Epidermal Growth Factor Radiopharmaceuticals: 111In Chelation

The ABC-Elite immunocytochemistry kit was obtained from Vector Labs (Burlingame, CA). ... Buffer A (0.05 M NaH2PO4/0.5 M NaCl/pH 7.0) was added to a v...
0 downloads 0 Views 181KB Size
502

Bioconjugate Chem. 1999, 10, 502−511

Epidermal Growth Factor Radiopharmaceuticals: 111In Chelation, Conjugation to a Blood-Brain Barrier Delivery Vector via a Biotin-Polyethylene Linker, Pharacokinetics, and in Vivo Imaging of Experimental Brain Tumors Atsushi Kurihara, Yoshiharu Deguchi, and William M. Pardridge* Department of Medicine, UCLA School of Medicine, Los Angeles, California 90095-1682. Received October 9, 1998; Revised Manuscript Received December 29, 1998

Epidermal growth factor (EGF) is a potential peptide radiopharmaceutical for detection of brain tumors, because many human gliomas overexpress the EGF receptor (EGFR). The transport of EGF to the brain, however, is restricted by the blood-brain barrier (BBB). The purpose of the present study was to develop a vector-mediated brain delivery system for radiolabeled EGF. Human EGF was monobiotinylated with NHS-PEG3400-biotin, where NHS is N-hydroxysuccinimide and PEG3400 is poly(ethylene glycol) of 3400 Da molecular mass. EGF-PEG3400-biotin was radiolabeled with either 125I or 111In through the metal chelator, diethylenetriaminepentaacetic acid (DTPA). The radiolabeled EGF was then conjugated to a BBB delivery vector comprised of a complex of the OX26 monoclonal antibody (MAb) to the rat transferrin receptor, which was coupled to streptavidin (SA). Following intravenous injection in rats, the 125I conjugate was rapidly degraded in vivo, while the 111In conjugate was metabolically stable. The brain delivery of [111In]DTPA-EGF-PEG3400-biotin was enabled by conjugation with OX26/SA and was optimized by co-injection of unlabeled EGF to saturate EGF receptors in the liver. The specific binding of the [111In]DTPA-EGF-PEG3400-biotin conjugated to OX26/SA to the EGF receptor was confirmed in C6 rat glioma cells, which had been transfected with a gene encoding for the human EGF receptor under the regulation of a dexamethasone-inducible promoter. In vivo studies of C6-EGFR experimental tumors in Fischer 344 rats demonstrated successful brain imaging only when the peptide radiopharmaceutical was conjugated to the BBB delivery system, although the C6EGFR tumors did not express EGFR in vivo. In conclusion, these studies describe the molecular formulation of a peptide radiopharmaceutical that can be used for imaging brain tumors behind the BBB.

INTRODUCTION

Primary human brain tumors frequently overexpress the receptor for epidermal growth factor (EGF) (1-3), and EGF peptide radiopharmaceuticals are potential imaging agents for the diagnosis of human brain gliomas (4). However, like most other neuropeptides, EGF does not cross the brain capillary endothelial wall (5), which makes up the blood-brain barrier (BBB) in vivo. Since the BBB is intact (6), until the end stage of most human brain tumors, it may be necessary to conjugate EGF to BBB transport vectors in order to use this peptide as a radiopharmaceutical for imaging human brain tumors. BBB transport vectors include the OX26 monoclonal antibody (MAb), which targets the rat transferrin receptor (7), or the anti-human insulin receptor MAb (HIRMAb), which is a BBB delivery vector active in humans or Old World primates such as Rhesus monkeys (8). Peptide radiopharmaceuticals may be conjugated to BBB transport vectors using avidin-biotin technology (7). Previous studies describe the biotinylation of human EGF at surface -amino moieties of lysine residues using a 14 atom bis(aminohexanoyl) spacer arm (-XX-) between the EGF and the biotin (9). However, when the EGFXX-biotin was bound to a conjugate of the OX26 MAb * To whom correspondence should be addressed. Phone: (310) 825-8858. Fax: (310) 206-5163. E-mail: wpardrid@ med1.medsch.ucla.edu.

and streptavidin (SA), which is designated OX26/SA, the EGF no longer bound to the EGF receptor on C6 rat glioma cells (9). These C6 cells had been transfected with a gene encoding the human EGF receptor under the influence of a dexamethasone-inducible promoter and designated C6EGFRp cells (10). Conversely, when the -XX- linker was replaced with a >200 atom linker comprised of poly(ethylene glycol) of 3400 Da molecular mass, designated PEG3400, the binding of the peptide to the EGF receptor was restored despite conjugation to the OX26/SA BBB delivery system (9). The radionuclide used in previous studies was 125I, which was added to surface tyrosine residues on EGF. The present investigations extend these previous studies (9) and examine the pharmacokinetics and in vivo metabolic stability of the [125I]EGF-PEG3400-biotin conjugated to OX26/SA. However, the initial studies with the EGF peptide radiopharmaceutical formulated with the 125I radionuclide demonstrated a marked metabolic instability in vivo. Therefore, the present study also describes the preparation of a human EGF peptide radiopharmaceutical that is conjugated with both diethylenetriaminepentaacetic acid (DTPA), to enable chelation with the 111In radionuclide, and monobiotinylation using NHS-PEG3400-biotin, where NHS ) N-hydroxysuccinimide. The monobiotinylated form of EGF was purified by gel filtration fast-protein liquid chromatography (FPLC) using two columns in series and characterized using

10.1021/bc980123x CCC: $18.00 © 1999 American Chemical Society Published on Web 03/06/1999

Blood-Brain Barrier Transport of EGF

matrix-assisted laser desorption ionization mass spectrometry (MALDI) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting. The metabolic stability of the EGF conjugate formulated with the 111In radionuclide was demonstrated with FPLC analysis of rat serum taken 60 min after an intravenous (i.v.) injection. The affinity of the EGF peptide radiopharmaceutical conjugated to the BBB delivery system for the human EGF receptor was examined using the C6EGFR cells both in tissue culture and in vivo, wherein experimental brain tumors were formed by implanting the C6EGFR cells into the caudateputamen nucleus of Fischer 344 rats. Following initial pharmacokinetic analysis, the EGF peptide radiopharmaceutical conjugated to the BBB delivery system and formulated with the 111In radionuclide was used for neuroimaging of experimental brain tumors. EXPERIMENTAL PROCEDURES

Materials. [125I]Na was supplied by Amersham (Arlington Heights, IL). [111In]Cl3 was supplied by NEN Life Science Product Inc. (Boston, MA). NHS-PEG3400-biotin was obtained from Shearwater Polymers (Huntsville, AL), where NHS ) N-hydroxysuccinimide and PEG3400 ) poly(ethylene glycol) of 3400 Da molecular mass. Diethylenetriaminepentaacetic (DTPA) dianhydride was obtained from Aldrich Chemical Co., Inc. (Milwaukee, WI). Chloramine-T was purchased from MCB Reagents (Cincinnati, OH). Recombinant streptavidin (SA), recombinant human epidermal growth factor (EGF), and all other reagents were obtained from Sigma Chemical Company (St. Louis, MO). Superose 12 HR 10/30 FPLC columns and High-Trap copper affinity columns (1.0 mL) were obtained from Pharmacia Biotech (Piscataway, NJ). C6 rat glioma cells transfected with the gene for the human EGF receptor were kindly provided by Dr. Robert Fenstermaker of the Roswell Park Cancer Institute (Buffalo, NY), and these cells are designated C6EGFRp (10). The mouse hybridoma line producing the 528 MAb to human EGFR was obtained from the American Type Culture Collection (Rockville, MD). The anti-human EGFR mouse MAb was purchased from Upstate Biotechnology (Lake Placid, NY), catalog no. 05-101. The ABCElite immunocytochemistry kit was obtained from Vector Labs (Burlingame, CA). Mouse IgG1 and IgG2a isotype controls were obtained from Cappel Pharmaceuticals (Aurora, OH). Preparation of [125I]EGF-PEG3400-biotin. NHSPEG3400-biotin (65 nmol in 100 µL of 0.05 M NaHCO3) was added in an 8:1 molar ratio to 50 µg of EGF (8.0 nmol in 100 µL of 0.05 M NaHCO3, pH 8.3) followed by incubation at room temperature for 60 min. Of this mixture, 20 µL was added to 20 µL of 0.3 M NaH2PO4 (pH 7.0) followed by the addition of 2 mCi of 125I and 10 µL (0.8 µg) of chloramine-T. After a 60 s incubation, an additional 10 µL (0.8 µg) of chloramine-T was added, and 60 s later, the reaction was terminated by the addition of 50 µL (65 nmol) of sodium metabisulfite. Buffer A (0.05 M NaH2PO4/0.5 M NaCl/pH 7.0) was added to a volume of 1.0 mL, and the mixture was then applied to a 1.0 mL High-Trap copper affinity column to remove unreacted NHS-PEG3400-biotin, as described previously (11). The column was initially washed with 5.0 mL of water and charged with 0.5 mL of 0.1 M CuSO4 followed by washing with 5.0 mL of water, followed by application of the sample. The column was washed with 10 mL of buffer A, and the [125I]EGF-PEG3400-biotin was eluted with 5.0 mL of 50 mM imidazole in buffer A. One milliliter of this eluate (250 µCi) was added to a Centricon-3 microcon-

Bioconjugate Chem., Vol. 10, No. 3, 1999 503

centrator (Amicon, Inc., Beverly, MA), and the volume was reduced to 200 µL by centrifugation at 7500 rpm (7500g) for 40 min at room temperature. Then [125I]EGFPEG3400-biotin was purified by size-exclusion chromatography using a TSK-gel G2000SWXL column (TosoHaas, Montgomery, PA) with an elution buffer of 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4/0.05% Tween-20 at a flow rate of 0.5 mL/min. Preparation of [111In]DTPA-EGF-PEG3400-biotin. NHS-PEG3400-biotin (80 nmol in 50 µL of 0.05 M NaHCO3) was added in a 5:1 molar ratio to EGF (16 nmol in 200 µL of 0.05 M NaHCO3) followed by incubation at room temperature for 60 min. To this mixture, 50 µL of 0.43 M NaHCO3 and 850 nmol of DTPA dianhydride (17 µL of a dimethyl sulfoxide solution) were added, followed by incubation at room temperature for 60 min. The entire mixture was then applied to two Superose 12 HR 10/30 FPLC columns in series, followed by the elution in 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4 at a flow rate of 0.5 mL/ min for 90 min. EGF, containing a single DTPA group and a single PEG3400-biotin moiety and designated DTPAEGF-PEG3400-biotin, eluted at 52 min from the column, as demonstrated by MALDI (below). Two millicurie of 111In was added to the DTPA-EGF-PEG3400-biotin. After 30 min at room temperature, unreacted free 111In was removed by Sephadex G25 size-exclusion chromatography with an elution buffer of 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4 containing 0.05% Tween-20. Conjugation of [125I]EGF-PEG3400-biotin or [111In]DTPA-EGF-PEG3400-biotin to OX26/SA. The conjugate of the OX26 MAb and recombinant streptavidin (SA) was prepared with a thiol-ether linkage as described previously (12). Briefly, 17 mg of OX26 MAb was thiolated with a 10:1 molar ratio of 2-iminothiolane; 7 mg of recombinant SA was activated with a 20:1 molar ratio of m-maleimidobenzoyl N-hydroxysuccinimide ester (MBS). The activated SA and the thiolated OX26 were then mixed and incubated at room temperature for 3 h. The OX26/SA conjugate was labeled with 2.5 µCi of [3H]biotin and was purified on a 2.6 × 92 cm column of Sephacryl S300HR (Pharmacia) followed by elution in 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4/0.05% Tween-20 at 30 mL/h, and 3 mL fractions were collected. The conjugate peak was separated from either aggregates, unconjugated OX26 MAb, or unconjugated SA. The number of biotin binding sites per OX26/SA conjugate was 3.3 ( 0.3, as determined with the [3H]biotin binding assay (12). The conjugate of [125I]EGF-PEG3400-biotin and OX26/ SA was prepared by mixing 1 mL of the imidazole eluate from the High-Trap column containing [125I]EGFPEG3400-biotin (0.6 nmol) and 400 µg (0.2 nmol) of OX26/ SA (9). After reduction of the volume to 0.2 mL with a Centricon-30 microconcentrater, [125I]EGF-PEG3400-biotin conjugated to OX26/SA was separated from either [125I]EGF or unreacted [125I]EGF-PEG3400-biotin by size-exclusion chromatography using a TSK-gel G2000SWXL with an elution buffer of 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4/0.05% Tween-20 at a flow rate of 0.5 mL/min. The conjugate of [111In]DTPA-EGF-PEG3400-biotin and OX26/SA was prepared by mixing [111In]DTPA-EGFPEG3400-biotin (5 nmol) eluted from Sephadex G25 column with OX26/SA (8 nmol). Matrix-Assisted Laser Desorption Ionization Mass Spectrometry (MALDI). Mass spectra of DTPA-EGFPEG3400-biotin eluted from the two Superose 12HR 10/ 30 FPLC columns in series were recorded with a reflector time-of-flight instrument (PerSeptive Biosystems Voyager RP) used in the linear mode with stainless steel targets. An aliquot of sample solution (0.5 µL) or sample

504 Bioconjugate Chem., Vol. 10, No. 3, 1999

solution plus internal standard (myoglobin 5 pmol in 0.5 µL of water) was overlaid on top of an aliquot matrix (sinapinic acid), and after drying, the target was inserted into the mass spectrometer, and time/intensity data from 50 to 100 scans was averaged and converted to mass/ intensity data using external calibration from myoglobin. SDS-PAGE and Western Blotting. The change in molecular weight of EGF following conjugation with NHS-PEG3400-biotin was confirmed by SDS-PAGE. The reaction mixture of EGF and varying concentrations of NHS-PEG3400-biotin was applied to a 16.5% SDS-PAGE minigel (Bio-Rad, Rhichmond, CA), and electrophoresis was performed at constant voltage (100 V) for 2 h at room temperature. The gel slab was stained for protein content with Coomassie Blue G250. The incorporation of biotin at the tip of PEG strand was confirmed by Western blotting. Following SDS-PAGE, the gel slab was blotted to a nylon membrane without methanol (100 V, 60 min, 4 °C) and the membrane was subsequently stained for biotin content with the ABC kit (Vector Labs, Burlingame, CA) using avidin and biotinylated peroxidase and diaminobenzidine as chromagen (13). SDS-PAGE molecular weight standards were obtained from Bio-Rad (Richmond, CA). C6EGFRp Radioreceptor Assay. C6EGFRp cells were grown in Dulbecco’s modified Eagle medium (DMEM) with 4.5 g/L D-glucose, 4 mM glutamine, 10% fetal bovine serum, and 250 µg/mL G418, used as a selection agent, in 24-well cluster dishes. For the dexamethasone stimulation condition, the medium was changed to DMEM with 4.5 g/L D-glucose, 4 mM glutamine, 1% fetal bovine serum, and 1 µM dexamethasone 24 h prior to performing radioreceptor assays (10). Cells were washed with 0.01 M HEPES/0.15 M NaCl/pH 7.4/0.1% BSA (HBSB buffer), followed by 15 min to 120 min incubation at 37 °C with 200 µL of HBSB buffer containing 1.0 µCi/mL (0.2 nM) of [111In]DTPA-EGF-PEG3400-biotin with and without OX26/SA in the presence and absence of 1 µM unlabeled EGF. After incubation, supernatants were aspirated and the cells were washed two times with cold 0.01 M HEPES/0.15 M NaCl/pH 7.4, and solubilized by the addition of 0.5 mL of 1 N NaOH and incubation at 37 °C for 4 h. 111In radioactivity was counted by γ counter and protein content of the cells was measured with the bicinchoninic acid (BCA) protein assay (Pierce Chemical Co., Rockford, IL). Cell-associated binding (surface-binding and intracellular accumulation) was expressed as percent of medium radioactivity bound per milligram of cell protein. Pharmacokinetics and Brain Drug Delivery. Male Sprague-Dawley rats weighing 280-350 g were anesthetized with 100 mg/kg ketamine and 2 mg/kg xylazine intraperitoneally (ip). The femoral vein was cannulated with a PE50 cannula and injected with either 5 µCi of [125I]EGF-PEG3400-biotin or 5 uCi of [111In]DTPA-EGFPEG3400-biotin, with or without conjugation to OX26/SA and either 0 or 16 nmol of unlabeled human EGF. Blood samples (0.25 mL) were collected via a PE50 cannula implanted in the femoral artery at 0.25, 1, 2, 5, 15, 30 and 60 min after the injection of the isotope solution. The blood volume was replaced with an equal volume of saline. After 60 min, the animals were decapitated for removal of the brain, liver, kidney, heart, and lung. The radioactivity in the plasma and organ samples were counted using a γ counter (Beckman Instruments, Inc., Fullerton, CA). Pharmacokinetic parameters were calculated by fitting the plasma radioactivity data to a biexponential equation (14):

Kurihara et al.

A(t) ) A1e-k1t + A2e-k2t where A(t) ) % ID/mL plasma and ID ) injected dose. For [125I]EGF-PEG3400-biotin, TCA-precipitable radioactivity data was used. The biexponential equation was fit to plasma data by a derivative-free nonlinear regression analysis (PARBMDP, Biomedical Computer PSeries, developed at UCLA Health Science Computing Facilities). The area under the plasma concentrationtime curve (AUC), the steady-state volume of distribution (Vss), total plasma clearance (Cl), and the mean residence time (MRT) were calculated from A1, A2, k1, k2, and the body weight (kg) of the rat as described by Gibaldi and Perrier (15). The organ permeability-surface area (PS) product was calculated as

PS ) [Vd - V0]Cp(60min)/AUC(0-60min) where Cp(60min) is the terminal plasma concentration (dpm/µL) at 60 min after injection, Vd is the tissue volume of distribution determined from the ratio of disintegrations per minute per gram of the tissue to Cp(60min), and V0 is the organ plasma volume, which has been measured previously (14). The organ delivery of the EGF conjugate was determined as

% ID/g ) PS × AUC(0-60min) where % ID/g is the percent injected dose taken up per gram of organ. The metabolic stability in plasma of the [125I]EGFPEG3400-biotin was determined by trichloroacetic acid (TCA) precipitation of 50 µL aliquots of the plasma sample removed at each time point. The metabolic stability in plasma of the [111In]DTPA-EGFPEG3400-biotin bound to the OX26/SA conjugate was also examined by elution of plasma through a single Superose 12 HR 10/30 FPLC column. In this study, blood was removed at 60 min following intravenous injection of 60 µCi of [111In]DTPA-EGF-PEG3400-biotin conjugated to OX26/SA and 16 nmol of unlabeled EGF. Fifty microliters of serum sample was applied to the FPLC column, followed by the elution in 0.01 M NaH2PO4/0.15 M NaCl/ 0.05% Tween-20/pH 7.4 at a flow rate of 0.5 mL/min for 50 min. Column fractions (0.5 mL) were assayed for 111In radioactivity. Intracerebral Tumor Implantation and in Vivo Autoradiography. Male CD Fischer 344 rats weighing 170-180 g were implanted with C6EGFRp cells (4, 10). The rats were anaesthetized with a mixture of ketamine (100 mg/kg) and xylazine (2 mg/kg) ip and fixed in a stereotactic frame. A burr hole was drilled 3 mm to the right of midline, 1 mm posterior to bregma. C6EGFRp cells were suspended in serum-free DMEM containing 1.2% methylcellulose. Five microliters of cell suspension (1 × 105 cells) was injected into the right caudate nucleus at a depth of 4 mm over 10 s, using a 10 µL Hamilton syringe with fixed needle. At the end of the procedure, the viability of the remaining cells was determined by means of the trypan blue exclusion method. To allow for the implanted glioma cells to produce a tumor of a adequate size (∼5 mm), in vivo autoradiography studies were carried out 4 weeks following tumor cell implantation (4). Tumor-bearing rats were intravenously injected with either [111In]DTPA-EGF-PEG3400-biotin (100 µCi) with OX26/SA and 16 nmol of unlabeled EGF or [111In]DTPA-EGF-PEG3400-biotin (100 µCi) alone. Sixty minutes after isotope injection, rats were decapitated and the brain was rapidly removed from the cranium and sec-

Blood-Brain Barrier Transport of EGF

Bioconjugate Chem., Vol. 10, No. 3, 1999 505

Figure 1. (Right panel) Profile of the TCA-precipitable plasma radioactivity of either unconjugated [125I]EGF-PEG3400-biotin (open circle) or the [125I]EGF-PEG3400-biotin conjugated to the OX26/SA vector (closed circle) after an i.v. injection of 5 µCi/rat. (Left panel) Time course of the percent of plasma radioactivity that is precipitable with TCA after the i.v. injection. Each point represents the mean (SE of three rats.

tioned into 3 mm slabs. The slabs were plunged into powdered dry ice for rapid freezing over 30 min. Frozen sections of 15 µm were cut and thaw-mounted on glass cover slips. After drying, these cover slips were placed in apposition to Kodak Biomax MS X-ray film and exposed for 4 days at -70 °C with intensifying screen. The film was scanned with a Hewlett-Packard ScanJet IIcx/T flatbed scanner and transferred to Adobe Photoshop on a Power Macintosh 7100/66 microcomputer, followed by colorization with NIH image software, and prints were generated with a Kodak printer. Immunocytochemistry. The expression of the EGF receptor in the C6EGFRp glioma cells that had been transplanted into the caudate-putamen nucleus of Fischer 344 rats was investigated with immunochemistry at 4 weeks after the implantation of 1 × 105 C6EGFRp cells. Rats were sacrificed and the brain was rapidly removed and frozen in pulverized dry ice. The frozen brain was then embedded in tissue OCT (Tissue Tek II, Miles, Elkhart, IN) and 15 µ frozen sections were obtained with a Bright cryostat. Sections were mounted to glass slides and then fixed in either 100% acetone at -20 °C for 20 min or fixed by drying in an oven at 60 °C for 30 min. Endogenous peroxidase activity was inactivated with 1% H2O2 for 5 min at room temperature, and the slides were blocked with 3% horse serum for 30 min at room temperature. Two different primary antibodies were used. The 528 mouse MAb to the human epidermal growth factor receptor was generated from conditioned medium produced from 528 hybridoma cells. The 528 antibody has been demonstrated in previous studies to illuminate the human EGF receptor in experimental brain tumors with immunocytochemistry (16). A second primary antibody that was used in these studies was the anti-human EGF receptor mouse monoclonal antibody from Upstate Biotechnology (Lake Placcid, NY). The monoclonal antibodies were used at a concentration of 3 and 10 µg/mL and mouse IgG1 or mouse IgG2a isotype controls were employed in parallel. Following reaction with a biotinylated horse anti-mouse antibody that had been preabsorbed with rat immunoglobulin (Vector Laboratories, Burlingame, CA, catalog no. BA-2001), the slides were exposed to avidin and biotinylated peroxidase using the vector ABC method (Vector Labs, Burlingame, CA). These slides were lightly counterstained with hematoxylin and mounted. As a control, human glioblastoma multiforme U87 cells, which express the human EGF receptor (16), were obtained from the American Type

Figure 2. Tissue uptake of unconjugated [125I]EGFPEG3400-biotin (open column) or the [125I]EGF-PEG3400-biotin conjugated to the OX26/SA vector (closed column) 60 min after the i.v. injection of 5 µCi/rat. Data are mean (SE (n ) 3 rats).

Culture Collection and examined in parallel. These cells formed tumors in the brains of nude rats and the U87 cells, both in cell culture and in vivo as brain tumors, highly expressed immunoreactive human EGF receptor. RESULTS

Pharmacokinetics and Brain Uptake of [125I]EGF Conjugate. The [125I]EGF-PEG3400-biotin was prepared and injected intravenously into anesthetized rats with or without conjugation to OX26/SA and the plasma radioactivity profile is shown in Figure 1 (left panel). The EGF radiolabeled with 125I radionuclide was rapidly removed from the plasma compartment despite conjugation to OX26/SA and exhibited reduced metabolic stability based on measurement of plasma radioactivity that was precipitable with trichloroacetic acid (TCA), as demonstrated in Figure 1 (right panel). Analysis of distribution of the EGF conjugate into organs (brain, liver, and kidney) demonstrated the expected redirection of the EGF radiopharmaceutical from kidney to liver following conjugation to the anti-transferrin receptor MAb (Figure 2). However, the brain uptake of the EGF radiopharmaceutical formulated with a 125I radionuclide was low with or without conjugation to OX26/SA (Figure 2). As noted in previous studies (13), it is difficult to achieve effective brain uptake of drugs, despite conjugation to BBB delivery systems, when the peptide is inherently metabolically labile. These results indicated

506 Bioconjugate Chem., Vol. 10, No. 3, 1999

Figure 3. Delivery of radiolabeled EGF through the bloodbrain barrier and binding to EGF receptor (EGF-R) overexpressed in glioma cells. The delivery system is comprised of three domains: (i) a glioma binding domain consisting of EGF radiolabeled stably with 111In through the metal chelator, diethylenetriaminepentaacetic acid (DTPA); (ii) a linker domain consisting of single strand of PEG3400 spacer arm attached to EGF and the biotin moiety, which is in turn bound to streptavidin (SA); (iii) and a BBB transport domain consisting of the OX26 anti-rat transferrin receptor (TfR) monoclonal antibody (MAb), which is conjugated to SA through stable thioether linkage. The OX26 MAb undergoes receptor-mediated transcytosis through the blood-brain barrier via the brain capillary endothelial TfR (21). The EGF moiety binds to EGF-R positive glioma cells. The PEG3400 linker releases steric hindrance of the OX26 MAb on binding of the EGF to its cognate receptor (9). The radioligand 111In is suitable for imaging and confers metabolic stability of the radionuclide moiety compared to 125I.

the necessity of reformulating the EGF peptide radiopharmaceutical with a radionuclide that provided a greater degree of metabolic stability. Synthesis and Characterization of [111In]EGF Conjugate. The structure of the EGF peptide radiopharmaceutical conjugated to the BBB delivery system and formulated with 111In radionuclide though a DTPA chelator moiety is shown in Figure 3. The placement of a >200 atom linker comprised of PEG3400 reduces any steric hindrance between the two components of the conjugate (9). The extended linker allows for maintenance of bifunctionality of the conjugate with binding to both the EGF receptor, for imaging brain tumors, and to the transferrin receptor, to enable receptor-mediated transyctosis through the BBB in vivo. Following reaction of DTPA dianhydride with NHS-PEG3400-biotin, the human EGF was eluted through two Superose 12HR gel filtration FPLC columns in series and the elution profile is shown in Figure 4. The identity of the various peaks was determined by either MALDI (Experimental Procedures), or by parallel injection of standards. The EGF peak that contained a single DTPA residue and a single PEG3400-biotin residue was found in peak C, and the measured molecular mass, 10 193 Da, was the expected size, which is approximately 10 200 Da (inset, Figure 4). Following reaction with NHS-PEG3400-biotin, the EGF was also examined with SDS-PAGE, shown in Figure 5. The Coomassie blue stain of the SDS-PAGE gel (Figure 5A) shows the molecular weight of the unconjugated EGF at 6200 Da, and a second band is formed on the gel when NHS-PEG3400-biotin is added, and this band is seen in lanes 1-4 of Figure 5A and migrates at a molecular size of 10 000 Da, which approximates the expected molecular weight, 9900 Da, of EGF with a single PEG3400-biotin residue attached. At the higher molar ratios of NHS-PEG3400-biotin:EGF, an additional band is detected which migrates at 16 000 Da and approximates the expected molecular weight (14 000 Da) of EGF with 2 PEG3400-biotin moieties attached. The attachment of the

Kurihara et al.

Figure 4. Purification of monobiotinylated DTPA-EGF-PEG3400biotin by FPLC. Reaction mixture of EGF, NHS-PEG3400-biotin, and DTPA dianhydride was applied to two Superose 12 HR 10/ 30 FPLC column in series, followed by the elution in 0.01 M NaH2PO4/0.15 M NaCl/pH 7.4 at a flow rate of 0.5 mL/min for 90 min. Peak A, EGF-(PEG3400-biotin)3; peak B, DTPA-EGF(PEG3400-biotin)2; peak C, DTPA-EGF-PEG3400-biotin; peak D, DTPA-EGF; peak E, DTPA; peak F, NHS-PEG3400-biotin; peak G, solvent peak. (Insert) MALDI mass spectra shows mean molecular mass of 10 193 Da for peak C (DTPA-EGFPEG3400-biotin, theoretical MW, 10 240). Peak A, EGF-(PEG3400biotin)3 and peak B, DTPA-EGF-(PEG3400-biotin)2 showed molecular masses of 17 500 Da (theoretical 17 542 Da) and 13 801 Da (theoretical 13 891 Da), respectively, with the MALDI analysis. The (PEG3400-biotin)n nomenclature, where n ) 1, 2, or 3, refers to the number of PEG3400 strands conjugated to the EGF.

biotin residue was confirmed by Western blotting as shown in Figure 5B. No signal is detected in lanes 5 or 6, which represent reactions that contained either no NHS-PEG3400-biotin or no EGF, respectively. The reaction with a 1:1 molar ratio of NHS-PEG3400-biotin:EGF yields a single prominent band of 9200 Da, and this corresponds to the approximate molecular mass of EGF with a single PEG3400-biotin residue attached. At the higher molar ratios (lanes 1-3) of NHS-PEG3400-biotin:EGF, a second higher molecular weight band is detected of 14 600 Da (Figure 5B), and this approximates the molecular mass of EGF with 2 PEG3400-biotin moieties attached. Radioreceptor Assays with [111In]EGF Conjugate and C6 Glioma Cells in Tissue Culture. The EGFPEG3400-biotin formulated with 111In and DTPA radionuclide still actively bound to the EGF receptor on the C6EGFRp cells with or without conjugation to OX26/SA, as shown in Figure 6. Radioreceptor assays performed with the C6EGFRp cells pretreated with 1 µM dexamethasone for 24 h are shown in the right-hand panel of Figure 6, and radioreceptor assays performed with these cells without dexamethasone pretreatment are shown in the left-hand panel of Figure 6. In the absence of dexamethasone pretreatment, there is little induction of the EGF receptor in these cells, which explains the minimal binding of unconjugated [111In]EGF (Figure 6, left-hand panel). However, there was significant binding, even in the C6EGFRp cells that were not pretreated with

Blood-Brain Barrier Transport of EGF

Bioconjugate Chem., Vol. 10, No. 3, 1999 507

Figure 5. SDS-polyacrylamide gel electrophoresis (A) and Western blotting (B) of EGF-PEG3400-biotin. Reaction mixture of EGF and varying concentration of NHS-PEG3400-biotin was applied to a 16.5% SDS-PAGE minigel. Gel slab was either stained with Coomassie Blue G250 for protein content (panel A), or blotted to a nylon membrane followed by staining with avidin and biotinylated peroxidase to visualize the attachment of PEG3400-biotin to the EGF molecule (panel B). Molar ratio of NHS-PEG3400-biotin: EGF in the reaction mixture was 5:1 (lane 1), 4:1 (lane 2), 2:1 (lane 3), 1:1 (lane 4), 0:1 (lane 5), and 5:0 (lane 6).

Figure 6. Radioreceptor assay showing time course of binding of either [111In]DTPA-EGF-PEG3400-biotin conjugated to OX26/SA (circles) or binding of [111In]DTPA-EGF-PEG3400-biotin (triangles) to C6 rat glioma cells transfected with the human EGF receptor gene. Assays were performed in the presence (right panel) or absence (left panel) of prior dexamethasone (1 µM) stimulation and in the presence (open symbols) or absence (closed symbols) of 1 µM unlabeled EGF. Each point represents the mean (SE of three wells.

dexamethasone, of the [111In]EGF-PEG3400-biotin following conjugation to OX26/SA, owing to the expression of transferrin receptor on these cells. The conjugate bound to both transferrin receptor and EGF receptor on the cells, when the EGF receptor was induced by dexamethasone pretreatment (Figure 6, right-hand panel). A saturation analysis demonstrated that the ED50 of EGF competition of the binding of [111In]EGF-PEG3400-biotin conjugated to OX26/SA was approximately 1 nM. Pharmacokinetics and Brain Uptake of [111In]EGF Conjugate. The pharmacokinetic analysis showed that unconjugated [111In]EGF- PEG3400-biotin was rapidly removed from the bloodstream (Figure 7), similar to the rapid removal of [125I]EGF- PEG3400-biotin from the circulation (Figure 1). For example, the % ID/mL of plasma was approximately 0.1% ID/mL at 10 min for either form of the EGF, indicating that >90% of the peptide had been removed from the circulation within the first 10 min after injection. Following conjugation of [111In]EGF- PEG3400-biotin to OX26/SA, there was some improvement in the plasma AUC (Table 1), but the EGF-

PEG3400-biotin conjugated to OX26/SA and formulated with DTPA/111In was still relatively rapidly removed from the bloodstream (Figure 7). Since human EGF is avidly removed from the circulation by the hepatic rat EGF receptor (17, 18), we co-administered a loading dose of unlabeled human recombinant EGF in parallel with the intravenous injection of the EGF-PEG3400-biotin conjugated to OX26/SA and radiolabeled with 111In (Figure 7). The co-injection of the unlabeled EGF resulted in a >4fold increase in the plasma AUC (Table 1), and the EGF loading also optimized the brain uptake of the conjugate, as shown in Figure 8. In the presence of the unlabeled EGF, the brain uptake of the [111In]EGF-PEG3400-biotin conjugated to OX26/SA was increased more than 10-fold compared to the brain uptake of [111In]EGFPEG3400-biotin without attachment to the BBB delivery system (Figure 8). The metabolic stability of the EGF-PEG3400-biotin conjugated to OX26/SA and formulated with DTPA/111In was confirmed by gel filtration FPLC analysis of rat serum taken 60 min after an intravenous injection of the

508 Bioconjugate Chem., Vol. 10, No. 3, 1999

Kurihara et al.

Figure 8. Brain uptake of unconjugated [111In]DTPA-EGFPEG3400-biotin, [111In]DTPA-EGF-PEG3400-biotin conjugated to the OX26/SA vector, or [111In]DTPA-EGF-PEG3400-biotin conjugated to the OX26/SA vector and co-injected with 16 nmol of unlabeled EGF 60 min after the i.v. injection of 5 µCi/rat. Data are mean (SE (n ) 3 rats). Figure 7. Profile of plasma radioactivity of either unconjugated [111In]DTPA-EGF-PEG3400-biotin (closed triangle), the [111In]DTPA-EGF-PEG3400-biotin conjugated to the OX26/SA vector (open circle) or [111In]DTPA-EGF-PEG3400-biotin conjugated to the OX26/SA vector and co-injected with 16 nmol of unlabeled EGF (closed circle) after an i.v. injection of 5 µCi/rat. Each point represents the mean (SE of three rats. Table 1. Pharmacokinetic Parametersa

parameter

[111In]EGF

[111In]EGF + OX26/SA

[111In]EGF + OX26/SA + unlabeled EGF

A1 (% ID/mL) 3.3 ( 0.7 2.6 ( 0.2 3.5 ( 0.3 A2 (% ID/mL) 0.11 ( 0.02 0.31 ( 0.03 1.7 ( 0.1 k1 (min-1) 1.1 ( 0.2 0.60 ( 0.11 0.43 ( 0.06 k2 (min-1) 0.044 ( 0.016 0.021 ( 0.001 0.024 ( 0.002 1 t1/2 (min) 0.71 ( 0.14 1.2 ( 0.2 1.7 ( 0.2 2 t1/2 (min) 33 ( 8 33 ( 2 29 ( 2 AUC(0-60) (% 6.6 ( 0.6 15 ( 1 63 ( 4 ID min/mL) AUC(0-∞) (% 8.0 ( 0.4 19 ( 1 79 ( 6 ID min/mL) Vss (mL/kg) 1297 ( 376 636 ( 7 155 ( 7 Cl (mL/min/kg) 44 ( 3 18 ( 2 4.2 ( 0.3 MRT (min) 30 ( 9 38 ( 5 37 ( 3 a Computed from the plasma data in Figure 7. In all cases, the [111In]EGF represents [111In]EGF-PEG3400-biotin.

conjugate (Figure 9). No free EGF was detected, and no significant transfer of the 111In radionuclide to circulating transferrin was detected (peak 4 and 3, respectively, Figure 9). Although there were some transfer of radioactivity to high molecular weight fraction in serum that comigrated with lipoproteins, the principal peak remaining in serum 60 min after intravenous injection was the intact conjugate (peak 2, Figure 9). In Vivo C6 Glioma Tumor Imaging with [111In]EGF Conjugate. The quantitative autoradiography (QAR) and tumor imaging using an EGF peptide radiopharmaceutical that is radiolabeled with 111In and administered either with or without the OX26/SA BBB delivery system is shown in Figure 10. The brain images obtained following the intravenous injection of [111In]EGF-PEG3400-biotin without attachment to the OX26/SA BBB delivery system are shown in the bottom panels (E-H) of Figure 10 and indicate that there is no measurable BBB transport of the peptide radiopharmaceutical into either normal brain or the brain tumor. When the [111In]EGF-PEG3400-biotin is conjugated to OX26/SA and injected intravenously into C6EGFR tumor-bearing rats,

Figure 9. Elution from single Superose 12 HR 10/30 FPLC column of rat serum obtained 60 min after an i.v. injection of 60 µCi of [111In]DTPA-EGF-PEG3400-biotin conjugated to the OX26/SA vector with 16 nmol of unlabeled EGF. Fifty microliters of pooled serum was applied to the FPLC column, followed by the elution in 0.01 M NaH2PO4/0.15 M NaCl/0.05% Tween20/pH 7.4 at a flow rate of 0.5 mL/min for 50 min. Column fractions (0.5 mL) were assayed for [111In]radioactivity. Arrows 1, 2, 3, and 4 indicate elution time of lipoproteins, [111In]DTPAEGF-PEG3400-biotin conjugated to the OX26/SA, [111In]transferrin, and [111In]DTPA-EGF-PEG3400-biotin, respectively. The result shows that [111In]DTPA-EGF-PEG3400-biotin conjugated to OX26/SA is stable in the circulation at 60 min after i.v. injection in rats.

there is a marked increase in the generalized brain uptake of the peptide radiopharmaceutical owing to transport through the BBB as shown in the top panels (A-D) of Figure 10. The regions of the tumor are clearly demarcated owing to the reduced radionuclide signal emanating from the tumor region. This reduced signal suggested that the C6EGFRp cells, which produce the EGF receptor under the influence of a dexamethasone inducible promoter (10), do not express the human EGF receptor in vivo. This was confirmed by immunocytochemistry using two different monoclonal antibodies to the human EGF receptor. Immunocytochemistry of frozen sections of the C6EGFRp tumor showed no visible immunoreactive EGF receptor in the tumor cells in vivo. The positive control cells, U87 human glioma cells, demonstrated high expression of immunoreactive human EGF receptor using either of the monoclonal antibodies (Experimental Procedures).

Blood-Brain Barrier Transport of EGF

Bioconjugate Chem., Vol. 10, No. 3, 1999 509

Figure 10. Film autoradiography of brain sections obtained from C6EGFRp tumor-bearing rats injected i.v. with 100 µCi of either [111In]DTPA-EGF-PEG3400-biotin with OX26/SA and 16 nmol of unlabeled EGF (panels A-D) or [111In]DTPA-EGF-PEG3400-biotin alone (panels E-H). The C6EGFRp rat glioma cells (1 × 105 cells/rat) were implanted into the brains of CD Fischer 344 rats and grown for 4 weeks prior to brain distribution experiments. Brains were removed 60 min after the isotope injection. Biomax MS X-ray films were exposed for 4 days at -70 °C with intensifying screen. DISCUSSION

The results of the present investigation are consistent with the following conclusions. First, formulation of the EGF conjugate with the 125I radionuclide yields a radiopharmaceutical that is metabolically unstable (Figure 1). Second, biotinylation of EGF with NHS-PEG3400-biotin yields a mixture of mono- and dibiotinylated peptide, even at a 1:1 molar ratio of EGF:NHS-PEG3400-biotin, as demonstrated by either SDS-PAGE (Figure 5) or gel filtration FPLC (Figure 4); the EGF construct with a single DTPA residue and a single PEG3400-biotin moiety, such as that depicted in Figure 3, can be purified using two Superose 12HR gel filtration FPLC columns in series (Figure 4). Third, the [111In]EGF-PEG3400-biotin conjugated to OX26/SA binds to both transferrin receptor and EGF receptor in the C6EGFRp cells under the influence of 1 µM dexamethasone, but no specific EGF receptor activity is detected in the cells that are not exposed to dexamethasone (Figure 6). Fourth, the EGFPEG3400-biotin conjugated to OX26/SA and formulated with the 111In radionuclide shows a high degree of metabolic stability, as low molecular weight 111In-labeled metabolites are not detected in plasma (Figure 9). In contrast, there is a rapid appearance of TCA-soluble low molecular weight 125I metabolites formed in plasma when the 125I radionuclide is used to label the EGF (Figure 1, right panel). Fifth, the [111In]DTPA- EGF-PEG3400-biotin conjugated to OX26/SA is rapidly removed from plasma (Figure 7), and this rapid uptake by peripheral tissues is greatly reduced by co-injection of 16 nmol/rat (100 µg/ rat) of unlabeled EGF (Figure 7, Table 1). Sixth, imaging of brain with radiolabeled forms of EGF is not possible when the peptide is not coupled to a BBB drug delivery system (Figure 10, panels E-H), but brain structures and tumors can be imaged with radiolabeled EGF that is conjugated to a BBB drug delivery system using a PEG3400 linker (Figure 10, panels A-D). The plasma AUC and whole body Vss values are comparable for EGF conjugates when either the 125I or

the 111In radionuclide is used. For example, the plasma AUC|∞0 for [125I]EGF-PEG3400-biotin conjugated to OX26/SA or [111In]DTPA-EGF-PEG3400-biotin conjugated to OX26/SA are 23 ( 2% ID min/mL (data not shown) and 19 ( 1% ID min/mL (Table 1), respectively. When the EGF is conjugated to OX26/SA and co-injected with unlabeled EGF, the plasma AUC|∞0 for the 125I and the 111In formulations are 61 ( 8% ID min/mL (data not shown) and 79 ( 6% ID min/mL (Table 1), respectively. However, after uptake by peripheral tissues (e.g., liver and kidney), the 125I radionuclide is rapidly exported to the plasma in the form of low molecular weight TCA soluble metabolites (Figure 1, right panel), but there is no export to plasma of low molecular weight 111In metabolites (Figure 9). These low molecular weight metabolites may include either radiolabeled iodide or amino acids such as iodotyrosine, which are capable of transport through the BBB via carrier-mediated transport and give artifactually high basal uptakes of brain radioactivity (11, 19). Such artifacts are seen in the present study; the basal brain uptake of [125I]EGFPEG3400-biotin, 0.026 ( 0.001% ID/g (Figure 2), is 10-fold greater than the basal brain uptake of [111In]DTPA-EGFPEG3400-biotin (Figure 8). The brain uptake of these low molecular weight 125I metabolites would make brain imaging very difficult as it would reduce the signal/noise ratio of the image. Conversely, the brain uptake of radioactivity is negligible when the EGF is administered in the form of the 111In radionuclide, and EGF is not conjugated to a BBB drug delivery system such as OX26/ SA (Figure 10, panels E-H). These considerations are further evidence that the measurement of brain uptake of 125I-labeled peptides following peripheral administration is driven principally by the artifactual uptake of low molecular weight metabolites rapidly generated by the degradation of the peptide in noncerebral tissues (11). The synthetic scheme for [125I]EGF-PEG3400-biotin yields a mixture of mono- and dibiotinylated peptide

510 Bioconjugate Chem., Vol. 10, No. 3, 1999

Kurihara et al.

Table 2. PS Values (µL/min/g)a

organ

[111In]EGF

[111In]EGF + OX26/SA

heart lung liver kidney

6.3 ( 0.4 8.9 ( 2.8 635 ( 33 1006 ( 32

2.9 ( 0.7 11.1 ( 1.9 390 ( 12 145 ( 13

[111In]EGF + OX26/SA + unlabeled EGF 0.17 ( 0.11 4.0 ( 0.5 74 ( 4 76 ( 6

a Determined at 60 min after intravenous injection. Mean ( S. E. (n ) 3 rats). Computed from the plasma data in Figure 7. In all cases, the [111In]EGF represents [111In]EGF-PEG3400 biotin.

(Experimental Procedures). Owing to the multivalency of streptavidin binding of biotin, there is a potential for aggregates to form when the ligand is multibiotinylated and bound to OX26/SA. However, in the present studies of the 125I radionuclide formulation, significant aggregates are minimized, as the molar ratio of the [125I]EGF(PEG3400-biotin)2 and the OX26/SA was approximately 1:1 (Experimental Procedures). Moreover, when the monobiotinylated form of EGF-PEG3400-biotin was purified by two Superose 12HR gel filtration FPLC columns in series (Figure 4) and then radiolabeled with 125I and chloramine T (9), there were no significant differences in the plasma radioactivity profiles (unpublished observations). The conjugation of EGF-PEG3400-biotin to OX26/SA resulted in a redirection of the EGF peptide from kidney to liver (Figure 2, Table 2). The molecular mass of EGFPEG3400-biotin is approximately 10 000 Da, and the molecular mass of OX26/SA is 200 000 Da. Therefore, following conjugation to OX26/SA, the effective molecular mass of the EGF increases from 10 000 Da to 210 000 Da, and this explains the marked reduction in renal clearance of the EGF conjugate. However, the EGF is still rapidly removed from plasma despite conjugation to OX26/SA (Figure 7). Previous studies have shown that human recombinant EGF is rapidly taken up by the rat EGF receptor in liver in vivo (17, 18). Therefore, we considered that co-injection of unlabeled EGF (100µg/rat) would saturate hepatic uptake of the EGF construct, prolong the circulation time of the imaging agent, and have a beneficial pharmacokinetic effect. This was observed as the plasma AUC|∞0 was increased 4-fold from 19 ( 1 to 79 ( 6% ID min/mL following EGF loading (Table 1). The systemic volume of distribution (Vss) was also decreased 4-fold from 637 ( 7 to 157 ( 7 mL/kg following EGF loading (Table 1). The plasma AUC|∞0 of [111In]DTPA-EGF-PEG3400-biotin, 8.0 ( 0.4% ID min/mL, is comparable to the plasma AUC for [125I]EGF, 6.8% ID min/mL, which is calculated from the data of Kim et al. (18). The principal effect of EGF loading is a 5-fold reduction in the organ clearance by liver, although EGF clearance by heart, lung, and kidney is also decreased by EGF loading (Table 2). The pharmacokinetic data for [111In]DTPA-EGFPEG3400-biotin in Tables 1 and 2 allow for comparison with similar data reported previously for [125I]EGF (17, 18). The fact that the plasma clearance for the two formulations of EGF are comparable means that placement of a single PEG moiety on a peptide such as EGF does not significantly reduce the rate of plasma clearance of the peptide. Pegylation of proteins is well-known to delay plasma clearance, but this requires the placement of multiple PEG moieties on the surface of the protein (11). The EGF peptide radiopharmaceutical depicted in Figure 3 and used for the brain imaging studies shown in Figure 10 embodies three aspects of the formulation that allow for optimization of the brain imaging signal.

First, a >200 atom PEG3400 linker is placed between the EGF and the biotin to release any steric hindrance caused by binding to streptavidin (9). This long flexible arm allows for free binding of the construct to either the EGF receptor or the BBB transferrin receptor. Second, the radionuclide moiety is 111In rather than the 125I used in previous studies (9). The use of the 111In radionuclide confers a greater degree of metabolic stability (Figure 9) and a very low basal uptake of the unconjugated peptide (Figure 10). Third, the co-injection of unlabeled EGF inhibits hepatic uptake of the conjugate approximately 5-fold (Table 2), and this optimizes the plasma pharmacokinetics (Table 1, Figure 7) and results in a proportionate increase in brain uptake (Figure 8). The brain imaging studies in Figure 10 allow for the following conclusions. First, [111In]DTPA-EGFPEG3400-biotin does not cross either the BBB in normal brain or the blood-tumor barrier (BTB) in the brain tumor (Figure 10, panels E-H). The uptake of the [111In]EGFPEG3400-biotin in brain is 0.0033 ( 0.0007% ID/g (Figure 8), and this approximates the brain uptake of sucrose (7). The failure of EGF to cross the BBB in normal brain is expected based on previous studies (5). Although the BTB is more permeable than the normal BBB to small molecules (20), the BTB in the C6 glioma is clearly not permeable to the unconjugated EGF peptide radiopharmaceutical, which has a molecular mass of approximately 10 000 Da, as shown in the bottom panels (E-H) of Figure 10. Second, the [111In]DTPA-EGF-PEG3400-biotin does cross the BBB and enter brain following conjugation to OX26/SA. The in vivo QAR studies (Figure 10, panels A- D) show increased brain uptake of the EGF peptide radiopharmaceutical that is attached to the OX26 MAb BBB delivery system. The transcytosis of the OX26 MAb through the BBB has been demonstrated previously at the electron microscopic level with conjugates of OX26 and 5 nm gold (21) and at the light microscopic level using autoradiography (22). Third, the EGF peptide pharmaceutical conjugated to OX26/SA also crosses the BTB, although the signal in the tumor brain is reduced compared to normal brain (Figure 10, panels A-D). This reduced signal suggested that the human EGF receptor was not expressed by the C6EGFRp cells following implantation in brain and growth as a brain tumor. The absence of EGF receptor in these cells in vivo was confirmed by immunocytochemistry (Results) and is consistent with the tissue culture experiments (Figure 6) showing minimal binding to the EGF receptor in the C6EGFRp cells that were not exposed to 1 µM dexamethasone for 24 h. The expression of the human EGF receptor in the C6EGFRp cells is driven by the dexamethasone-inducible mouse mammary tumor virus long terminal repeat (MMTV-LTR) promoter (10). We administered 1 mg/rat of dexamethasone 24 h prior to sacrifice but still could not detect any immunoreactive EGF receptor in these brain tumors in vivo (unpublished observations). Another reason the signal in the brain tumor region is decreased compared to normal brain most likely arises from the decreased vascular density in the brain tumor compared to normal brain. In previous studies with peptide radiopharmaceuticals that were enabled to undergo transport through the BBB by attachment to BBB drug delivery systems, there was a marked increase in the image in gray matter versus white matter in the primate brain in vivo (8). This arose from the 3-fold greater vascular density in gray matter versus white matter and, accordingly, the 3-fold greater vascular density of BBB receptor being targeted by the delivery system. The vascular density in the C6 glioma

Blood-Brain Barrier Transport of EGF

is probably comparable to the vascular density in white matter, since the rate of cerebral blood flow (CBF) in the C6 glioma approximates white matter CBF, which is only 44% of the CBF in gray matter (23). The present studies show that the C6EGFRp cells express undetectable levels of human EGF receptor when the cells form experimental brain tumors in vivo based on both the in vivo imaging (Figure 10) and the EGFR immunocytochemistry (Results). Experimental tumors such as the U87 human glioblastoma multiforme, which grows as experimental brain tumors in immuno-compromised rodents (16), may prove to be a more suitable model for imaging the human EGF receptor in experimental brain tumors in vivo. Factors that will affect the tumor imaging using EGF peptide radiopharmaceuticals that are enabled to cross the BTB include (i) the vascular density and local concentration of transferrin receptor at the BTB and (ii) the local expression of the EGF receptor in the tumor cells and the rate of endocytosis of the EGF into the tumor cells. These receptor systems may be investigated in future imaging of brain tumors using peptide radiopharmaceuticals such as that depicted in Figure 3. The present studies demonstrate that peptide radiopharmaceuticals such as EGF may undergo special molecular formulation to enable both metabolic stability of the radionuclide and transport of the peptide radiopharmaceutical through the BBB in vivo. ACKNOWLEDGMENT

Margarita Tayag provided expert technical assistance and the manuscript was skillfully prepared by David Kim and Daniel Jeong. This work was supported by a grant from the U.S. Department of Energy. Dr. Kym Faull performed the MALDI analyses. Dr. Dafang Wu asisted with the brain tumor implants. LITERATURE CITED (1) Libermann, T. A., Nusbaum, H. R., Razon, N., Kris, R., Lax, I., Soreq, H., Whittle, N., Waterfield, M. D., Ullrich, A., and Schlessinger, J. (1984) Amplification, enhanced expression and possible rearrangement of EGR receptor gene in primary human brain tumours of glial origin. Nature 313, 144-147. (2) Wong, A. J., Bigner, S. H., Bigner, D. D., Kinzler, K. W., Hamilton, S. R., and Vogelstein, B. (1987) Increased expression of the epidermal growth factor receptor gene in malignant gliomas is invariably associated with gene amplification. Proc. Natl. Acad. Sci. U.S.A. 84, 6899-6903. (3) Torp, S. H., Helseth, E., Dalen, A., and Unsgaard, G. (1991) Epidermal growth factor receptor expression in human gliomas. Cancer Immunol. Immunother. 33, 61-64. (4) Capala, J., Barth, R. F., Bailey, M. Q., Fenstermaker, R. A., Marek, M. J., and Rhodes, B. A. (1997) Radiolabeling of epidermal growth factor with 99m-Tc and in vivo localization following intracerebral injection into normal and gliomabearing rats. Bioconjugate Chem. 8, 289-295. (5) Nave, K. A., Probstmeier, R., and Schachner, M. (1985) Epidermal growth factor does not cross the blood-brain barrier. Cell Tissue Res. 241, 453-457. (6) Blasberg, R. G., and Groothuis, D. R. (1986) Chemotherapy of brain tumors: physiological and pharmacokinetic considerations. Sem. Oncol. 13, 70-82. (7) Pardridge, W. M. (1997) Drug delivery to the brain. J. Cereb. Blood Flow Metab. 17, 713-731. (8) Wu, D., Yang, J., and Pardridge, W. M. (1997) Drug targeting of a peptide radiopharmaceutical through the primate blood-brain barrier in vivo with a monoclonal anti-

Bioconjugate Chem., Vol. 10, No. 3, 1999 511 body to the human insulin receptor. J. Clin. Invest. 100, 1804-1812. (9) Deguchi, Y., Kurihara, A., and Pardridge, W. M. (1999) Retention of biologic activity of human epidermal growth factor following conjugation to a blood-brain barrier drug delivery vector via extended poly(ethylene glycol) linker. Bioconjugate Chem. 10, 32-37. (10) Fenstermaker, R. A., Capala, J., Barth, R. F., Hujer, A., Kung, H.-J., and Kaetzel, D. M. (1995) The effect of epidermal growth factor receptor (EGFR) expression on in vivo growth of rat C6 glioma cells. Leukemia 9, S106-S112. (11) Sakane, T., and Pardridge, W. M. (1997) Carboxy-directed pegylation of brain-derived neurotropic factor markedly reduces systemic clearance with minimal loss of biologic activity. Pharm. Res. 14, 1085-1091. (12) Wu, D., and Pardridge, W. M. (1996) Central nervous system pharmacologic effect in conscious rats after intravenous injection of a biotinylated vasoactive intestinal peptide analogue coupled to a blood-brain barrier drug delivery system. J. Pharmacol. Exp. Ther. 279, 77-83. (13) Pardridge, W. M., Wu, D., and Sakane, T. (1998) Combined use of carboxyl-directed protein pegylation and vector-mediated blood-brain barrier drug delivery system optimizes brain uptake of brain-derived neurotrophic factor following intravenous administration. Pharm. Res. 15, 576-582. (14) Kang, Y.-S., and Pardridge, W. M. (1994) Use of neutral avidin improves pharmacokinetics and brain delivery of biotin bound to an avidin-monoclonal antibody conjugate. J. Pharmacol. Exp. Ther. 269, 344-350. (15) Gibaldi, M., and Perrier, D. (1982) Pharmacokinetics, Marcel Dekker, Inc., New York. (16) Huang, H.-J., Nagane, M., Klingbeil, C. K., Lin, H., Nishikawa, R., Ji, X.-D., Huang, C.-M., Gill, G. N., Wiley, H. S., and Cavenee, W. K. (1997) The enhanced tumorigenic activity of a mutant epidermal growth factor receptor common in human cancers is mediated by threshold levels of constitutive tyrosine phosphorylation and unattenuated signaling. J. Biol. Chem. 272, 2927-2935. (17) Kim, D. C., Sugiyama, Y., Fuwa, T., Sakamoto, S., Iga, T., and Hanano, M. (1989) Kinetic analysis of the elimination process of human epidermal growth factor (hEGF) in rats. Biochem. Pharmacol. 38, 241-249. (18) Kim, D. C., Hanano, M., Kanai, Y., Ohnuma, N., and Sugiyama, Y. (1992) Localization of binding sites for epidermal growth factor (EGF) in rat kidney: evidence for the existence of low affinity EGF binding sites on the brush border membrane. Pharm. Res. 9, 1394-1401. (19) Pardridge, W. M. (1981) Enkephalin and blood-brain barrier: studies of binding and degradation in isolated brain microvessels. Endocrinology 109, 1138-1143. (20) Sugita, M., and Black, K. L. (1998) Cyclic GMP-specific Phosphodiesterase Inhibition and Intracarotid Bradykinin Infusion Enhances Permeability into Brain Tumors. Cancer Res. 58, 914-920. (21) Bickel, U., Kang, Y.-S., Yoshikawa, T., and Pardridge, W. M. (1994) In vivo demonstration of subcellular localization of anti-transferrin receptor monoclonal antibody-colloidal gold conjugate within brain capillary endothelium. J. Histochem. Cytochem. 42, 1493-1497. (22) Skarlatos, S., Yoshikawa, T., and Pardridge, W. M. (1995) Transport of [125I]-transferrin through the rat blood-brain barrier in vivo. Brain Res. 683, 164-171. (23) Hiesiger, E. M., Voorhies, R. M., Basler, G. A., Lipschutz, L. E., Posner, J. B., and Shapiro, W. R. (1986) Opening the blood-brain and blood-tumor barriers in experimental rat brain tumors: the effect of intracarotid hyperosmolar mannitol on capillary permeability and blood flow. Ann. Neurol. 19, 50-59.

BC980123X