Bioconjugate Chem. 2009, 20, 1171–1178
1171
In Vitro and in Vivo Metabolism of Lu-AMBA, a GRP-Receptor Binding Compound, and the Synthesis and Characterization of Its Metabolites Karen E. Linder,* Edmund Metcalfe, Thangavel Arunachalam, Jianqing Chen, Stephen M. Eaton, Weiwei Feng, Hong Fan, Natarajan Raju, Aldo Cagnolini, Laura E. Lantry, Adrian D. Nunn, and Rolf E. Swenson Bracco Research USA Inc., 305 College Road East, Princeton, New Jersey 08540. Received January 26, 2009; Revised Manuscript Received April 27, 2009
The metabolism of 177Lu-AMBA (AMBA ) DO3A-CH2CO-G-(4-aminobenzoyl)-QWAVGHLM-NH2), a radiotherapeutic compound in clinical development that binds to GRP and NMB receptors, was studied in vitro (mouse, rat and human plasma, mouse kidney homogenate) and in vivo (by analysis of mouse and rat plasma and urine following IV injection of 177Lu-AMBA). The primary metabolites were Lu-DO3A-CH2CO-G-Abz4-R, where R ) -Q-OH (A), -QW-OH (B), and -QWAVGH-OH (C). Minor amounts of (D) where R ) -QWAVGHLM-OH and (E) -QWAVGHL-OH were also observed. Clearance of 177Lu-AMBA and of radioactivity from mouse and rat blood was rapid in vivo. In mouse and rat urine, only metabolites Lu-A and Lu-B were foundsno parent drug was excreted. Unmetalated ligands and natLu and 177Lu complexes for Lu-AMBA metabolites A-E were synthesized, characterized by HPLC and MS, and used to perform in vitro competition and direct binding studies on GRP receptor-positive PC-3 (human prostate) cancer cells. Biodistribution studies with 177Lu-labeled metabolites A-E were performed in PC-3 tumor-bearing mice and the results compared with intact 177Lu-AMBA. IC50 values for unmetalated metabolite ligands A-E were >400 nM in PC-3 cells in competition binding studies against 177LuAMBA. No direct binding to PC-3 cells was observed with 177Lu-labeled A-C, confirming IC50 results. 177Lulabeled metabolites A-E showed no uptake in GRP-receptor positive tumor or pancreas in PC-3 tumor bearing mice. All metabolites were rapidly excreted via the renal route (∼78-87%) within 1 h. These results demonstrate that the tumor uptake observed with 177Lu-AMBA is due to parent drug and not due to any of its identified metabolites.
INTRODUCTION Bombesin, a 14-amino acid peptide, is an analogue of human gastrin-releasing peptide (GRP) and binds GRP receptors with high affinity and specificity (1). 177Lu-AMBA (Figure 1) is a bombesin (7-14) derivative that binds to and is internalized by cells expressing gastrin-releasing peptide (GRP) and Neuromedin B (NMB) receptors (2, 3). GRP receptors are overexpressed in many types of cancers, notably breast, prostate, pancreatic, and lung (4-7). They are an attractive target for molecular imaging and targeted radiotherapy of cancer, as they are located on the cell membrane and are frequently expressed in high numbers. When an agonist ligand binds to them, the receptor/ligand complex is internalized, thus facilitating either peptide receptor scintigraphy or targeted radiotherapy (8-10). Many radiolabeled bombesin derivatives have been investigated by different groups, using full-length peptide, bombesin (7-14), or analogues modified to bind to all four receptor subtypes of the bombesin receptor family (2, 8, 9, 11-13). Lantry et al. (2) have demonstrated that 177Lu-AMBA (AMBA ) DO3A-CH2CO-G-(4-aminobenzoyl)-QWAVGHLM-NH2) binds with nanomolar affinity to GRP receptors. Preclinical studies demonstrated its therapeutic efficacy in a GRP-R positive PC-3 human prostate tumor-bearing mouse model. 177Lu-AMBA is now in clinical trials for the radiotherapeutic treatment of prostate cancer. This paper describes the identification of the metabolites of 177 Lu-AMBA, their receptor-binding characteristics, and their * Communicating author. Karen E. Linder, Bracco Research USA, 305 College Road East, Princeton, NJ 08540, E-mail: karen.linder@ bru.bracco.com, Phone: (609) 514-2416, Fax: (609) 514-2446.
biodistribution and routes of excretion. The synthesis and characterization of the proposed metabolites and their unmetalated ligands and in vitro and in vivo studies with these authentic standards are also described.
EXPERIMENTAL PROCEDURES Synthesis of AMBA and Metabolite Ligands A, B, C, D, and E. AMBA [(DO3A-CH2CO-G-(4-aminobenzoyl)-QWAVGHLM-NH2), DO3A ) (1,4,7,10-tetraaza-4,7,10-tris(carboxymethyl)-cyclododecyl)-acetyl] was synthesized using solid-phase peptide synthesis chemistry as described by Lantry et al. (2). The Fmoc-protected amino acids used were purchased from Nova-Biochem (San Diego, CA), Advanced ChemTech (Louisville, KY), Chem-Impex International (Wood Dale, Ill), and Multiple Peptide Systems (San Diego, CA). Other chemicals, reagents, and adsorbents were procured from Aldrich Chemical Co. (Milwaukee, WI) and VWR Scientific Products (Bridgeport, NJ). The peptide chains for unmetalated metabolite ligands A-E (Table 1) were assembled by either of two methods (methods A and B, below) of solid-phase peptide synthesis. For all constructs, the coupling of the glycine to the 4-aminobenzoyl attached resin and the DO3A-acetyl chelating moiety and cleavage from the resin were performed manually as indicated below. Method A. DIC (4 equiv) was added to a solution of amino acid (4 equiv, 0.25 M in NMP) and HOBT (4 equiv, 0.5 M in NMP). After 5 min, the solution was added to the resin in NMP (15 mL NMP/g of resin). After 4 h of shaking, the resin was filtered and washed 3× with NMP. For the next coupling, the Fmoc group was removed from the resin-bound peptide with 20% piperidine in DMF (15 mL/g resin) with shaking 2× for
10.1021/bc9000189 CCC: $40.75 2009 American Chemical Society Published on Web 05/29/2009
1172 Bioconjugate Chem., Vol. 20, No. 6, 2009
Figure 1. Chemical structure of
177
Linder et al.
Lu-AMBA showing the proposed coordination of the lutetium chelate.
Table 1. Summary of Ligands Synthesized ligand (DO3A-CH2CO-G-Abz4-R) R ) A: -Q-OH B: -QW-OH C: -QWAVGH-OH D: -QWAVGHLM-OH E: -QWAVGHL-OH
mass spectrum (API-ES) +
[M+H] : 709.2 ([M+2H]/2)2+: 355.1 [M+H]+: 895.3 ([M+2H]/2)2+: 448.2 [M+H]+: 1259.6 ([M+2H]/2)2+: 630.3 [M+H]+: 1503.5 ([M+2H]/2)2+: 752.3 [M+H]+: 1372.6 ([M+2H]/2)2+: 687.0
HPLC systema
retention time (purity)
3
2.5 min (100%) 4.7 min (100%) 4.8 min (100%) 5.9 min (>95%) 6.1 min (100%)
3 3 4 3
a System 3: XTerra MS C-18 (4.6 × 50 mm, 5 µ particle size): A ) H20 (0.1%TFA), B ) ACN (0.1% TFA). Gradient ) 0-40% B over 8 min, 3 mL/min. UV detection 220 nm. System 4: Initial condition: 10% B. Gradient ) 10-40% B over 10 min.
15 min. The resin was filtered and washed 4× with NMP. The next coupling was repeated as above until the Fmoc-4aminobenzoic acid had been added to the resin. Method B. Synthesis of the peptides was carried out using a 0.25 mmol FastMoc protocol on an ABI 433A peptide synthesizer under N2. In each cycle of this protocol, 1.0 mmol of a dry protected amino acid in a cartridge was dissolved in a solution of 0.9 mmol of HBTU, 2 mmol of DIEA, and 0.9 mmol of HOBT in DMF with additional NMP added. The activated Fmoc-amino acid was transferred directly to the reaction vessel. In the reaction vessel, 0.18 mmol of the C-terminal amino acid on 2-chlorotrityl resin (Novabiochem, resin substitution 0.42 mmol/g) was used. All couplings in this protocol were of 21 min duration. Fmoc deprotection was carried out with 20% piperidine in NMP. The peptide resin was finally washed with NMP for further manual derivatization. Coupling of Glycine to 4-Aminobenzoic Acid Portion of the Peptide Chain. Fmoc-Gly-OH (10 equiv) was dissolved in NMP (5.0 mL/g) and treated with 10 equiv of HATU. The mixture was stirred under N2 at 0 °C. DIEA (20 equiv) was directly added to the mixture and stirred for 10 min more before transferring the activated acid to the resin. The coupling was continued for 10 h at RT. The resin was drained of the reactants, and the above coupling procedure was repeated one more time. After draining the resin, it was washed with NMP (3×) and the resin was then taken to the cleavage step. General Coupling Procedure for DOTA-tri-t-Butyl Ester. DOTA-tri-t-butyl-OH (8 equiv) was dissolved in NMP (5 mL/ g). HBTU (8 equiv) was added and stirred under N2. DIEA (16 equiv) was added neat to the above mixture and stirred for 10 min at RT before transferring the activated acid to the resin. The total volume of the reaction mixture was adjusted to about 15 mL/g of the resin used. The coupling was continued overnight (20 h). The resin was drained, washed with NMP (3×), and then washed twice with CH2Cl2. Cleavage Procedure. Reagent B (15.0 mL, 88:5:5:2 TFA/ water/phenol/TIPS v/v/wt/v) was added to about 1.0 g of the resin, and the vessel was shaken for 4 h at RT. The resin was filtered and then washed twice with TFA (5 mL/g of resin). The filtrates were combined, concentrated to a syrup, triturated with 20 mL of ether/g of the resin used, and centrifuged. The supernatant was decanted, and the process was repeated three times. Crude peptides were purified with a Shimadzu HPLC
system using a YMC C-18 or Waters X Terra MS-C18 preparative column. The fractions were analyzed by reversedphase HPLC and MS; appropriate fractions were combined and lyophilized. Synthesis of 177Lu and natLu Complexes. All commercial reagents and solvents were of reagent grade or better and used as received. L-(+)-Selenomethionine (Se-Met) was obtained from Sabinsa Corp. Bacteriostatic 0.9% sodium chloride injection (U.S.P.) was purchased from Abbott Laboratories. ASCOR L500 ascorbic acid injection (U.S.P.) [containing 500 mg/mL ascorbic acid and 0.025% (w/v) Na2EDTA] was obtained from McGuff Pharmaceuticals, Inc. 177LuCl3 in 0.05 N HCl was purchased from Missouri University Research Reactor (MURR). Lutetium trichloride · 6H2O, 99.9%, was obtained from Alfa Aesar. 177 Lu-AMBA. 177Lu-AMBA was prepared and characterized as described by Chen et al. (14). For HPLC-purified product, the radiocomplex was separated from free ligand and radioactive and nonradioactive impurities by HPLC (Zorbax Bonus-RP Column, 5 µm, 80 Å pore, 250 mm × 4.6 mm, Agilent). For non-HPLC-purified product, the reaction mixture was diluted to the desired concentration using a 9:1 mixture of Bacteriostatic 0.9% sodium chloride injection USP and ASCOR L500 ascorbic acid injection USP to prevent radiolysis. HPLC-purified product had a specific activity of ∼3-4 Ci/µmol and a radiochemical purity (RCP) of >95%. Non-HPLC-purified product had a specific activity of ∼1.4 Ci/µmol and an RCP of g91%. nat Lu-AMBA and natLu-Labeled Metabolites A-E. LuCl3 · 6H2O (14 mg, 0.036 mmol, 1.5 equiv) was added to AMBA (36.5 mg, 0.024 mmol, 1 equiv) dissolved in 5 mL of 0.2 M NaOAc buffer, pH 4.5. The solution was refluxed for 15 min and then cooled to room temperature. The metalated product was purified by HPLC on a Vydac C-18 column (10 × 250 mm) using a gradient of 80% H2O (0.1% TFA)/20% CH3CN (0.1% TFA) to 76%/24% in 15 min, a flow rate of 4.5 mL/min and a column temperature of 37 °C. The desired fractions were lyophilized to yield 38 mg (75%) of Lu-AMBA · 2TFA · 10H2O as a white solid, based on elemental analysis (C, H, N, S, F, H2O). Authentic standards of natLu-labeled metabolites natLuA, -B, -C, -D, and -E were similarly prepared and isolated. Preparation of 177Lu-Labeled Metabolites A-E for Biodistribution Studies. Ligand C (10 µg in 10 µL of 0.2 M NaOAc, pH 4.8) was mixed with 177LuCl3 (5 µL, 2.11 mCi) and 0.2 M
In Vitro and in Vivo Metabolism of Lu-AMBA
NaOAc pH 4.8 (100 µL). The solution was heated at 100 °C for 10 min, cooled for 2 min in a water bath, and 5 µL of 5% Na2EDTA solution in water was added. The reaction mixture was purified by HPLC to remove unchelated ligand, and the desired peak was collected into a 9:1 mixture of bacteriostatic saline/Ascor solution containing 0.2% human serum albumin (HSA), evaporated to near dryness, and diluted to 50 µCi/mL for biodistribution studies. 177Lu-labeled metabolites A, B, D, and E were similarly prepared using the appropriate ligand. Hydrolysis of 177Lu-AMBA in Kidney Homogenate. Kidneys from nude mice were washed in PBS (all operations on ice), minced with scissors in Tris buffer (25 mM Tris, 150 mM NaCl, pH 7.4), and then Dounce-homogenized. The total volume of the homogenate was adjusted to contain 4 mL of Tris buffer for each gram of original tissue and then centrifuged at ∼4000 × g for 10 min at 4 °C. The supernatant was aliquoted (200 µL) and stored at -80 °C until use. HPLC-purified 177Lu-AMBA (25 µL, 3.50 Ci/µmol, 2.25 µCi, 0.00064 nmol) was incubated at 37 °C in 200 µL of kidney homogenate that was diluted 1:5 in Tris buffer immediately before use. Aliquots were removed and heat-denatured (100 °C, 10 min) at 10, 20, 30, and 60 min. The control was no homogenate, heated at 100 °C for 10 min. All samples were centrifuged at 20 000 × g for 6 min at 4 °C, and the supernates were analyzed by HPLC, using radiodetection. Hydrolysis of natLu-AMBA in Kidney Homogenate. Kidney homogenate (25 µL), Tris buffer (15 µL), and natLu-AMBA (30 µL, 6.9 mg/mL) were incubated at 37 °C. At 0, 30, or 60 min, 23 µL of the mixture was removed, combined with 46 µL of cold methanol on ice, and centrifuged at 4 °C for 10 min at 20 000 × g to remove the protein. Preliminary LC/MS analyses were performed on the supernatants using an Agilent 1100 Series LC/MSD and a Waters XTerra C18 column (50 × 4.6 mm, 5 µm pore size) eluted with a gradient of H2O (0.1% TFA)/ACN (0.1% TFA) at a flow rate of 0.5 mL/min. To confirm the identity of the metabolites, each metabolite peak was isolated from the supernate solution by preparative HPLC on a Vydac C18 column (5 µm, 4.6 × 250 mm) eluted at 1.5 mL/min with a gradient of 0-36% B over 26 min, where A ) H2O (0.1% TFA) and B ) ACN (0.085% TFA). The desired peaks were detected at 280 nm. Solvents were evaporated; the samples were redissolved in 1:1 H2O/ACN and analyzed by direct-injection MS. In Vitro Metabolism in Mouse, Rat, and Human Plasma. Pooled normal human, rat, and mouse plasma (sodium citrate form) was obtained from Innovative Research, Inc. (Southfield, MI). 177Lu-AMBA was radiolabeled as described above, but bacteriostatic saline and Ascor were omitted to avoid possible interference from EDTA and ascorbic acid. Immediately after labeling, 10 µL (1 mCi) of 177Lu-AMBA was mixed with 0.99 mL of plasma and incubated at 37 °C (n ) 4 per time point). At various times, sample solution (200 µL) was mixed with 10 µL of 10 mM DTPA, proteins were precipitated by treatment with 0.1 mL of 50% ice-cold EtOH (containing 5 mg/mL SeMet) and 0.3 mL of ice-cold ACN, followed by centrifugation for 20 min at 20 000 × g at 4 °C. Recovery of radioactivity in the supernates was ∼90%; the isolated supernates showed good stability (18 h at RT). These samples were analyzed using HPLC System 1. HPLC Analysis of in Vitro Plasma Samples (System 1). HPLC analysis of in vitro plasma supernates was performed on a Zorbax Bonus-RP column (4.6 × 250 mm; 5 µm, Agilent) using the gradient in Table 2, where A, H2O; B, 30 mM (NH4)2SO4/0.1% TFA (v/v); C, acetonitrile (ACN); D, methanol (MeOH); flow rate ) 1.5 mL/min, with a column temperature of 37 °C. The amounts of 177Lu-AMBA (% RCP) and metabolites were determined at each time-point from HPLC radiochromatograms.
Bioconjugate Chem., Vol. 20, No. 6, 2009 1173 Table 2 mobile phases and percentage (%) time
A
B
C
D
0-5 min 15 min 30 min 35-40 min 45 min
70 44 30 0 70
30 30 30 30 30
0 13 20 35 0
0 13 20 35 0
A one- or two-compartment open model was used to fit the standard deviation-weighted, mean % RCP data in human, rat, and mouse plasma, using eq 1 [C ) Ae-at + Be-bt ], which was solved using the SAAM II program (19-21). The LuAMBA half-life was then interpolated from the fitted curves. 177 Lu-AMBA Metabolism in Normal Mouse and Rats. Animal studies were conducted in accordance with the U.S. Public Health Service Policy on Human Care and Use of Laboratory Animals as well as institutional guidelines. NonHPLC purified 177Lu-AMBA was diluted to 100 µCi/ 0.1 mL using 9:1 bacteriostatic saline/Ascor solution (RCP ) 91%) for studies in mice and 750 µCi/ 0.1 mL (RCP ∼94.0%) for studies in rats. 177Lu-AMBA (100 µCi, 0.1 mL) was injected into a lateral tail vein of each of six mice (CD-1 normal males, Charles River Laboratories, ∼8 weeks old, 35 g). Two mice each were sacrificed at 2 and 10 min. Urine was collected from the bladders of the sacrificed animals at t ) 10 min and pooled. At 30 and 60 min, urine samples were taken from the bladder of one additional mouse each. Blood was taken from the descending aorta following CO2 asphyxiation and stored on ice. Within 10 min, it was centrifuged at 4 °C for 15 min at 1000 × g to remove the cellular fraction, and the plasma fraction was removed and kept on ice. Urine samples were immediately placed on ice and treated in the same fashion as the plasma samples. An accurate aliquot of the plasma or urine was vortexed on ice with an equal volume of 50% EtOH containing 0.1% TFA (v/v), 0.05 mg/ mL of Se-Met, and 0.05 mg/mL EDTA; treated with 3 parts of ice-chilled ACN and centrifuged at 4 °C for 15 min at 20 000 × g. Extracts were transferred to glass autosampler vials for HPLC analysis. For studies in normal male CD rats (Charles River Laboratories), 0.1 mL (750 µCi) of 177Lu-AMBA was injected into a lateral tail vein. In the first study, three rats were injected, each ∼4 weeks in age and 90 g. In the followup study, a single rat ∼6 weeks in age and 150 g was injected in the same manner. Blood was collected from conscious animals restrained in a soft plastic cone from the contralateral tail vein using heparin-rinsed syringes fitted with 21-gauge infusion sets. A separate syringe and infusion set was used for each time point. Urine samples were collected at 30 min and 2 h post injection (n ) 1/time point). HPLC analysis of 177Lu-Labeled Metabolites in Vivo (HPLC System 2). Plasma extracts and urine samples were analyzed by HPLC using an Agilent 1100 series quaternary HPLC equipped with an autosampler and variable wavelength detector. Radioactivity was monitored using a Canberra NaI detector (model 802-2 × 2W) with high-voltage power supply (model 3102D), single channel analyzer (model 2015A), and a linear/logarithmic ratemeter (model 1481LA). Column: Vydac C18 (250 mm × 4.6 mm); column temp 37 ( 1 °C; mobile phase A, H2O with 0.1% TFA (v/v); mobile phase B, ACN with 0.085% TFA; flow rate, 1.5 mL/min. Gradient: 0% B at 0 min, 20% B at 15 min, 23% B at 27 min, 67% B at 30 min, return to initial conditions at 36 min. In Vivo Biodistribution of Isolated Metabolites. TAC:Cr: (NCr)-Foxn1nu homozygous 4-5 week old male mice (Taconic Farms Inc.) were xenografted with human prostate cancer PC-3 cells (2 × 106). Studies were performed with tumors ranging
1174 Bioconjugate Chem., Vol. 20, No. 6, 2009
Linder et al.
Figure 2. HPLC chromatogram of 177Lu-AMBA (177Lu-DO3A-CH2CO-QWAVGHLM-NH2) after 30 min incubation in mouse kidney homogenate at 37 °C. Metabolites A, B, and C are 177Lu-DO3A-CH2CO-R, where R ) -Q-OH, -QW-OH, and -QWAVGH-OH. HPLC System 2 was used for this analysis.
from 125-1800 mm3. Each subject (n ) 4/group) was administered, via i.v. tail vein, a 0.1 mL (5 µCi) dose of HPLC-purified 177 Lu-AMBA or 177Lu-labeled metabolite A-E (RCP > 96%, specific activity 1.9-3.54 Ci/µmol). Animals were sacrificed at 1 or 24 h via cervical dislocation, and tissues were harvested, weighed, and assayed for residual radioactivity in a Perkin-Elmer model 1480 Wizard gamma counter. In Vitro Binding Studies - Cell Culture. Competition and direct binding studies were performed at 4 °C using the human prostate cancer cell line PC-3 (androgen-resistant prostate adenocarcinoma; American type Culture Collection) following procedures described previously (2). Plates were used for competition and direct binding studies on day 3 postplating, at which time the cells were confluent. Assays were performed in binding buffer containing protease inhibitors (RPMI-1640 supplemented with 0.2% BSA, 0.5 mM [4-(2-aminoethyl) benzenesulfonyl fluoride] (AEBSF) and Bacitracin (50 mg/500 mL), pH 7.4). IC50 Determination. AMBA or metabolite ligands (A-E) (without added Lutetium) were diluted to concentrations ranging from 1.25× 10-9 M to 5× 10-7 M in binding buffer. Competition binding studies were performed to determine their inhibitory concentration (IC50) against binding of 177Lu-AMBA. After the addition of the appropriate solutions, plates were incubated for 1 h at 4 °C (n ) 3 wells/data point). Incubation was ended by the addition of 200 µL of ice-cold binding buffer. Plates were washed five times, blotted dry, and counted in either an LKB CompuGamma counter or a microplate scintillation counter. Bound cpm (177Lu) was plotted against concentration of metabolites to determine the concentration of the metabolite at 50% inhibition (IC50). Direct Binding of 177Lu-AMBA Metabolites to PC-3 Cells. HPLC-purified 177Lu-labeled metabolite A, B, or C (75 µL, RCP >98%) was added to triplicate wells at final concentrations of 0-10 µCi/mL in binding buffer. The cells were incubated for 1 h at room temperature. After washing off the unbound radioactivity (0.1% BSA in PBS, pH 7.4), total radioactivity bound to the cells was determined. In a parallel experiment, cells were incubated with various concentrations of 177Lu-labeled metabolites in the presence of their respective unmetalated ligand (2 µM) to determine any nonspecific binding.
RESULTS In Vitro Metabolism in Liver and Kidney Homogenate. The enzymatic cleavage of 177Lu-AMBA was initially studied in mouse kidney homogenates for up to 60 min, followed by TLC analysis. These studies indicated that 177Lu-AMBA was readily degraded in a time-dependent manner into radioactive products that were more polar than the parent 177Lu-AMBA (data not shown). To identify the hydrolysis/cleavage products, nonradioactive natLu-AMBA or 177Lu-AMBA was incubated with mouse kidney homogenate at 37 °C for 0, 30, and 60 min, and the incubation mixtures analyzed by LC-MS or HPLC. The
Figure 3. Comparison of the in vitro metabolism of 177Lu-AMBA at 37 °C (retention time ) 25 min) in mouse plasma (t ) 0 and 6 h) and in human plasma (t ) 0, 6, and 72 h). Metabolism is significantly slower in human plasma. The metabolites observed were Lu-DO3A-CH2COG-Abz4-R, where R ) Q-OH (A), QW-OH (B), QWAVGH-OH (C), QWAVGHLM-OH (D), and QWAVGHL-OH (E). HPLC System 1 was used for this analysis.
major products were identified as Lu-DO3A-CH2CO-G-AbzQ-OH [m/z 881.3, M+H+], Lu-DO3A-CH2CO-G-Abz-QW-OH [m/z 1067, M+H+], and Lu-DO3A-CH2CO-G-Abz4-QWAVGHOH [m/z 1431, M+H+] based on LC-MS results. The time course for this study (data not shown) suggested that Lu-AMBA was first cleaved to Lu-DO3A-CH2CO-G-Abz4-R, where R ) QWAVGH-OH (Metabolite C), which was in turn further hydrolyzed into smaller metabolites, where R ) QW-OH and -Q-OH (Metabolites B and A). After 30 min at 37 °C, 177LuAMBA had been completely converted to 177Lu-containing metabolites Lu-A, -B, and -C, as shown in the radiochromatogram in Figure 2. In Vitro Studies in Mouse, Rat and Human Plasma. In contrast to the results obtained with kidney and liver homogenate, the metabolism of 177Lu-AMBA in plasma in vitro was very slow. Figure 3 shows typical radiochromatograms of the time course in these studies in mouse and human plasma. Full data sets for studies in mouse, rat, and human plasma are provided in the Supporting Information (Tables S1-2 and Figures 1-3S). 177 Lu-AMBA was metabolized faster in mouse and rat plasma than in human plasma. Radiochemical purity (RCP) data from these studies were used to calculate the half-lives of 177Lu-
In Vitro and in Vivo Metabolism of Lu-AMBA
Figure 4. Disappearance of intact 177Lu-AMBA in human, rat, and mouse plasma in vitro. Error bars on experimental data are indicated.
AMBA in human, rat and mouse plasma in vitro. Figure 4 shows the fits to the data. The half-lives of 177Lu-AMBA in human, rat, and mouse plasma in vitro were found to be 38.8 ( 1.3 h, 8.1 ( 3.8 h, and 3.1 ( 0.1 h, respectively. Nature of Metabolites. Table 3 summarizes the five 177LuAMBA metabolites observed in plasma in vitro. Lu-AMBA has a retention time of 25 min in the HPLC system used for analysis of in vitro plasma samples (HPLC System 1). The peaks at tR )18 min and tR ) 17 min corresponded to the metabolites identified by LC/MS from studies in mouse kidney homogenate: Lu-DO3A-CH2CO-G-Abz4-QW-OH (B) and Lu-DO3A-CH2COG-Abz4-QWAVGH-OH (C), respectively. The metabolite that elutes after Lu-AMBA (D, tR ) 26 min) coeluted with an authentic standard of Lu-DO3A-CH2CO-G-Abz4-QWAVGHLMOH, a compound in which the terminal amide of Lu-AMBA is converted to a carboxylic acid. This metabolite was especially predominant in rat plasma in vitro. The polar metabolite labeled A, tR) 9 min, coeluted with an authentic standard of Lu-DO3ACH2CO-G-Abz4-Q-OH. Minor metabolite E coeluted with an authentic standard of Lu-DO3A-CH2CO-G-Abz4-QWAVGHLOH (Lu-AMBA without its terminal methioninamide residue). All assignments were confirmed by coinjection of authentic standards of the proposed metabolites. Chemical Synthesis of Metabolites. Standards of the unmetalated ligands for Lu-AMBA metabolites A-E were readily synthesized by automated solid-phase peptide synthesis and/or manual methods as described above. HPLC analysis of the peptides showed them all to be >95% pure; the appropriate molecular weight was observed by API-ES mass spectra in all cases (Table 1). Their natLu analogues were prepared by heating excess LuCl3 · 6H2O with the appropriate ligand in acetate buffer, followed by HPLC purification and lyophilization. These natLumetabolite standards showed the appropriate mass spectra and coeluted with their 177Lu analogs, using HPLC Systems 1 and 2 (Table 3). In Vitro Cell Binding Studies with Synthesized Metabolite Ligands. In vitro competition binding of AMBA metabolites to GRP receptors in PC-3 cells were performed by competing 177 Lu-AMBA with increasing concentrations of each of the unmetalated metabolite ligands. The results are shown in Table 4. While AMBA ligand competed with 177Lu-AMBA at 5 nM (IC50), the five metabolite ligands showed little or no competition, with IC50 >400-500 nM. Furthermore, direct binding studies performed with the 177Lu-labeled metabolites A, B, and C did not demonstrate any specific binding to PC-3 cells. In Vivo PK Studies on 177Lu-AMBA - Metabolites in Mouse/Rat Plasma and Urine. The clearance and metabolism of 177Lu-AMBA following injection into male CD-1 mice was studied to determine its pharmacokinetics and the rate at which metabolites formed in vivo. HPLC traces obtained from mouse plasma samples obtained at 2 and 10 min are shown in Figure
Bioconjugate Chem., Vol. 20, No. 6, 2009 1175
5. Fit-functions for 177Lu-AMBA in rats showed an initial clearance phase (t1/2 of 0.66 min) and a terminal clearance phase with a t1/2 of 8.64 min. Modeling results for the metabolism of Lu-AMBA in rat plasma are shown in Figure 6. A radiochromatogram of mouse urine at 10 min post-injection is shown in Figure 5. Mouse urine at 60 min and rat urine at 30 min had a similar appearance. No parent 177Lu-AMBA was found in the urine at any time point; only 177Lu-A and 177Lu-B were observed. In Vivo Biodistribution of Lu-AMBA and Its Metabolites. Biodistribution results for 177Lu-AMBA and its five metabolites at 1 and 24 h are shown in Table 5. The major elimination route for all 177Lu-AMBA metabolites was by renal excretion (78-87%). As reported previously (2), 177Lu-AMBA showed high uptake in PC-3 tumors and mouse pancreas, both of which express high numbers of GRP receptors. No significant uptake in tumor or pancreas was noted for any of the 177Lu-AMBA metabolites, confirming the in vitro IC50 and direct binding study results.
DISCUSSION In Vitro Plasma. The metabolites formed from 177Lu-AMBA in mouse, rat, and human plasma in vitro were the same across species, but the relative amounts of the various metabolites formed were different. At intermediate time points in mouse and rat plasma, Lu-labeled metabolites C and D predominated, with only small amounts of Lu-A and -B observed. In contrast, human plasma samples in vitro formed only trace amounts of Lu-labeled D; Metabolites Lu-B and Lu-C were the predominant species observed. Lu-A formed slowly in all cases. In vitro metabolism was slow in all three types of plasma, with observed half-lives of hours (Figure 4). In vitro metabolism half-lives were significantly slower if performed in plasma containing EDTA as an anticoagulant (data not shown). Inhibition of 177Lu-AMBA metabolism by EDTA strongly suggests that metalloenzymes are responsible for the metabolism of this peptide complex, as EDTA can bind to the metals found in the active site of metalloenzymes, thus inactivating them. In Vivo Mouse and Rat. In contrast to in vitro results, the metabolism of 177Lu-AMBA in vivo was rapid. Two minutes after IV administration to normal mice, only ∼45% of the radioactivity in the plasma was present as 177Lu-AMBA. The remainder was present as metabolites Lu-A, -B, -C, -D, and -E, the products previously identified from in vitro studies. By 10 min post-injection, 78-87% of the radioactivity in mouse plasma had been converted to these metabolites. Similar results were obtained in the normal rat, although the rate of 177LuAMBA metabolism was somewhat slower. By 1 h post-injection, no 177Lu-AMBA was detectable. In both species, metabolites Lu-A, -B, and -C were found in the plasma. Lu-D was also a predominant species in mice, but was much less prominent in samples of rat plasma. Lu-E was found only in trace amounts. Radioactivity cleared rapidly into the urine in both mouse and rat. For all urine samples analyzed, ∼95-100% of the radioactivity detected was due to metabolites Lu-A and -B. No parent 177Lu-AMBA was excreted into any urine sample. Nature of the Metabolites. All metabolites identified were due to selective cleavage of the -QWAVGHLM-NH2 peptide chain responsible for binding to the GRP receptor, including removal of the terminal amide from the methioninamide residue to form metabolite D, loss of methionine to form E, cleavage of the HvL bond to form C, cleavage of the WvA bond to form B, and cleavage of the QvW bond, forming A. Consistent with the concept of enzymatic control of small bioactive peptides, we note that the metabolites reported here are inactive. Thus, it is clear that attempts to stabilize such
1176 Bioconjugate Chem., Vol. 20, No. 6, 2009
Linder et al.
Table 3. Summary of Lu-AMBA Metabolites and Their Sequencesa Lu metabolite
Lu-DO3A-CH2CO-G-Abz4-R R )
MW
mass spectrum
Parent Lu-AMBA
-QWAVGHLM-NH2
1674.9
Metabolite Lu-A
-Q-OH
Metabolite Lu-B
-QW-OH
1066.9
Metabolite Lu-C
-QWAVGH-OH
1431.3
Metabolite Lu-D
-QWAVGHLM-OH
1675.9
Metabolite Lu-E
-QWAVGHL-OH
1544.4
[M+H]+: 1675.3 [M+2H]2+/2: 837.9 [M+Na+H]2+/2: 848.8 [M+H]+: 881.3 [M+Na]+: 903.4 [M+2Na]2+/2: 463.2 [M+H]+: 1067.5 [M+Na]+: 1089.5 [M+2Na]2+/2: 556.2 [M+H]+: 1431.4 [2M+3H]3+/3: 954.9 [M+2H]2+/2: 716.2 [M+H]+: 1675.4 [M+2H]2+/2: 837.8 ND
a
880.7
HPLC System 1 tR (min) (in vitro studies)
HPLC System 2 tR (min) (in vivo studies)
25.2
22.5
9.0
7.9
17.8
14.8
17.0
15.6
26.6
23.2
23
19
ND ) not determined; NI ) not identified.
Table 4. Competition of Unmetalated Lu-AMBA Metabolite Ligands with 177Lu-AMBA for GRP-R Binding in PC-3 Cells symbol
metabolite ligands
IC50 [nM]
AMBA A B C D E
DO3A-CH2CO-G-Abz4-QWAVGHLM-NH2 DO3A-CH2CO-G-Abz4-Q-OH DO3A-CH2CO-G-Abz4-QW-OH DO3A-CH2CO-G-Abz4-QWAVGH-OH DO3A-CH2CO-G-Abz4-QWAVGHLM-OH DO3A-CH2CO-G-Abz4-QWAVGHL-OH
5 >500 >500 >500 >500 400
derivatives should focus on stabilization of these bonds. Garoyoa et al. attempted such a strategy in the synthesis of 99m Tc(CO)3 derivatives bound to a modified histidyl residue bearing various analogues of -QWAVGHLM-NH2. Replacement of Leu with the unnatural amino acid cyclohexylalanine (Cha) and/or the replacement of Met with norleucine (Nle) reportedly led to derivatives that had significantly improved stability when incubated in human plasma at 37 °C (22). These compounds still showed significant binding to the GRP receptor in vitro, and comparable or improved uptake in GRP receptor-positive pancreas and tumor in PC-3 tumor-bearing mice relative to unstabilized analogues, especially when a suitable spacer was inserted between the targeting group and the retroNR-carboxylmethyl histidine residue responsible for coordination to the Tc(CO)3 core. However, for the biodistributions reported, washout of the compounds from such
Figure 5. Radiochromatograms of in vivo mouse plasma and urine samples. HPLC System 2 was used for this analysis.
tissues was faster than that observed for 177Lu-AMBA (ref 2 and Table 5), a finding that is not ideal for radiotherapeutic applications. Similarly Ho¨hne et al. (23) chose to modify the Leu residue to Statin [(3S,4S)-4-amino-3-hydroxy-6-methylheptanoic acid], along with N-methylation of the glycyl amide bond for their 18F-labeled derivative 2-(4-(di-tertbutylfluorosilyl)phenyl)acetyl-Arg-Ava-Gln-Trp-Ala-ValNMeGly-His-Sta-Leu-NH2. This group reported little metabolism of this construct either in vitro or in vivo. However, this agent showed relatively low uptake in GRP-receptor positive pancreas in nude mice (4.67% I.D./g at 1 h), and even lower uptake in PC-3 tumors (0.63% I.D./g), although retention in these tissues over 4 h was good. Thus, the stabilizing effects of the modifications made will need to be balanced by their effects on receptor-binding affinity, internalization, biodistribution, and clearance characteristics. That bombesin and its analogues are readily metabolized has been known for some time. Rapid cleavage is not unexpected, as bombesin-like peptides have potent biological effects, stimulating gastric acid and pancreatic enzyme secretion, gall bladder contraction, modifying gut motility and affecting behavior, food and water intake, and thermoregulation (1). Rapid and highly selective proteolytic cleavage of bioactive peptides is common and allows close control over their effects (15). Two of the major metabolites observed are consistent with those observed by Shipp et al. (16), who studied the hydrolysis of bombesin, GRP, and GRP (14-27) with purified samples of the metalloprotease known as neutral endopeptidase 24.11 (NEP) or neprilysin. They found that cleavage occurred at the Trp-Ala and His-Leu peptide bonds in the conserved 7-amino acid C terminus region of these peptides [i.e., Q/H-WvAVGHvLM-NH2. In Lu-AMBA, such metabolism would give rise to metabolites Lu-B and Lu-C. Little has been reported on the metabolism of other radiolabeled bombesin derivatives. Zhang et al. (9) found that the modified bombesin (6-14) derivative 111In-DTPAGABA-[D-Tyr6-QWAV-β-Ala11-H-Thi13Nle14 was cleaved in human serum between Ala11 and the His residue, followed by metabolism to 111In-DTPA-GABA-D-Tyr-Q-OH and 111InDTPA. Addition of EDTA slowed the rate of metabolism, suggesting the involvement of a metalloprotease. The enzyme carnosinase was suggested to be the active enzyme responsible for cleavage. The 111In-DTPA-GABA-D-Tyr-Q-OH metabolite is analogous to metabolite Lu-A with Lu-AMBA, but no other similarities were seen. Nock et al. (17) reported that the 99mTc complex of a tetraamime ligand coupled to the bombesin antagonist sequence DPhe6-QWAVGHL-NHEt showed two polar metabolites in the urine following IV
In Vitro and in Vivo Metabolism of Lu-AMBA
Bioconjugate Chem., Vol. 20, No. 6, 2009 1177
Table 5. Biodistribution of the Five Known Metabolites of Lu-AMBA in PC-3 Tumor-bearing Tac:Cr:(NCr)-Foxn1 Homozygous Male Nude Mice (1 and 24 h data). 1 h 1 Hour parent organ tumor (%ID/g) liver (%ID) kidney (%ID) skin (%ID/g) muscle (%ID/g) pancreas (%ID) GI (%ID) blood (%ID) blad/urine (%ID)
177
Lu-AMBA
4.92 ( 1.57 0.26 ( 0.12 3.93 ( 0.85 0.28 ( 0.04 0.07 ( 0.02 17.76 ( 1.45 10.93 ( 0.53 0.25 ( 0.05 52.41 ( 6.43
Lu-AMBA metabolites (177Lu-A) n ) 4
(177Lu-B) n ) 4
(177Lu-C) n ) 4
0.35 ( 0.08 0.14 ( 0.05 0.85 ( 0.33 0.27 ( 0.10 0.07 ( 0.04 0.08 ( 0.05 1.41 ( 1.77 0.38 ( 0.22 86.76 ( 2.88
0.39 ( 0.08 0.13 ( 0.02 1.31 ( 1.41 0.26 ( 0.11 0.07 ( 0.05 0.03 ( 0.00 0.33 ( 0.11 0.29 ( 0.15 78.32 ( 25.01
0.33 ( 0.12 0.12 ( 0.03 1.27 ( 0.71 0.75 ( 1.06 0.07 ( 0.04 0.09 ( 0.10 0.25 ( 0.05 0.22 ( 0.11 86.10 ( 6.27
(e
Lu-D) n ) 4
177
(177Lu-E) n ) 4
0.62 ( 0.07 0.15 ( 0.04 1.36 ( 0.72 0.34 ( 0.12 0.19 ( 0.24 0.26 ( 0.03 0.57 ( 0.27 0.49 ( 0.020 86.45 ( 6.49
0.25 ( 0.16 0.12 ( 0.02 0.74 ( 0.22 0.15 ( 0.02 0.10 ( 0.11 0.04 ( 0.02 0.25 ( 0.07 0.21 ( 0.04 86.33 ( 5.06
24 Hour parent organ tumor (%ID/g) liver (%ID) kidney (%ID) skin (%ID/g) muscle (%ID/g) pancreas (%ID) GI (%ID) blood (%ID)
177
Lu-AMBA
3.21 ( 0.51 0.22 ( 0.27 1.03 ( 0.08 0.11 ( 0.04 0.01 ( 0.01 13.25 ( 1.03 6.38 ( 0.34 0.01 ( 0.01
Lu-AMBA metabolites (177Lu-A) n ) 4
(177Lu-B) n ) 4
(177Lu-C) n ) 4
(177Lu-D) n ) 4
(177Lu-E) n ) 4
0.04 ( 0.01 0.04 ( 0.02 0.23 ( 0.05 0.06 ( 0.01 0.01 ( 0.01 0.00 ( 0.00 0.04 ( 0.01 0.01 ( 0.01
0.06 ( 0.01 0.04 ( 0.00 0.20 ( 0.04 0.07 ( 0.02 0.03 ( 0.01 0.00 ( 0.00 0.16 ( 0.10 0.01 ( 0.01
0.06 ( 0.01 0.05 ( 0.00 0.28 ( 0.01 0.04 ( 0.01 0.04 ( 0.03 0.01 ( 0.00 0.05 ( 0.01 0.00 ( 0.01
0.16 ( 0.05 0.05 ( 0.001 0.30 ( 0.06 0.06 ( 0.01 0.02 ( 0.01 0.14 ( 0.01 0.14 ( 0.04 0.01 ( 0.01
0.07 ( 0.01 0.04 ( 0.01 0.23 ( 0.04 0.07 ( 0.01 0.03 ( 0.01 0.01 ( 0.00 0.07 ( 0.01 0.01 ( 0.01
injection, but the nature of these compounds was not determined, except to confirm that the products were not 99m TcO4-. In summary, Lu-AMBA is a derivative of bombesin and is expected to be metabolized relatively rapidly in vivo. Our in vivo results confirm this expectation. Of the radioactivity remaining in mouse plasma in vivo at 10 min post-injection, an average of 18% was parent 177Lu-AMBA, supporting a plasma half-life of
