Bbconjwte Chem. 1991, 2, 171-179
171
Astatine-211 Labeling of an Antimelanoma Antibody and Its Fab Fragment Using N-Succinimidylp-Astatobenzoate: Comparisons in Vivo with the p - [12SI]IodobenzoylConjugate Stephen W. Hadley,+ D. Scott Wilbur,'J Mary Ann Gray,+lgand Robert W. Atcherll University of Washington Medical Center, Seattle, Washington 98195, NeoRx Corporation, Seattle, Washington 98119, and Argonne National Laboratory, Argonne, Illinois 60439. Received January 7, 1991 Astatine-211 labeling of an antimelanoma antibody, NR-ML-05, and its Fab fragment with N-succinimidyl p - [211At]astatobenzoate(2a)has been described. Preparation of the astatinated intermediate 2a was accomplished by distilling astatine-211 from an irradiated bismuth target directly into a reaction mixture containing an organometallic compound, N-succinimidyl p-(tri-n-butylstanny1)benzoate(11, and an oxidant, N-chlorosuccinimide, in 5 % HOAc/MeOH. Trapping of distilled astatine as 2a was found to be efficient, resulting in 70-90% yields based on the amount of astatine-211 in the reaction mixture. The dry distillation technique employed gave recoveries of astatine-211 which ranged from 20% to 75%. Conjugation of 2a to NR-ML-05 and its Fab fragment was accomplished in 40430% yields. The [211At]astatoben~~yl-c~njugated antibodies were found to be stable in vitro when challenged by strong denaturants and nucleophilic reagents. Coinjected dual-labeled studies of the 2a astatinated antibodies and the same antibodies labeled with N-succinimidyl p-[12SI]iodobenzoate(2b) in athymic mice bearing the human tumor xenograft A375 Met/Mix demonstrated that both radiolabeled antibodies had equivalent tumor localization. Data from the dual-labeled biodistribution of the intact antibody suggests that the astatine is stably attached. Data from the dual-labeled Fab fragment suggests that a portion of the astatine label is released as astatide, either from the astatinated Fab or from a catabolite.
INTRODUCTION The application of a emitting radionuclides to radioimmunotherapy (RIT) of cancer is attractive due to the unique cytotoxic properties of the a particle ( I , 2). While there are concerns about general application to RIT, antibodies labeled with radionuclides that decay by a emission may have specific applications such as treatment of blood-borne cancers (e.g. leukemia), intracavitary administration for some types of cancers (e.g. peritoneal administration for ovarian carcinoma), or adjuvant treatment of disseminated micrometastatic disease (3). Of the many radionuclides that have a emissions, only bismuth212 ( t l p = 60.1 min) and astatine-211 ( t l 2 = 7.2 h) are presently feasible for application to labele antibodies (4, 5). Application of astatine-211-labeledantibodies for cancer therapy was suggested in 1954 (6). Since astatine is in the halogen group of elements it was reasoned that its chemistry would parallel radioiodine (7). Early attempts to radiolabel proteins ( 4 9 )and tyrosine (10)with astatine provided evidencethat its chemistry was different. Indeed, unlike radioiodine, labeling proteins directly with astatine produced a labile astatine-protein bond, which appeared to be formed from the reaction of astatine with the protein sulfhydryl groups (11). These studies indicated that another method of antibody labeling was needed. Friedman et al. (12) demonstrated that labeling of a protein with an astatinated benzoic acid resulted in a much less labile attachment of the astatine-211 label under the conditions encountered in vivo. In their study astatine-
ci
~~~
* Address correspondence to D. Scott Wilbur, Department of Radiation Oncology, RC-08, University of Washington Medical Center, 1969 N.E. Pacific St., Seattle, WA 98195. + NeoRx Corp. t University of Washington. I Current address: Schering-Plough Research, Department of Tumor Biology, Bloomfield, N J 07003. I Argonne National Laboratory. 1043-1802/91/2902-0171$02.50/0
211 was incorporated into benzoic acid via a diazonium salt decomposition. Protein conjugation of the resultant p - [211At]astatobenzoicacid was accomplished by preparing a mixed anhydride and subsequent reaction with the protein under basic conditions. Several other astatinated protein studies have been reported which used this method of labeling (13-15). A dramatic increase in the in vivo stability was obtained; however, there were several aspects of the labeling procedure which did not make it attractive for routine labeling. For example,diazonium salt reactions in aqueous solution have the potential to produce phenols (16,17)which could compete for the unreacted astatine211, resulting in an astatinated product that would be expected to be less stable than p-astatobenzoyl conjugates. Further, the labeling procedure required a number of chemical manipulations, such as ether extraction, which were time consumingand potentially dangerous. Perhaps the most important limitation was the overall time for the labeling process, taking up to 3 h to complete, which resulted in a loss of nearly 25% of the astatine-211 by decay alone. More recently, astatination of benzoic acid derivatives has been accomplished in high yield, without the concerns for phenolic byproducts, by using organometallic intermediates (18-21). For example, Zalutsky et al. (20,21) have labeled antibodies by using a protein-reactive organometallic intermediate (N-succinimidyl m-(trialkylstannyl)benzoate, termed ATE or MeATE), which could be astatinated readily and subsequently conjugated to the antibody. In this investigation, we have studied the use of the isomeric organometallic intermediate N-succinimidyl p - (tri-n-butylstannyl)benzoate (1) for astatine-211 labeling of antibodies as shown in Scheme I. The organometallic intermediate 1 used to obtain the proteinreactive astatinated product N-succinimidylp-astatobenzoate ( t a ) ,like the m-(trialkylstanny1)benzoylcompound employed by Zalutsky (22), had been previously used in 0 1991 American Chemical Society
Hadley et al.
172 Bloconlugate Chem., Vol. 2, No. 3, 1991
Scheme 1. Reaction Sequence for Astatine-211 (and Iodine-125) Labeling of Antibodies
0 1
/
/ /
0 2. X-At-211 2b X-1-125
Protein-NHP
3. X-At-211 3b X-1-125
radioiodinations of antibodies and had been shown to stabilize radioiodine from in vivo dehalogenation (23). The primary objective of the investigation described herein was to develop a method of labeling antibodies with astatine-211 for subsequent investigation to their application to therapy of metastatic lesions or blood-borne cancers. The goal was to obtain an astatine-211-labeling method which was rapid, high yielding, and could be conducted in a safe manner. The organometallic intermediate 1 was used to provide a facile incorporation of astatine-211 into the protein-reactive benzoate 2a. To simplify the chemical manipulations in the astatinations, we employed a method of distilling astatine (24)directly from the irradiated target into a reaction mixture containing 1. The astatine-labeling studies were carried out on an antimelanoma antibody, NR-ML-05, and its Fab fragment,l which had been previously studied employing the p-iodobenzoyl (PIB) conjugates (23). To determine if the labeling resulted in a method in which the desired antibody properties were retained, the in vitro stability and in vivo distribution of the astatinated antibodies were examined. To accomplish this, the astatinated antibodies were evaluated in vitro under a variety of challenge conditions and in vivo in athymic mice bearing a human tumor xenograft, A375 Met/Mix.2 For comparative purposes, NR-ML-OB, or its Fab fragment, was radioiodinated with 1261-labeledp-iodobenzoate (2b)and was coinjected with the astatinated antibody. EXPERIMENTAL PROCEDURES
General. All reagents used were analytical-reagent grade or better and were used as obtained. HPLC solvents were obtained as HPLC grade and were filtered (0.2 pm) prior to use. N-Succinimidyl p-(tri-n-butylstanny1)benzoate (1) was prepared as previously described (23). Disposable reverse-phase C-18 columns used for purification of radiolabeled intermediates were obtained from J.T. Baker, Inc. (6 mL, #7020-6). Astatine-21 1 was produced at Argonne National Laboratory on the 6041-1.cyclotron (25). The fixed-frequency 42 MeV beam was degraded with aluminum foils to ca. 28 MeV to avoid production of astatine-210 (26). Irradiation of aluminum-backed 0.5-0.8 mm thick bismuth targets NR-ML-OB is an Ig&b antibody that is reactive with a high molecular weight antigen on most melanoma cells (50). The F(ab’h fragment could not be prepared. a The A375 Met/Mix xenograft is a metastatic variant of the A375 cell line isolated from nude mice. (51).
was generally done a t 10-12 PA and produced 13 mCi in 4 h. Once irradiated, the targets were shipped to Seattle for use in the experiments conducted the following day. Radioiodine was obtained from Du Pont/NEN (Billerica, MA) as high-concentration, high specific activity solutions in 0.1 N NaOH. Phosphate-buffered saline was purchased from Gibco Laboratories (Baltimore, MD) as Dulbecco’s phosphate-buffered saline (#310-4190)or from Whittaker Bioproducts, Inc. (Walkersville, MD). Protein Concentrations. Protein concentrations in solution were measured by UV spectrophotometry a t 280 nm. The concentrations were obtained with the BeerLambert equation with the molar absorptivity being equal to 1.43 for whole antibody and 1.53 for Fab. Astatine-211 Measurements. Radioactive materials were counted on a Packard Autogamma 5650 y scintillation counter. Multiple standards were used in each counting to assess the quantity of astatine and radioiodine injected. Radioactivity counting was carried out with a low-energy window (20-80 keV) for both astatine-211 (211PoX-rays) and iodine-125. Initial astatine-211 activity in the irradiated target was determined (at Argonne) using an EG&G Ortec Ace multichannel analyzer (MCA) with a modified version of the program Maestro. The amount of activity was determined with an Ortec Gamma-X Intrinsic Ge detector a t a distance of 80-150 cm from the target. The integral of the 687 keV y peak which is present in 0.245% abundance from the electron decay branch of 211At was used as the basis for quantification. The measurement of astatine activity was calibrated with a standard traceable to NIST, standard set 2200 R2317. To help cross-calibrate the instruments a t Argonne and in these laboratories, a sample was distilled at Argonne and quantified by MCA. The sample was then placed in a CRC-7 dose calibrator which was adjusted with the potentiometer until it read the correct value. That derived calibration factor, 25, was set on the CRC-7 dose calibrator in Seattle for readings of distilled astatine quantities. For quantification of the percent recovery of astatine from the irradiated target, the activity was measured by a Canberra Series 20 Model 2802 MCA using a Canberra Model 2007P (preamp) and Model 802-3 thin NaI crystal detector to obtain the integrated area under 687 keV peak a t a set distance (ca. 60 cm) before and after distillation. These values were compared with the quantities of astatine in the reaction vial and in other parts of the distillation apparatus. Total astatine activity balances of 95-120% were obtained. Analytical Chromatography. High-performance liquid chromatography (HPLC) was used to determine the purity of the radiolabeled species obtained from the reaction mixtures and to assist in the development/ assessment of preparative purification methods (see Astatine Distillation and Labeling Procedure). HPLC was conducted on either a gradient system for separation of small molecules (1,2a,b)or an isocratic system for separations of labeled proteins (3a,b). The gradient system consisted of two Beckman Model llOB pumps, a Beckman Model 153 UV detector, a Rheodyne Model 7125 injector, and a Beckman Model 170radioisotope detector. The isocratic system was identical with the gradient except it had only one solvent pump. All separations were run at a flow rate of 1mL/min. Spectrophotometric analysis of the HPLC effluent was performed with UV detection at 254 nm. Reverse-phase chromatography was conducted on a 4.6 mm X 12.5 cm C-18 column (Whatman) using a binary gradient solvent system which minimized outgassing. Eluant A in the gradient was a mixture of 98% MeOH and
pAstatobenzoyl Antibody Labeling
Bioconlugte Chem., Vol. 2,
No. 3, 1991 173
L B
K A
Figure 1. Schematic of astatine distillation setup. Components are (A) reaction vessel; (B) charcoal filter; (C) Teflon tubing, 15 cm X 0.8 mm i.d. X 1.5 mm o.d., fitted on one end with a l/rZ8 threaded polypropylene bushing; (D) coupling, l/4-28 thread polypropylene; (E)Tefzel adapter, l/4-28 thread to a female Luer adapter; (F) Pyrex Luer tip; (G) graded seal; (H) quartz distillation tube; (I) irradiated bismuth target; (J) quartz tube boat; (K) standard ball/socket joint (50/30);(L) carrier gas inlet.
2 % of a 1% HOAc/H20 solution. Eluant B was a mixture of 20% MeOH and 80% of a 1% HOAc/H20 solution. The gradient was begun by eluting for 2 min at 50% of each eluant. After 2 min the solvent mixture was increased in eluant A over the next 8 min to a final composition of 98% eluant A and 2% eluant B. The eluant mixture remained at that composition for the next 10 min to complete the gradient. Retention times for N-succinimidyl p - (tri-n-butylstannyl) benzoate (1) was 14.6 min and for astatine-211-labeled benzoate 2a was 4.1 min with the gradient described. The retention time for p-astatobenzoic acid was 4.9 min and astatide ion came off at the solvent front, 1.2 min. Size-exclusion chromatographywas conducted on a Zorbax Bio Series GF-250 9.4 mm X 24 cm column (Du Pont) with 0.2 M phosphate buffer, pH 6.8, as the eluant. The retention times for NR-ML-05 labeled with 2a or 2b on the size-exclusion column were 8.5 min for intact IgG and 9.8 min for Fab. Thin-layer chromatography of protein conjugates was conducted on silica gel impregnated glass-fiber strips (ITLC, Gelman) eluting with 80% MeOH/H20. In this system, the protein is denatured at the origin and the free halide or radiolabeled small molecules move with the solvent front. Radiochemical purity of the radiolabeled Fab preparations was defined as the percent protein-bound activity, which was estimated by dividing the counts at the TLC origin by the total counts on the TLC strip. Astatine Distillation/Reaction Setup. The astatination reactions were conducted in a glovebox that was vented through a charcoal filter into an adjacent fume hood. The astatine distillation and labeling reactions employing 1 were conducted in the glassware setup depicted in Figure 1. The quartz distilling tube was prepared with dimensions to fit the commercially available tube oven (Thermolyne Model 21100). Prior to insertion into the distillation oven the 17 mm X 132 mm X 1.25 mm irradiated targets were placed in a single-use 22 mm 0.d. X 180-200 mm quartz tube (boat). This was done to alleviate the problem of cracking caused by the cooling of dripped molten bismuth (mp 271 "C) in contact with the quartz distilling tube. As an additional precaution, a larger quartz tube, 25 mm i.d., was kept in the distilling tube as aliner. The quartz distilling tube was coupled to the glass tube by a Tefzel l/4-28 thread (Alltech, #20055) to a female Luer fitting. Distillation of astatine was accomplished at 600-700 "C with a stream of argon to carry the astatine into the reaction vessel. The distillation apparatus was vented to the inside of the glovebox through a charcoal filter prepared from a packed syringe barrel. Introduction of reagents into the reaction vessel was done by prefilled syringe. Distillation glassware, Teflon tubing, and fittings were soaked in concentrated HN03 and rinsed well with deionized H20 prior to each distillation.
Astatine Distillation and Labeling Procedure. The distillation apparatus was placed in the glovebox and assembled. The reaction vial was removed from the quartz distillation tube by disconnecting the Tefzel adapter. In the experiments with the best recovery of astatine-211 from the irradiated target, the furnace was preheated to ca. 650 "C under a stream of argon and allowed to equilibrate for 30-60 min. During the equilibrium process the argon gas was allowed to pass through the heated still and exit into the glove box. While this was occurring, the reaction mixture was prepared in the following manner. Into the reaction vial was placed 100 pL of a 1mg/mL solution of 1 (0.20 mmol) in 1% glacial HOAc/MeOH, 80 pL of a 1mg/mL solution of N-chlorosuccinimide (NCS) (0.60 mmol) in MeOH, 400 pL of a 5% glacial HOAc/ MeOH mixture, and 50 pL of phosphate-buffered saline (PBS). After the oven temperature had equilibrated, the reaction vial was reconnectedto the distillation apparatus. The flow rate of argon gas was metered to between 25 and 75 mL/min. An irradiated bismuth target was wiped and placed into a quartz boat. The preheated quartz distilling tube was opened by disconnecting the carrier gas inlet adapter at the ball joint end of the still. The quartz boat containing the irradiated target was carefully slid into the center of the heated quartz distillation tube with tongs. The ball joint was quickly reconnected and clamped to secure it. The distillation was allowed to proceed for 1h with the reaction vial at ambient temperature. During the distillation, the volume in the reaction vial was maintained by addition of two 25-pL aliquots of MeOH. The distillation was terminated by turning off the furnace and the argon flow. A 10-pL aliquot of a 5.8 mg/mL aqueous solution of Na2S205 (0.61 mmol of NaHS03) was added to the reaction mixture to reduce any volatile astatine species. The reaction vial was then removed from the distilling apparatus by disconnecting the Tefzel adapter from the Luer tip. The labeled 2a was separated from [211At]astatide and the aryltin compound 1prior to its use in the conjugation reaction as follows. Two reverse-phase C-18 disposable columns (J.T. Baker, SPE octadecyl C-18) were activated by elution with 5 mL of 95% EtOH, followed by 5 mL of PBS. The solution containing astatinated benzoate 2a prepared above was then loaded onto one of the columns. The reaction vial was rinsed with less than 0.5 mL of EtOH/PBS (1:l)and six 1-mLfractions were eluted withEtOH/PBS (6:4). The activity in each of the eluted fractions was measured and those fractions with the highest levels of activity (typically 60-70% in fractions 3 and 4) were used in the following steps. To each of the 1-mL fractions with the highest activity was added 2 mL of PBS. The entire volume of the diluted fractions was loaded onto the second C-18 column and was eluted through the column. The astatinated benzoate 2a remained on the column. The column
174 Bloconlugste Chem., Vol. 2, No. 3, 1991
Figure 2. HPLC radiochromatogram of astatination reaction mixture after purification by two C-18 columns. Retention times indicate that peak A (4.14 min) is N-succinimidyl p-astatobenzoate (2a), peak B (4.94 min) is p-astatobenzoic acid, peak C (solvent front) is likely to be remaining astatide, and peaks D (2.88 min) and E (9.28 min) are unidetermined.
was then dried by passing a stream of nitrogen gas over it for ca. 5 min. The activity was eluted from the dried column with dry CHsCN, collecting three l-mL fractions. Typically >90% of the activity was in fractions 1 and 2. These fractions were combined and the CH3CN was then removed under a gentle stream of dry nitrogen. A typical radio-HPLC chromatogram of the combined fractions is shown in Figure 2. The concentrated material, composed principally of 2a, was used for antibody labeling by using the following procedure. To the vial containing 2a was added 500 pL of a solution containing 1mg of NR-ML-05 antibody in 200 pL of PBS and 300 pL of 1.0 M, pH 9.2, NaHC03/Na&O3 buffer. After 10 min at room temperature, the crude labeled antibody was purified by size-exclusion chromatography (Sephadex G-25, PD-10) eluting with PBS. Overall antibody astatine-211-labeling yields of 34-41 7% (decay corrected) have been achieved. Radioiodine Labeling. Labeling of NR-ML-05 and its Fab fragment with iodobenzoate2b used in the in vitro and in vivo experiments was conducted as previously described (231,except that 2b was separated from 1 and [1BI]iodidevia the two (2-18columns prior to conjugation as described in the procedure above. Overall antibody radioiodine-labelingyields of 42-47 7% were obtained using this procedure. In Vitro Stability Experiments. The stability of the astatine-211-labeled NR-ML-05 was examined by subjecting it to solutions of PBS and a variety of reagents at 37 OC for several hours. In the experiments test tubes were charged with 0.5 mL of (a) PBS, (b) 1M glycine in PBS, (c) 1M NazCOs/NaHC03 buffer, pH 9.2, and (d) 1 M glycine + 1 M NazCOs/NaHC03. To each test tube was added 65 fig of astatinated antibody. After this addition, the samples were lightly vortexed and an aliquot was removed and analyzed by TLC (80% MeOH) as the t = 0 time point. The tubes were then incubated at 37 OC for 20h. Aliquota were removed periodically and examined by TLC. Animals and Tumor Model. Athymic (nude) mice were obtained from Simonsen Laboratories, Inc. (Gilroy,
liadky et ai.
CA) and were used at 10-12 weeks of age. Female athymic mice (nu/nu) were housed in microisolatorcaging with filter bonnet tops and maintained on sterilized chow and water in a controlled environment. The 18 mice used in the experiment with intact NR-ML-05 weighed 24.04 g f 2.04 g and the 18 mice used in the experiment with the NR-ML-05 Fab weighed 24.61 g f 1.55 g. Tumors were established by injection of lo7A375 Met/Mix2 melanoma cells in the subscapular region of the back. The cells were allowed to grow for 7-10 days to produce tumors which had an average tumor weight of 18 mg f 9 mg for the study involving intact NR-ML-05 and 34 mg f 13 mg for the study involving NR-ML-05 Fab. Biodistribution Studies. In each experiment athymic mice bearing human melanomaA375 Met/Mix tumor xenografts were injected intravenously (ca. 100 pL) via the lateral tail vein. The mice were each marked for identification and were handled in the same manner. Replicate (4X) 5-pL aliquots of the injectate were prepared and counted to serve as standards for the calculation of the total injected dose. Mice were weighed, restrained, and injected;weighingthe syringe before and after injection to determine the volume injected. At designated times postinjection, groups of mice were sacrificed by cervical dislocation and dissected. Blood samples were collected just prior to sacrificing the mouse by obtaining a sample by retroorbital bleeding. Eleven additional tissues were excised from each mouse. These tissues were as follows: tumor, skin (from back), muscle, bone (femur),lung, liver, spleen,stomach (includingcontents),neck (softtissue from the ventral portion of the neck, includingthyroid),kidneys, and intestines (includingcontents). Organs were removed, blotted, weighed, and counted in a y scintillation counter. Initial radioactivity counting was conducted shortly after sacrifice of the animals to obtain combined lZaIand zllAt counts. The tissues were recounted 6-7-days postsacrifice when only lZsIcounts remained. Data analysis was conducted by subtracting the lzaIcounts from the total initial countsto obtain the astatine-211counts,subtracting the background counts from the lZaIcounts, and fiially decay correcting every sample counted. Further analyees of the data included calculation of the percent injected dose per gram (%ID/g)for each isotope in each tissue, an average and standard deviation for the 5% ID/g in each set of six animals,the ratioe of the % ID/g for the two isotopes (zllAt/lZaI) in the tissues of each animal, and the average ratio and standard deviation for each tissue. Three animal biodistributions were performed. In one study, NR-ML-05 IgG labeled with purified 2a and (separately)with purified radioiodinated 2b was coinjected into 18 athymic mice with A375 Met/Mix human tumor xenografts. The NR-ML-05 labeled with 2a used in the experiment had a specific activity of 0.3 pCi/pg and a 99 9% radiochemicalpurity. The radioiodinated 2b labeled NR-ML-05 coinjected had a specific activity of 0.5 pCi/pg and a radiochemical purity of 98%. The injectate contained 7 pCi of astatine-211 and 1.5 pCi of iodine-125 on a total of 24 pg of protein. Groups of six mice were sacrificed at 1, 4, and 22 h postinjection. The tissue distribution data are given in Table 11. A second biodistribution study was conducted by coinjection of NR-ML-05 Fab labeled with purified 2a and (separately) with purified 2b into 18 athymic mice with A375 Met/Mix human tumor xenografts. The astatinated Fab employed had a specific activity of 0.5 pCi/pg and a radiochemical purity of >98%. The radioiodinated NRML-05 Fab employed had a specific activity of 0.5 pCi/pg and a radiochemical purity of >98%. The injectate
pAstatobenroy1 Antibody Labeling
Bloconlugate Chem., Vol. 2, No. 3, 1991 175
Table I. Astatine-211 Distillation Yields. oven temp, OC
estimatedb astatine, pCi
reaction vial
650-700 650-700 650-675 600-625 600 625-675 600-650 600-650 650-700 650-700 600-650
1320 1610 916 1390 1134 1021 1624 1193 682 1250 1644
988 (75) 837 (52) 275 (30) 448 (44) 325 (29) 194 (19) 341 (21) 692 (58) 307 (45) 250 (20) 296 (18)
uCi (94 recovered) fittings/glassware 68 (7) 70 (8) 276 (27) 407 (36) 310 (30) 457 (29) 112 (9) 350 (51) 87 (7) 219 (13)
total recove@ 1064 (81) 912 (69) 371 (41) 740 (73) 763 (68) 515 (51) 825 (51) 865 (72) 669 (98) 349 (28) 531 (32)
0 Values were obtained from Capintec CRC-7 dose calibrator and should be considered estimates. b Values are estimates based on decay correction from end-of-bombardmentof target to time of beginning distillation. The remainder of the activity was present on the distillation glassware which could not be fit into the dose calibrator.
Table 11. Distribution of Radioactivity for Coinjected [211At]2a-and [lMI]2b-LabeledNR-ML-05 in Athymic Mice with A375 Met/Mix Xenografts. l h 4h 22 h tissue 2llAt 1WI 211At 125 '1 211At 1WI 36.78 f 1.75 blood 36.07 f 1.65 27.90 f 1.51 28.45 f 1.69 18.48 f 2.85 21.27 f 3.66 tumor 3.82 f 1.32 4.65 f 0.93 8.63 f 3.03 9.43 f 3.72 19.64 f 6.57 23.62 f 8.36 skin 3.06 f 0.70 3.03 f 0.66 4.20 f 1.18 4.05 f 1.12 5.57 f 1.42 5.62 f 1.43 1.01 f 0.09 muscle 0.97 f 0.10 1.27 f 0.26 1.30 f 0.24 1.89 f 0.35 1.72 f 0.26 2.83 i 3.23 bone 2.86 f 0.52 2.56 f 0.48 2.59 f 0.40 2.81 f 0.69 2.70 f 1.01 lung 13.79 f 3.12 13.54 f 3.23 14.06 f 2.92 13.70 f 3.02 9.23 f 2.36 8.23 f 2.45 5.03 f 1.60 liver 8.54 f 1.47 8.57 f 1.54 6.67 f 0.86 6.69 f 0.92 4.66 f 1.52 spleen 7.64 f 1.83 7.38 f 1.83 6.62 f 2.07 5.75 f 1.77 6.88 f 1.66 5.15 f 1.98 stomach 1.69 f 0.42 1.09 f 0.29 3.09 f 0.51 1.32 f 0.34 4.81 f 0.87 0.92 f 0.17 neck 7.11 f 2.57 7.13 f 2.72 7.28 f 1.45 6.99 f 1.41 7.28 f 2.00 6.67 f 1.93 kidney 7.70 f 0.60 7.95 f 0.64 6.14 f 0.92 6.62 f 0.96 5.08 f 1.05 5.49 f 1.11 intestine 1.34 f 0.14 1.39 f 0.13 1.47 f 0.32 1.38 f 0.36 1.44 f 0.21 1.32 f 0.18 Results were obtained for n = 6 mice and are given as mean f standard deviation of the percent injected dose/g (% ID/g). Table 111. Distribution of Radioactivity for Coinjected [:11At]2a- and [lMI]2b-LabeledNR-ML-05 Fab in Athymic Mice with A375 Met/Mix Xenografts.
6.50 f 0.79 6.26 f 0.72 blood 1.74 f 0.19 1.17 f 0.09 0.44 f 0.06 0.14 f 0.01 5.53 f 0.68 tumor 5.20 f 0.92 7.56 f 1.35 5.68 f 1.05 4.16 f 1.21 3.07 f 0.95 skin 2.80 f 0.53 2.34 f 0.46 2.31 f 0.40 0.96 f 0.15 0.91 f 0.12 0.06 f 0.01 1.18 f 0.19 1.05 f 0.19 muscle 0.67 f 0.13 0.36 f 0.05 0.22 f 0.04 0.02 f 0.01 1.30 f 0.34 1.05 f 0.27 bone 0.98 f 0.13 0.31 f 0.06 0.42 f 0.12 0.02 f 0.02 5.66 f 0.98 4.12 f 0.85 4.52 f 0.57 1.06 f 0.08 2.04 f 0.40 0.18 f 0.12 lung liver 1.90 f 0.24 2.04 f 0.24 0.96 f 0.16 0.57 f 0.09 0.34 f 0.06 0.10 f 0.02 2.63 f 0.56 1.35 f 0.26 spleen 3.65 f 0.77 0.34 f 0.03 2.06 f 0.54 0.06 f 0.02 2.92 f 0.94 0.94 f 0.31 stomach 12.14 f 2.56 0.85 f 0.16 5.44 f 1.55 0.07 f 0.02 neck 3.54 f 0.53 2.86 f 0.43 3.37 f 0.92 0.85 f 0.09 1.40 f 0.59 0.14 f 0.12 34.38 f 6.35 38.56 f 6.84 kidney 9.82 f 2.59 13.92 f 4.14 1.49 f 0.74 0.89 f 0.48 intestine 0.94 f 0.18 0.82 f 0.19 0.89 f 0.13 0.37 f 0.03 0.39 f 0.11 0.04 f 0.01 Results were obtained for n = 6 mice and are given as mean f standard deviation of the percent injected dose/g (%ID/$).
Table IV. Distribution of Astatine-211in Athymic Mice. tissue 2h 4h blood 1.06 f 0.08 1.06 f 0.16 skin 2.90 f 0.35 2.76 f 0.44 muscle 0.87 f 0.05 0.50 f 0.08 bone 1.14 f 0.10 1.05 f 0.16 lung 7.05 f 0.40 5.49 f 0.89 liver 1.10 f 0.13 0.99 f 0.13 spleen 8.95 f 1.69 6.38 f 0.62 stomach 23.42 f 5.53 19.07 f 3.98 neck 4.29 f 0.83 3.22 f 0.26 kidneys 2.65 f 0.19 2.29 f 0.23 intestines 1.45 f 0.14 1.28 f 0.17 Resulta were obtained for n = 4 mice and are given as mean f standard deviation of the percent injected dose/g (%ID/g).
contained 9 pCi of astatine-211 and 4 pCi of iodine-125 on a total of 25 Mg of protein. Groups of six mice were sacrificed at 1, 4, and 20 h postinjection. The tissue distribution data are given in Table 111. A third biodistribution study was carried out to evaluate the in vivo distribution of astatide ion in athymic mice. In that study, eight mice were each injected with 9.3 pCi of [211At]astatidein 100 p L of PBS solution. Sets of four animals were sacrificed at 2 and 4 h postinjection. The tissue distribution data are given in Table IV. RESULTS
Astatine-211 was distilled into the reaction mixture containing the organometallic intermediate 1,the oxidizing agent N-chlorosuccinimide (NCS), and acetic acid in methanol by using the distillation setup shown in Figure 1. It was observed that the best distillation yields were
,
obtained when the oven was preheated to the distilling temperature (or slightly higher) prior to insertion of the irradiated target. A compilation of the recovery of astatine
176 Bioconlugete Chem., Vol. 2, No. 3, 1991
from several distillations is given in Table I. While the overall efficiency of the dry distillations was not high in many of the distillations, the efficiency of trapping the astatine appeared to be very good, as typically only a few percent (1-6 5% ) of the distilled astatine was found in the charcoal trap on the reaction vessel. The reason for the low recovery of astatine in some of the distillations is not fully understood. In the distillations with low recovery of activity, the astatine activity was associated with yellow and black deposits (presumably bismuth) on the glass surfaces of the distilling apparatus. Routinely, the glassware and associated fittings were cleaned prior to use to avoid potential problems with surface contaminants and to clean off the deposits. In addition, use of HF to etch and clear the quartz distilling tube of deposited materials was studied. These procedures did not alleviate the deposition of visible material and astatine activity. Another consideration investigated was the use of oxygen in the carrier gas to oxidize the bismuth and make it less volatile (27). Initial distillations had been conducted under Argon gas (flowing) atmosphere. A premixed quantity of argon/oxygen (11ppm) was studied to try to improve the recovery yields of astatine-211; however, this also did not improve the overall recovery yields. Since there were some concerns about the oxidation of astatine in the presence of oxygen a t the elevated temperatures, all further studies have been conducted under argon atmospheres without having oxygen present. Another factor that may have contributed to the drop in recovery yields was the fact that the initial targets had a bismuth thickness of 0.5 mm, but later ones had a thickness of 0.8 mm. Thickness of target had previously been cited as a possible reason for low recovery of astatine in a dry distillation (24). Rather than pursuing the cause for the low recoveries further, vacuum distillations are presently being conducted which give higher and more consistent yields (28). A number of reactions were carried out to study astatination of the aryltin intermediate 1to produce 2a. The stoichiometry of reagents and solvent compositions used in the astatination reactions were based on optimization studies with the analogous radioiodination reactions (19, 23). The astatination of 1 was found to be facile with radiochemical yields of 2a ranging from 70 % to 90 % .The composition of reaction mixtures were analyzed routinely by reverse-phase HPLC. The gradient employed readily separated starting material 1 from the astatine-labeled materials. The amount of astatinated impurities such as free astatide and p-[211At]astatobenz~icacid observed in the crude reaction mixtures varied, but combined were generally less than 20%. Recovery of injected astatine activity from the HPLC was nearly quantitative (e.g. 92 % 1, indicating that there were no major portions of activity unaccounted for. Concern about the larger quantity of reagents used in the astatination reactions as compared to those in the radioiodination reactions (23)and their potential impact on antibody conjugation led to development of a simple and rapid method of separating astatinated benzoate intermediate 2a (and p-astatobenzoic acid) from impurities. The separation was accomplished with two disposable C-18 columns. The first-18 column was used to separate the nonpolar impurities, including unreacted 1, from the more polar components which included 2a. Immediately following the elution from the first column, the eluant was diluted with PBS and loaded onto the second C-18 column. Under these conditions 2a was quantitatively retained on the column while the most polar impurities, including [211At]astatide,eluted. Removal of
Hadley et ai.
the water from the column and elution of 2a with dry acetonitrile provided a method of isolation in which very little hydrolysis of the N-succinimidyl ester occured. Overall recovery of activity from the columns was 6070%, but was not optimized. HPLC analysis of the purified product showed it to be the desired 2a and a small amount of p-astatobenzoic acid, free of aryltin precursor 1 and most of the [211At]astatide(see Figure 2). Conjugation of 2a to the intact antibody and ita Fab fragment were conducted under identical conditions and resulted in conjugation yields ranging from 40 % to 60 % The yields obtained in the conjugations were consistent with those obtained with the correspondingradioiodinated N-succinimidylp-iodobenzoate (2b) under similar reaction conditions. The nature of the astatinated benzoyl bonding or association with the antibody was of interest. Previous studies with radioiodinated iodobenzoyl conjugates had shown that they were very stable in serum and under a variety of harsh denaturing conditions in vitro (23). Although the denaturing conditions of the TLC (80% MeOH) indicated that the astatine was bonded to the protein in a stable manner, we were concerned that differences in lipophilicity or reactivity might lead to a portion of astatinated 2a (or p-astatobenzoate) being associated with the protein or potentially reacting with functional groups other than lysine amines. To evaluate this, astatinated antibodies (intact IgG and Fab) were subjected to the highly nucleophilic and strongly basic solutions of 1M glycine and 1M Na&Os/NaHCOs at 37 "C for up to 24 h. Upon evaluation, initial TLC purities (97-98%) did not change in PBS (control), under highly denaturing conditions, or with 1 M glycine over an 1824-h p e r i ~ d .A~ small decrease (less than 3%)in purity was observed in the highly basic 1 M Na&Os/NaHCOs mixtures over that time period. Evaluation of the in vivo distribution of [211At]2a-labeled NR-ML-05 and its Fab fragment was conducted in two biodistribution experiments. In the biodistribution studies coinjection of [1251]2b-labeled NR-ML-05 or its Fab fragment, previously shown to be stable in vivo (23),was carried out for comparison purposes. The distribution of zllAt and lZsIactivity when labeled to intact NR-ML-05 were found to be quite similar. The only appreciable difference in the concentrations of radionuclides was in the stomach at 4 and 22 h postinjection. A slight difference was noted in the tumor and blood concentrations at 22 h. The distribution of zllAt and lZ5Iactivity when labeled to NR-ML-05 Fab was similar at 1h postinjection, but large differences in tissue activity were seen a t 4 and 20 h postinjection. Although the immunoreactivities of the astatinated antibodies were not measured due to difficulties with the assay, nearly equivalent localization of activity between the astatinated and radioiodinated antibodies was observed in the tumor, indicating that the preparations were likely to have been immunocompetent. A third animal study was conducted to evaluate the biodistribution of [211At]astatidein athymic mice. A limited study was conducted, with sacrifices a t 2 and 4 h postinjection, for comparison with data from the literature. Concentrations of zllAt appreciably higher than for blood were seen in skin, lung, spleen, stomach, neck, and kidneys.
.
Similar experiments with another IgGab antibody, NR-LU10, and its Fab fragment demonstrated that when astatinated with 28, they were stable in PBS and serum at 37 O C over 18 h. Further, challenges with 95% EtOH, 12% TCA, 6 M urea, and 85 mM NaHSO3 at 37 O C over 18 h did not release astatine from the labeled antibody or its Fab fragment.
pAstatobenroyl Antlbody Labeling
The highest concentrations were seen in the stomach, spleen, lung, and neck. DISCUSSION
The cytotoxic properties of a particle emission of astatine-211 have been known for many years (29,30),as has the potential for tumor-site-selective delivery of different radionuclides by monoclonal antibodies (31-33). However, coupling of astatine to antibodies for use in radioimmunotherapy has presented a formidable challenge as the chemistry of astatine is not fully understood (34, 35). Studies of the chemistry of astatine have been difficult because there are no nuclides of astatine with a half-life greater than 8.3h (36). Introduction of astatine into amino acid functional groups on native proteins has led to attachments which were labile in vivo (7-9). Thus, our group, and other investigators, have studied an alternative method of protein labeling whereby astatine is attached to a molecule which can subsequently be attached (conjugated) to an antibody. The goal of this investigation was to evaluate a method for astatination of the antimelanoma antibody NR-ML05 and its Fab fragment' for subsequent studies of its potential application in therapy of metastatic melanoma. On the basis of our previous studies with p-iodobenzoyl conjugates of antibodies in animals (23,37,38),clinical evaluations (39),and the reported stability of the p-astatobenzoyl conjugate of proteins (12-15), we chose to investigate a method of producing p-astatobenzoyl-labeled antibodies which could be accomplished safely in a minimum time (e.g. 95%) of the activity which was recovered was distilled from the target within the first 30 min. Reaction of the organometallic intermediate 1was found to be facile as predicted by our preliminary studies (19). In the experiments, labeling using the method described herein resulted in good incorporation of astatine (70-905% ) into the protein-reactive intermediate 28,with only small quantities of astatide remaining unreacted. To preclude the reaction of 1with the antibody, the crude astatination (and iodination) product was rapidly separated from 1via a disposable C-18 cartridge elution. Likewise, the remaining astatide was removed from the crude reaction mixture by a second (2-18cartridge separation. The two separation steps using C-18 cartridges were conducted in a manner that they could be used to separate out the undesired components with a minimal loss (
