinylamido)]hexanoate (Pierce Chemical Co.) with BR96 in 50 mM sodium bicarbonate and 0.1 M NaCl (pH 8). The biotinylated antibody was then purified by Sephadex G-25m column chromatography. Nonspecific human IgG (Sigma Chemical Co. catalog no. I-4506) was purified from the aggregates and modified under conditions identical to those described above. The purity of modified antibodies was verified by SDS-PAGE using 7.5 and 10% gels and autoradiography. Cell Culture Experiments. Human small cell lung carcinoma LX-1 (kindly provided by E. Neuwelt, University of Oregon, Portland, OR) was propagated in RPMI and 10% FBS. Cells were detached from the plastic by aspiration, sedimented, and resuspended in fresh medium at 2 ×105 cells/mL. Indium-labeled MPEGs-PL-DTPA and BR96-DTPA were added to 1 mL of cell suspension at the concentration of 0.1-70 µg/mL and incubated with agitation at 37 °C. Cells were washed three times by centrifugation through a cushion of 30% Histopaque-1077 in HBSS (Sigma Chemical Co.) at 800g, for 10 min, and counted. The associated radioactivity was determined by γ-counting. Fluorescent microscopy was performed at 37 °C using live cells attached to the glass coverslips (39). Images were collected using a SenSys (Photometrics) CCD camera attached to a Zeiss Axiovert 100TV fluorescent microscope equipped with an XF22 filter set (Omega Optical Inc.). For intravesicular fluorescence measurements, images of 50-100 cells were digitized and segmented using a histogram method provided by IP LabSpectrum software (Signal Analytics Co., Vienna, VA). Fluorescence intensity of segments (intracellular vesicles, n ) 100-200 per data point) was measured. Segment areas were limited to a median of 10-20 arbitrarily chosen vesicles to exclude nonvesicular structures and cell fragments from consideration. The linearity of the CCD response was checked using standard X-RITC solutions. The response of the device was linear in the range of 503000 nM X-RITC. Animal Model of Human Lung Carcinoma. Female nude mice (n ) 36; 25-30 g; Edwin L. Steele Laboratory, MGH, Boston, MA) bearing 0.5-0.9 cm LX-1 human small cell lung carcinomas were used as a tumor model for all comparative experiments. In each animal, 1 × 106 LX-1 human lung small cell carcinoma cells were implanted subcutaneously in the left abdominal flank. Mice were used for biodistribution studies 15-17 days after implantation. Twelve were used for imaging studies. Six were used to process tumors for avidin-peroxidase histology. Six were used for the cell uptake experiment.

186 Bioconjugate Chem., Vol. 9, No. 2, 1998

Figure 1. Binding of BR96-[111In]DTPA and MPEGs-PL[111In]DTPA to LX-1 cells (1 h, 37 °C) treated in the presence of 100 µM N-ethylmaleimide: (1) BR96-DTPA and (2) MPEGsPL-DTPA. (Inset) Scatchard analysis of curve 1. Results are presented as mean ( SD (n ) 3).

Animal Experiments. Biodistribution studies in 36 mice were performed after intravenous (iv) injection of the agents via the tail vein. Animals were divided into two experimental groups of 18. Animals were injected in the tail vein with 200 µg of BR96-[111In]DTPA (60 nmol of mAb/kg, 30 µCi/animal) dissolved in 100 µL of sterile HBSS or with an equimolar amount of MPEGs-PL[111In]DTPA (ca. 60 µg of polymer/animal, 25 µCi/animal). In each group, sets of three animals were sacrificed at 6, 12, 48, and 72 h time points, and a set of six was sacrificed at the 24 h time point. Samples of blood, major normal tissues, and tumors were collected, and the tissues were blotted dry and weighed. Radioactivity of the samples was measured using a well-type γ-counter (LKB model 1282, Wallac, Turku, Finland). To correct for radioactive decay and to permit the calculation of activity in each organ as a fraction of the administered dose, aliquots of standard stock solutions were counted simultaneously. Biodistribution results were expressed as a percentage of the injected dose per gram of tissue (% ID/g). Tumor Fractionation Experiment. To determine the intracellular and extracellular fractions of the graft copolymer and monoclonal antibody within tumors, the tumor tissue obtained from six animals was processed (three for each group, 24 h postadministration). The tumor (200-300 mg of tissue) was excised, minced to small fragments on ice, suspended in 20 mL of PBS supplemented with 1 mM PMSF, 5 mM NaN3, and 1 mM N-ethylmaleimide (pH 7), and passed three times through #4 wire mesh (Sigma Chemical Co.) at 4 °C. Samples were centrifuged at 18000g for 15 min. The integrity of cells released from the tissue was verified with the trypan blue exclusion method. The radioactivity of the pellet was determined in a γ-counter and was assumed to be associated with intracellular compartment, and radioactivity of the supernatant was assumed to be contained within the extracellular fluid (40). Scintigraphy. Mice (n ) 12, 6 in each group) were used for imaging 6, 12, and 24 h postinjection (2 of the animals in each group were imaged at all time points). Mice were anesthetized subcutaneously (sc) using a mixture of 90 mg/mL ketamine (Ketalar, Parke Davis, Morris Plains, NJ) and 9 mg/mL xylazine (Rompun, Miles, Shawnee Mission, KA). Images were acquired by using 20% windows over a 247 keV photopeak of 111In (10 min at 6 h and 20 min at 12 and 24 h). All scintigrams were obtained by using a small-field of view

Marecos et al.

Figure 2. Time course of the uptake of BR96-[111In]DTPA (open symbols) and MPEGs-PL-[111In]DTPA (closed symbols) in LX-1 cells. Cells were incubated for the time indicated in the presence of equimolar concentrations (0.4 µM) of BR96 (60 µg/ mL) and the copolymer (165 µg/mL) at 37 °C. Results are presented as mean ( SD (n ) 3).

γ-camera equipped with a pinhole collimator (Sigma 410, Ohio Nuclear, Solon, OH) and analyzed using NucLear Mac 2.9 software (Scientific Imaging Inc.). Histology. Mice (nu/nu, n ) 6) were divided into two groups. The first group of animals received iv ca. 100 µg of BR96-biotin/animal. The animals of the second group were injected with 100-200 µg of MPEGs-PLbiotin. Animals were sacrificed at 24 h postinjection. Tumors were excised, and tumor tissue was quickly frozen and cryosectioned into 10-20 µm sections. Sections were fixed for 1 h in a mixture of ethyl acetate/ acetone/acetic acid (6/3/1), washed with PBS, blocked in solution of 10% goat serum in PBS, and incubated with the avidin-peroxidase complex (ABC kit, Vector Laboratories, Burlingame, CA) as described by the manufacturer. Sections were stained using diaminobenzidine/ hydrogen peroxide and counterstained with hematoxylin/ eosin. RESULTS

Interaction of the Copolymer and BR96 with LX-1 Cells. These experiments were conducted to determine whether the binding to plasma membrane may mediate the uptake of both MPEGs-PL-DTPA and BR96 in LX-1 cells. To abolish internalization, cells were preincubated in the presence of 100 mM NEM, an inhibitor of endocytosis (41). Indium-labeled BR96-DTPA showed a typical binding to N-ethylmaleimide-treated cells (Figure 1). Using Scatchard analysis, we established that LX-1 cells bear ca. 1 million BR96-DTPA binding sites/cell, with an apparent Kd of 4.51 × 10-7 M. The copolymer, MPEGs-PL-[111In]DTPA, incubated with cells under identical conditions showed no detectable binding to the cell surface (Figure 1). Kinetics of the labeled compound uptake by LX-1 cells was studied in live cells at equimolar concentrations (Figure 2). Labeled antibody was rapidly associated with LX-1 cells (16 pmol of antibody was taken up by 1 million cells during the first hour of incubation). The radioactivity associated with mAb-labeled cells slowly diminished with time. Graft copolymer was gradually taken up by cells, with the relatively fast initial phase (6 ng/h per 1 million cells) and slower later accumulation (Figure 2).

Antibody-Mediated versus Nontargeted Delivery

Bioconjugate Chem., Vol. 9, No. 2, 1998 187

Figure 4. Concentration dependence of vesicular fluorescence intensity measured by a CCD device in LX-1 cells: BR96-XRITC (1, open symbols) and MPEGs-PL-X-RITC (2, closed symbols). Cells were incubated with labeled compounds for 3 h at 37 °C and subjected to microfluorometry as described in Materials and Methods. Figure 3. Fluorescent microscopy of live LX-1 cells incubated in the presence of BR96-X-RITC (1 and 3) or MPEGs-PL-X-RITC (2 and 4) for 15 min (1 and 2) or 3 h (3 and 4) at 37 °C. Equal concentrations of fluorophore were used (500 nM).

Table 1. Biodistribution of 111In-Labeled MPEGs-PL-DTPA and 111In-Labeled BR96-DTPA in Mice Bearing Human Small Cell Lung Carcinoma Xenografts (in % ID/g) 24 h

Approximately 150 ng (ca. 0.1% of added copolymer) was associated with 2 × 105 cells after incubation for 15 h at 37 °C, whereas a total of 2500 ng of BR96 (16.7 pmol, 4.1% of the added antibody) was associated with the same number of cells. Using fluorescence microscopy and rhodamine-labeled BR96 antibody, we observed a characteristic fluorescent “halo” indicating association of antibodies with the cell surface at 15 min (Figure 3, panel 1). The majority of cells exhibited typical extracellular fluorescence with concomitant uptake in vesicles. At 3 h, most of the cellassociated fluorescence appeared to be distributed inside the vesicles (Figure 3, panel 3). For MPEGs-PL-DTPA, virtually no association of the polymer with plasma membrane was seen at the initial stage of incubation (Figure 3, panel 2). At 3 h, however, the morphology of intracellular distribution of fluorescence was almost identical to that of BR96 cells (Figure 3, panel 4). The main difference detected was in the number of fluorescent vesicles per cell which was lower in the case of the copolymer. To obtain a semiquantitative measure of the uptake of both fluorescently labeled BR96 and MPEGPL, we digitized fluorescent images of intracellular vesicles. This made it feasible to study the concentration dependence of the uptake only and allowed the exclusion of the fraction of plasma membrane-bound antibody or copolymer that was not internalized. Normalized vesicular fluorescence intensity of the cells incubated in the presence of MPEGs-PL-X-RITC showed a nearly linear increase with the concentration of fluorophore (Figure 4, closed symbols). Vesicular fluorescence intensity of BR96-treated LX-1 cells was very similar to that of cells treated with the copolymer in the broad range of rhodamine concentrations (Figure 4). In the higher concentration range of BR96 (>20 µg/mL, >1500 nM rhodamine-X), LX-1 cells showed signs of cytotoxicity which were manifested by the presence of fragmented cells. The cytotoxicity resulted in a dramatic increase of fluorescence (Figure 4, open symbols). Biodistribution. Biodistribution experiments were performed to compare kinetics of 111In-labeled BR96DTPA and MPEGs-PL-DTPA accumulation in LX-1 tumor xenografts (Table 1). At the dose used for intra-

tissue blood tumor liver lung heart spleen kidney muscle pancreas intestine fat bone marrow

mAb

MPEGsPL-DTPA

48 h mAb

MPEGsPL-DTPA

hIgG

7.7 ( 2.2 10.3 ( 1.8 4.0 ( 0.5 8.8 ( 0.3 9.3 ( 0.5 6.0 ( 3.2 2.8 ( 1.5 5.8 ( 0.4 5.2 ( 1.7 4.1 ( 1.4 5.8 ( 3.7 2.7 ( 1.5 4.7 ( 0.8 3.0 ( 0.3 5.5 ( 2.9 2.1 ( 1.0 3.4 ( 1.5 2.0 ( 0.2 3.7 ( 0.4 4.9 ( 3.0 2.3 ( 0.3 4.6 ( 1.0 1.5 ( 0.3 3.9 ( 0.8 2.4 ( 0.4 4.2 ( 3.0 3.5 ( 2.1 3.0 ( 0.5 3.2 ( 0.5 3.9 ( 1.9 6.3 ( 2.1 4.2 ( 1.0 5.0 ( 0.8 4.5 ( 0.5 4.7 ( 1.8 0.8 ( 0.4 1.6 ( 1.6 0.8 ( 0.0 1.1 ( 0.2 2.0 ( 0.1 1.0 ( 0.3 0.6 ( 0.4 1.1 ( 0.1 1.0 ( 0.1 1.5 ( 0.3 0.9 ( 0.7 1.6 ( 1.4 1.3 ( 0.4 1.9 ( 0.7 1.2 ( 1.0 1.2 ( 1.7 0.5 ( 0.3 0.6 ( 0.5 0.8 ( 0.3 1.1 ( 0.8 0.3 ( 0.1 0.7 ( 0.4 0.9 ( 0.4 1.3 ( 0.3 1.7 ( 1.1

venous injections (60 nmol/kg), both 111In-labeled copolymer and antibody showed very long blood residence times in nude mice, with a T1/2 of 26.6 h for mAb (monoexponential fitting, r2 ) 0.925). The blood half-life of the copolymer could be approximated as 60 h. Indium-labeled BR96-DTPA accumulated in the tumor rapidly and exhibited the highest accumulation 6-12 h after the injection at 7.1-7.9% of the injected dose per gram of tissue. Graft copolymer showed a slower uptake in the tumor with the maximum accumulation reached at 48 h (Table 1). Corresponding γ-camera images reveal a clear tumor mass which could be detected by the antibody at 6 h postadministration (Figure 5), whereas the copolymer accumulation was evident at 12 h postinjection. Over time, BR96-DTPA showed good retention within the tumor with high tumor/muscle localization ratios at 48-72 h post-IV administration. The copolymer was taken up by the tumor ca. 5-fold more efficiently than in the muscle at the peak of tumor accumulation (at 48 h). At this time point, we did not observe any statistically significant differences between tumor/muscle ratios calculated for BR96 (range of 6.75-7.25) and MPEG-PL-DTPA (range of 2.7-7.7) (Table 1). Both ratios were higher than the tumor/muscle ratio of nonspecific hIgG (1.3-2.6). Compartmental Distribution of Labeled Agents in the Tumor. Tumor compartmental distribution of labeled BR96-DTPA and MPEGs-PL-DTPA was determined by separating the cellular fraction from the extracellular fluid. Twenty-four hours after injection,

188 Bioconjugate Chem., Vol. 9, No. 2, 1998

Marecos et al.

Figure 5. Nuclear imaging of mice bearing LX-1 xenografts. Animals were injected iv with MPEGs-PL-[111In]DTPA (images 1-3) or BR96-[111In]DTPA (images 4-6). Images collected at 6 (images 1 and 4), 12 (images 2 and 5), and 24 (images 3 and 6) h postinjection as described in Materials and Methods. Arrows indicate localization of the tumor.

43.9 ( 11.3% of the total amount of 111In-labeled copolymer located in tumor was associated with the cellular fraction. The remainder of the dose localized in the tumor was found in the extracellular compartment. The cell-associated fraction of the injected dose was 60.4 ( 3.9% of mAb, and 39.6 ( 3.6% of the antibody was located in the extracellular compartment. Histological examination of the tumor obtained from the animal injected with MPEGs-PL-biotin (Figure 6A) and biotinylated antibody (Figure 6B) showed the presence of extravasation of both antibody and the polymer in perivascular stroma and certain accumulation in cells located in the perivascular area. DISCUSSION

This study was undertaken with the goal of comparing, in a set of parallel experiments, the ability of a cancer cell-specific monoclonal antibody (BR96) and a nonspecific long-circulating polymer to localize within cancer cells (LX-1) in vitro and in vivo. The tumor-specific monoclonal antibody used in this study binds to a variant of a Ley antigen abundantly expressed (>200000 molecules per cell) at the plasma membrane of human colon, lung, breast, and ovary carcinomas (3, 42). This antibody rapidly internalizes in human carcinoma cells and is cytotoxic in vitro (3, 4). Intravenous administration of BR96 results in antitumor activity against tumor xenografts in mice (38). The copolymer, MPEGs-PL-DTPA, was used because of its controllable size, relatively narrow molecular mass distribution, and long circulation time (1, 43). Furthermore, the copolymer is relatively inert in nature and is not immunogenic (37). Additionally, the copolymer does not induce local changes in capillary permeability and accumulates within tumors in experimental cancer (2, 26). The latter property was explained by the extravasation from vessels with abnormally elevated permeability (44, 45). It has been shown before that intravenous administration of tumor-specific antibodies may result in a very high

Figure 6. Extravasation and uptake of biotinylated BR96 (A) and biotin-MPEGs-PL-DTPA (B) in LX-1 xenografts by histology. Staining of tumor slices was performed with the avidinperoxidase complex with hematoxylin/eosin counterstaining. Arrows indicate extravasation sites. Asterisks indicate vascular lumens. (Insets) Magnified image of cells positively stained by avidin-peroxidase.

percentage of dose accumulated in tumor tissue, and that this effect is not accompanied by efficient binding to tumor cells. Tumor-specific antibodies accumulate in vivo in tumoral perivascular stroma (40, 46), and the majority of the dose (76-80%) is not bound to cells, but rather is associated with tumor-shed antigens (40). The discovery of interstitial fluid pressure gradients in tumors has also been used to explain the low efficiency of cellular uptake of tumor-specific antibodies (47-49). These findings prompted us to compare delivery to tumors and cellular uptake of a tumor-specific antibody and a copolymer which can associate with tumor cells only as a result of endocytosis. For “inert” particulates and macromolecules which do not bind to cell surfaces, constitutive fluid phase endocytosis is the predominant pathway of uptake (33, 34, 50-52). Therefore, we hypothesized that MPEG-PL-DTPA may accumulate in tumor cells as a result of fluid phase uptake. The results of in vitro experiments clearly indicate that MPEGs-PL-DTPA does not bind to the LX-1 cell plasma membrane, whereas BR96, labeled with either DTPA or rhodamine-X (Figures 1 and 3), binds to the plasma membrane at a very high density (1 million sites/cell were detected) and rapidly internalizes in live cells. After NEM treatment which blocks fluid phase uptake (41), the cells avidly bound the antibody while no binding of polymer was observed (Figure 1). Fluorescence microscopy suggested that nearly all association of MPEGs-PL with LX-1 cells is a result of fluid phase uptake; the copolymer appears within the cells at a much later stage than surface-bound antibody and associates almost ex-

Antibody-Mediated versus Nontargeted Delivery

clusively with intracellular vesicles (Figure 3). In addition, semiquantitative intravesicular fluorescence measurements revealed a linear dependence of fluorescence intensity on fluorophore concentration. This observation is consistent with the fluid phase uptake (Figure 4). Vesicular fluorescence of BR96-treated cells showed certain level of saturation within the range of 250-750 nM rhodamine-X, presumably because of the saturation of antibody binding sites and the predominance of membrane-bound antibodies in the vesicles. With the further increase in antibody concentration, fluorescence intensity of vesicles increased as a result of cell permeabilization by the cytotoxic antibody. It should be noted that an internalized molecule encounters a vesicular environment different from that of an extracellular medium in pH, microviscosity, and high protein concentrations. These factors can affect the quantum yield of the fluorescence of the given probe. It is also possible that the antibody and the polymer can be taken up and sorted into different vesicular compartments with different degrees of “maturation” and compaction (53). This may lead to different levels of fluorescence quenching. However, the results of measurements performed using radioactive BR96 and copolymer (Figure 2) were consistent with fluorescent microscopy data, revealing rapid association of the antibody with the cells and slow uptake of MPEGs-PL-DTPA. The results of the in vivo biodistribution study performed using an LX-1 xenograft model and the antibodycopolymer pair were remarkably similar to in vitro data. For example, tumor accumulation kinetics and corresponding tumor/muscle ratios were consistent with rapid accumulation of BR96 in cells (Figure 5). MPEGs-PLDTPA, lacking the ability to bind to tumor cells, showed a delayed accumulation in the tumor. At the 48 h time point, which corresponds to the peak of accumulation of MPEGs-PL-DTPA (Table 1), we observed only an insignificant difference between the antibody and copolymer accumulation (5.8 ( 0.4 vs 5.2 ( 1.7% ID/g, P > 0.05). These values were only marginally higher than that of hIgG (Table 1) but resulted in substantially higher localization (tumor/muscle ratios). Both BR96 and the copolymer were present in tumor interstitium and in tumor cells in animals injected with biotinylated mAb and MPEGs-PL-DTPA (Figure 6). Strongly positive cells (Figure 6, insets) were detected in both cases, suggesting the internalization of biotinylated copolymer and BR96 in vivo. As expected, higher amounts of the antibody were found to be associated with LX-1 cells in xenografts (more than 60% of tumor-associated antibody) than the copolymer (44% of the tumor-accumulated dose) at 24 h. This may be explained by rapid initial uptake of mAb in the tumor. We believe that the high tumor cell-associated activity cannot be explained by the uptake of plasma components labeled as a result of transchelation of 111In from the copolymer to plasma proteins since 111 In-labeled MPEGs-PL-DTPA is stable in plasma as well as in the presence of apotransferrin for at least 48 h (26). Our data indicate that nonspecific fluid phase uptake in tumor cells is the major pathway of tumor-directed delivery of nontargeted macromolecular agents exemplified in our study by the MPEGs-PL graft copolymer. There are obvious differences in the kinetics of tumor accumulation of BR96 and MPEGs-PL graft copolymer resulting in rapid tumor-specifc accumulation of the tumor-specific antibody. However, we showed that one can achieve a progressive increase of tumor delivery with increasing blood half-lives of the drug carriers with no

Bioconjugate Chem., Vol. 9, No. 2, 1998 189

specificity against tumor cells. Approximately 100 molecules of diagnostic label (e.g., DTPA) or an antitumor drug can be conjugated to a nonimmunogenic MPEGsPL copolymer. Thus, the nontargeted approach can potentially yield a high accumulation of the drug in the tumor and concomitant uptake by proliferating cancer cells, irrespective of the tumor antigenic status. ACKNOWLEDGMENT

This work was supported in part by the Center for Innovative Minimally Invasive Therapy, Massachusetts General Hospital (Fellowship Award to E.M. and Research Award to A.B.) and NIH Grant RO1 CA59649-01 (to R.W.). LITERATURE CITED (1) Bogdanov, A., Jr., Weissleder, R., and Brady, T. (1995) Longcirculating blood pool imaging agents. Adv. Drug Delivery Rev. 16, 335-348. (2) Bogdanov, A., Jr., Martin, C., Bogdanova, A. V., Brady, T. J., and Weissleder, R. (1996) An adduct of cis-diamminedichloroplatinum(II) and poly(ethylene glycol)poly(L-lysine)-succinate: synthesis and cytotoxic properties. Bioconjugate Chem. 7, 144-149. (3) Hellstrom, I., Garrigues, H. J., Garrigues, U., and Hellstrom, K. E. (1990) Highly tumor-reactive, internalizing, mouse monoclonal antibodies to Le(y)-related cell surface antigens. Cancer Res. 50, 2183-2190. (4) Garrigues, J., Garrigues, U., Hellstrom, I., and Hellstrom, K. E. (1993) Ley specific antibody with potent anti-tumor activity is internalized and degraded in lysosomes. Am. J. Pathol. 142, 607-622. (5) Buchsbaum, D. J. (1995) Experimental approaches to increase radiolabeled antibody localization in tumors. Cancer Res. 55, 5729s-5732s. (6) von Mehren, M., and Weiner, L. M. (1996) Monoclonal antibody-based therapy. Curr. Opin. Oncol. 8, 493-498. (7) Adams, G. P., McCartney, J. E., Tai, M. S., Oppermann, H., Huston, J. S., Stafford, W. F. d., Bookman, M. A., Fand, I., Houston, L. L., and Weiner, L. M. (1993) Highly specific in vivo tumor targeting by monovalent and divalent forms of 741F8 anti-c-erbB-2 single-chain Fv. Cancer Res. 53, 40264034. (8) George, A. J., Jamar, F., Tai, M. S., Heelan, B. T., Adams, G. P., McCartney, J. E., Houston, L. L., Weiner, L. M., Oppermann, H., Peters, A. M., et al. (1995) Radiometal labeling of recombinant proteins by a genetically engineered minimal chelation site: technetium-99m coordination by single-chain Fv antibody fusion proteins through a C-terminal cysteinyl peptide. Proc. Natl. Acad. Sci. U.S.A. 92, 83588362. (9) Dean, C. J., Eccles, S. A., Valeri, M., Box, G., Allan, S., McFarlane, C., Sandle, J., and Styles, J. (1993) Rat MAbs to the product of the c-erbB-2 proto-oncogene for diagnosis and therapy in breast cancer. Cell Biophys. 22, 111-127. (10) Biersack, H. J., Briele, B., Hotze, A. L., Oehr, P., Qian, L., Mekkawy, M. A., and Shih, W. J. (1992) The role of nuclear medicine in oncology. Ann. Nucl. Med. 6, 131-136. (11) Abdel-Nabi, H. H., and Doerr, R. J. (1992) Multicenter clinical trials of monoclonal antibody B72.3-GYK-DTPA 111In (111In-CYT-103; OncoScint CR103) in patients with colorectal carcinoma. Targeted Diagn. Ther. 6, 73-88. (12) Leitha, T., Walter, R., Schlick, W., and Dudczak, R. (1991) 99mTc-anti-CEA radioimmunoscintigraphy of lung adenocarcinoma. Chest 99, 14-19. (13) Saleh, M. N., Wheeler, R. H., Lee, J. Y., Khazaeli, M. B., Unger, M. W., Russell, C. H., and LoBuglio, A. F. (1991) In111 labeled monoclonal anti-carcinoembryonic antigen antibody (ZCE025) in the immunoscintigraphic imaging of metastatic antigen-producing adenocarcinomas. Clin. Nucl. Med. 16, 110-116. (14) Britton, K. E., Granowska, M., and Shepherd, J. (1985) Localization of cancer of the ovary and metastases using 123Ilabeled monoclonal antibody HMFG-2 compared to surgical

190 Bioconjugate Chem., Vol. 9, No. 2, 1998 findings. In Monoclonal Antibodies for Cancer Detection and Therapy, p 10, Academic Press, New York. (15) Surwit, E. A., Childers, J. M., Krag, D. N., Katterhagen, J. G., Gallion, H. H., Waggoner, S., and Mann, W. J., Jr. (1993) Clinical assessment of 111In-CYT-103 immunoscintigraphy in ovarian cancer. Gynecol. Oncol. 48, 285-292. (16) Holm, J., Hansen, S. I., Hoier-Madsen, M., Sondergaard, K., and Bzorek, M. (1994) Folate receptor of human mammary adenocarcinoma. APMIS 102, 413-419. (17) Franklin, W. A., Waintrub, M., Edwards, D., Christensen, K., Prendegrast, P., Woods, J., Bunn, P. A., and Kolhouse, J. F. (1994) New anti-lung-cancer antibody cluster 12 reacts with human folate receptors present on adenocarcinoma. Int. J. Cancer Suppl. 8, 89-95. (18) Coney, L. R., Mezzanzanica, D., Sanborn, D., Casalini, P., Colnaghi, M. I., and Zurawski, V. R., Jr. (1994) Chimeric murine-human antibodies directed against folate binding receptor are efficient mediators of ovarian carcinoma cell killing. Cancer Res. 54, 2448-2455. (19) Epstein, A. L., Khawli, L. A., Chen, F.-M., Hu, P., Glasky, M. S., and Taylor, C. R. (1995) Tumor necrosis imaging and treatment of solid tumors. In Targeted Delivery of Imaging Agents, pp 259-288, CRC Press, Boca Raton, FL. (20) Epstein, A. L., Chen, F.-M., and Taylor, C. R. (1988) A novel method for the detection of necrotic lesions in human cancers. Cancer Res. 48, 5842. (21) Pimm, M. V., Perkins, A. C., Strohalm, J., Ulbrich, K., and Duncan, R. (1996) Gamma scintigraphy of a 123I-labelled N-(2hydroxypropyl)methacrylamide copolymer-doxorubicin conjugate containing galactosamine following intravenous administration to nude mice bearing hepatic human colon carcinoma. J. Drug Targeting 3, 385-390. (22) Pimm, M. V., Perkins, A. C., Strohalm, J., Ulbrich, K., and Duncan, R. (1996) Gamma scintigraphy of the biodistribution of 123I-labelled N-(2-hydroxypropyl)methacrylamide copolymer-doxorubicin conjugates in mice with transplanted melanoma and mammary carcinoma. J. Drug Targeting 3, 375-383. (23) Jain, R. (1988) Determinants of tumor blood flow: A review. Cancer Res. 48, 2641-2658. (24) Trotter, M., Chaplin, D., Durand, R., and Olive, P. (1989) The use of fluorescent probes to identify regions of transient perfusion in murine tumors. Int. J. Radiat. Oncol., Biol., Phys. 16, 931-934. (25) Maeda, H., and Matsumura, Y. (1989) Tumoritropic and lymphotropic principles of macromolecular drugs. Crit. Rev. Ther. Drug Carrier Syst. 6, 193-210. (26) Bogdanov, A., Jr., Wright, S. C., Marecos, E. M., Bogdanova, A. V., Martin, C., Petherick, P., and Weissleder, R. (1997) A long-circulating co-polymer in “passive targeting” to solid tumors. J. Drug Targeting 4, 321-330. (27) Yuan, F., Leunig, M., Huang, S. K., Berk, D. A., Papahadjopoulos, D., and Jain, R. K. (1994) Microvascular permeability and interstitial penetration of sterically stabilized (stealth) liposomes in a human tumor xenograft. Cancer Res. 54, 3352-3356. (28) Yuan, F., Dellian, M., Fukumura, D., Leunig, M., Berk, D. A., Torchilin, V. P., and Jain, R. K. (1995) Vascular permeability in a human tumor xenograft: molecular size dependence and cutoff size. Cancer Res. 55, 3752-3756. (29) Oku, N., Namba, Y., and Okada, S. (1992) Tumor accumulation of novel RES-avoiding liposomes. Biochim. Biophys. Acta 1126, 255-260. (30) Litzinger, D. C., Buiting, A. M., van Rooijen, N., and Huang, L. (1994) Effect of liposome size on the circulation time and intraorgan distribution of amphipathic poly(ethylene glycol)-containing liposomes. Biochim. Biophs. Acta 1190, 99107. (31) Zou, Y., Ling, Y. H., Reddy, S., Priebe, W., and Perez-Soler, R. (1995) Effect of vesicle size and lipid composition on the in vivo tumor selectivity and toxicity of the non-cross-resistant anthracycline annamycin incorporated in liposomes. Int. J. Cancer 61, 666-671. (32) Francis, G. E., Delgado, C., Fisher, D., Malik, F., and Agrawal, A. K. (1996) Polyethylene glycol modification:

Marecos et al. relevance of improved methodology to tumour targeting. J. Drug Targeting 3, 321-340. (33) Pratten, M. K., Duncan, R., Cable, H. C., Schnee, R., Ringsdorf, H., and Lloyd, J. B. (1981) Pinocytic uptake of divinyl ether-maleic anhydride (pyran copolymer) and its failure to stimulate pinocytosis. Chem. Biol. Interact. 35, 319330. (34) Duncan, R., Pratten, M. K., Cable, H. C., Ringsdorf, H., and Lloyd, J. B. (1981) Effect of molecular size of 125I-labelled poly(vinylpyrrolidone) on its pinocytosis by rat visceral yolk sacs and rat peritoneal macrophages. Biochem. J. 196, 4955. (35) Quintart, J., Leroy-Houyet, M. A., Trouet, A., and Baudhuin, P. (1979) Endocytosis and chloroquine accumulation during the cell cycle of hepatoma cells in culture. J. Cell Biol. 82, 644-653. (36) Illinger, D., Italiano, L., Beck, J., Waltzinger, C., and Kuhry, J. (1993) Comparative evolution of endocytosis levels and of the cell surface area during the L929 cell cycle: a fluorescence study with TMA.-DPH. Biol. Cell. 79, 265-268. (37) Bogdanov, A. A., Jr., Weissleder, R., Frank, H. W., Bogdanova, A. V., Nossif, N., Schaffer, B. K., Tsai, E., Papisov, M. I., and Brady, T. J. (1993) A new macromolecule as a contrast agent for MR angiography: preparation, properties, and animal studies. Radiology 187, 701-706. (38) Schreiber, G. J., Hellstrom, K. E., and Hellstrom, I. (1992) An unmodified anticarcinoma antibody, BR96, localizes to and inhibits the outgrowth of human tumors in nude mice. Cancer Res. 52, 3262-3266. (39) Schulze, E., Ferrucci, J., Poss, K., Lapointe, L., Bogdanova, A., and Weissleder, R. (1995) Cellular uptake and trafficking of a prototypical magnetic iron oxide label in vitro. Invest. Radiol. 30, 604-610. (40) Lin, K., Nagy, J. A., Xu, H., Shockley, T. R., Yarmush, M. L., and Dvorak, H. F. (1994) Compartmental distribution of tumor-specific monoclonal antibodies in human melanoma xenografts. Cancer Res. 54, 2269-2277. (41) Schnitzer, J. E., Allard, J., and Oh, P. (1995) NEM inhibits transcytosis, endocytosis, and capillary permeability: implication of caveolae fusion in endothelia. Am. J. Physiol. 268, H48-H55. (42) Trail, P., Willner, D., Lasch, S., Henderson, A., Hofstead, S., Casazza, A., Firestone, R., Hellstro¨m, I., and Hellstro¨m, K. (1993) Cure of xenografted human carcinomas by BR96doxorubicin immunoconjugates. Science 261, 212-215. (43) Bogdanov, A., Jr., Callahan, R., Wilkinson, R., Martin, C., Cameron, J., Fishman, A., Brady, T., and Weissleder, R. (1994) A synthetic copolymer kit for radionuclide blood pool imaging. J. Nucl. Med. 35, 1880-1886. (44) Gupta, H., Weissleder, R., Bogdanov, A. A., Jr., and Brady, T. J. (1995) Experimental gastrointestinal hemorrhage: detection with contrast-enhanced MR imaging and scintigraphy. Radiology 196, 239-244. (45) Gupta, H., Wilkinson, R. A., Bogdanov, A. A., Jr., Callahan, R. J., and Weissleder, R. (1995) Inflammation: imaging with methoxy poly(ethylene glycol)-poly-L-lysine-DTPA, a longcirculating graft copolymer. Radiology 197, 665-669. (46) Shockley, T. R., Lin, K., Nagy, J. A., Tompkins, R. G., Yarmush, M. L., and Dvorak, H. F. (1992) Spatial distribution of tumor-specific monoclonal antibodies in human melanoma xenografts. Cancer Res. 52, 367-376. (47) Baxter, L. T., and Jain, R. K. (1991) Transport of fluid and macromolecules in tumors. IV. A microscopic model of the perivascular distribution. Microvasc. Res. 41, 252-272. (48) Baxter, L. T., and Jain, R. K. (1991) Transport of fluid and macromolecules in tumors. III. Role of binding and metabolism. Microvasc. Res. 41, 5-23. (49) Jain, R. K. (1990) Physiological barriers to delivery of monoclonal antibodies and other macromolecules in tumors. Cancer Res. 50, 814-819. (50) Duncan, R., Starling, D., Rypacek, F., Drobnik, J., and Lloyd, J. B. (1982) Pinocytosis of poly(R,β-(N-2-hydroxyethyl))DL-aspartamide and a tyramine derivative by rat visceral yolk sacs cultured in vitro. Ability of phenolic residues to enhance the rate of pinocytic capture of a macromolecule. Biochim. Biophys. Acta 717, 248-254.

Antibody-Mediated versus Nontargeted Delivery (51) Duncan, R., Kopeckova-Rejmanova, P., Strohalm, J., Hume, I., Cable, H. C., Pohl, J., Lloyd, J. B., and Kopecek, J. (1987) Anticancer agents coupled to N-(2-hydroxypropyl)methacrylamide copolymers. I. Evaluation of daunomycin and puromycin conjugates in vitro. Br. J. Cancer 55, 165-174. (52) Pratten, M., and Lloyd, J. (1986) Pinocytosis and phagocytosis: the effect of size of a particulate substrate on its mode

Bioconjugate Chem., Vol. 9, No. 2, 1998 191 of capture by rat peritoneal macrophages cultured in vitro. Biochim. Biophys. Acta 881, 307-313. (53) Racoosin, E. L., and Swanson, J. A. (1993) Macropinosome maturation and fusion with tubular lysosomes in macrophages. J. Cell Biol. 121, 1011-1020.

BC970146W