Technetium labeling of monoclonal antibodies with ... - ACS Publications

Mar 6, 1991 - which the boron substituent (R) was the protein-reactive m-phenyl isothiocyanate (PITC). The. "TcCl(dioxime)3PITC complexes [dioxime =...
0 downloads 0 Views 2MB Size
100

Bloconlugete Chem. 1991, 2, 160-170

Technetium Labeling of Monoclonal Antibodies with Functionalized BATOS. 1, TcCl(DMG)3PITC K. E. Linder, M. D. Wen, D. P. Nowotnik,' M. F. Malley, J. Z. Gougoutas, A. D. Nunn,' and W. C. Eckelman The Bristol-Myers Squibb Pharmaceutical Research Institute, One Squibb Drive, New Brunswick, New Jersey 08903. Received March 6,1991 BATO (boronic acid adduct of technetium dioximes) complexes, TcCl(dioxime)3BR, were prepared in which the boron substituent (R) was the protein-reactive m-phenyl isothiocyanate (PITC). The ggTcCl(dioxime)3PITCcomplexes [dioxime = dimethylglyoxime (DMG) or cyclohexanedione dioxime (CDO)]were prepared from gsTc(dioxime)~(p-OH)SnCl~ and characterized. The X-ray crystal structure of ggTcCl(DMG)3PITC was determined. The -Tc complexes were prepared from -TcO*-in a process using a freeze-dried kit, either in a one-step procedure or via gemTcCl(dioxime)3.Initial labeling studies with 99mTcCl(dioxime)3PITCwere performed on glycine and polylysine and, subsequently, on mouse IgG and the B72.3 monoclonal antibody. Covalent attachment of -TcCl(DMG)3PITC to B72.3 was demonstrated by SDS-PAGE electrophoresis. B72.3 labeled with 99mTcC1(DMG)3PITCdisplayed high binding to a TAG 72 affinity column and had a distribution in normal mice similar to that reported for iodine-labeled B72.3.

INTRODUCTION Due to the ideal properties of technetium-99m for medical imaging, there is wide spread interest in developing techniques for the rapid and specific labeling of monoclonal antibodies with this radionuclide. The attachment of m T c to monoclonal antibodies can be categorized as either direct attachment (where labeling is achieved by chelation of the reduced metal to unmodified or simply modified antibody) or indirect attachment (by way of a bifunctional chelating group). Examples of direct attachment have been reviewed (1). While direct labeling is simple to perform, without modification of the antibody, significant labeling of low-affinity sites does take place ( 2 ) . When technetium is bound to these low-affinity sites, the metal-antibody link can be labile and subject to in vivo breakdown. This lowers target uptake and elevates background levels of radioactivity compared to that seen with a radiolabeled antibody with a stable metal-antibody link. As free thiols have a high affinity for reduced technetium (3), a modification of the direct attachment approach involves the production of free thiols on the monoclonal antibody (or a F(ab')z fragment) by reduction, followed by radiolabeling of the reduced antibody with technetium by transfer of the metal from a labile technetium chelate (4-7). This approach significantly reduces the binding to low-affinity sites, but does not entirely eliminate such binding ( 2 , 8 ) . There are two ways in which bifunctional chelating agents (compounds which comprise of a multidentate ligand and a protein-reactive functional group) can be used in antibody labeling: (i) The ligand is first bound to the antibody, and the purified antibody-ligand conjugate is labeled with technetium. (ii) The technetium complex of the bifunctional chelate is formed initially and then reacted with the antibody. For the first of these approaches, several tetradentate ligands have been utilized. These include derivatives of cyclam (9, IO), the dithiasemicarbazones (DTS) (11,121, the diamide-disulfide ligand (DADS) (13), and the diamine-dithiol (DADT) ligand (14-1 7). Chelating macromolecules,such as metallothionein, have alsobeen attached to antibodies for technetium labeling (18-20). Chelation of technetium to the antibody-ligand complex suffers from 1043-180219 1/29O2-0160S02.50/0

the same problems as the indirect labeling approach; the binding to low-affinity sites is not entirely avoided, and so, some in vivo loss of label may result ( 2 , 8 ) . Previous studies involving the direct attachment of a preformed technetium chelate to the monoclonal antibody have concentrated on the use of three technetium cores: e9mTc(V)ON2Sz,diamido dithiol complexes; -Tc(V)ON3S, triamido thiols (21-23), derivitized with an active ester group for linkage to primary amines on the protein; and -Tc-cyclam complexes ( 9 , 2 4 , 2 5 )derivatized with a maleimide or vinylpyridine, for conjugation to free thiols (obtained by derivatization of the protein with 2-iminothiolane). While indirect labeling generally involves greater manipulation (purification steps) of the radiolabeled protein than is the case for the direct labeling methods, it has the advantage that the radiolabel exista in a known chemical form, with no low-affinity binding of technetium. This report describes studies of indirect protein labeling utilizing a functionalized derivative of the seven-coordinateTc(II1) BATO (boronic acid adduct of technetium dioximes) class of technetium complexes (26). The BATO complexes (generalstructure shown in Figure 1)are neutral, lipophilic complexes that are stable; i.e. the technetium atom cannot readily be removed by challange from other ligands. BATOs are formed by a template reaction involving a vicinal dioxime, a boronic acid, pertechnetate, and stannous ion (asreductant) at low pH (27). The versatility of this reaction has permitted the synthesis of a large number of BATOs (28). From this series, two compounds were selected for clinical evaluation, as myocardial and cerebral perfusion tracers (29,30).By using a boronic acid derivatized with a protein-reactive functional group, we felt that we could extend the application of BATO complexes to protein labeling. A suitable protein-reactive functionalgroup is isothiocyanate (R-N=C=S). Isothiocyanates react with free primary amine groups to form a stable thiourea link. Others have used isothiocyanates to link modified DTPA ligands to monoclonal antibodies (31). In this paper, we describe the synthesis and characterization of BATOs which contain an isothiocyanato functional group. The use of these bifunctional chelates in labeling amine-containing bio0 1991 American Chemical Society

Technetium Labeling of Monoclonal Antibodles

R

R

c1

Figure 1. The generalstructure of BATOs. Technetium is bound to six nitrogen atoms, from three vicinal dioximes, and a monodentate anion ((21- in this example). One end of the BAT0 is capped by a boronic acid. In this paper, the protein-reactive isothiocyanate group is part of the R' substitutent.

molecules and preliminary studies in animals of the monoclonal antibody B72.3 labeled with one of these reagents are also described. EXPERIMENTAL PROCEDURES

Materials and Reagents. Dimethylglyoxime (DMG) (Eastern Chemicals), cyclohexanedione dioxime (CDO), and stannous chloride (SnC12) (MCB) were used as received. Polylysine (MW 102 000) bovine serum albumin fraction V (BSA), mouse IgG, and rat serum were purchased from Sigma. Sephadex G-25 (medium grade) was purchased from Pharmacia. B72.3 monoclonal antibody (2.81 mg/mL) was purchased from Damon Biotech. PRP-1 resin (12-20 pm) was purchased from Alltech Associates. 3-Isothiocyanatophenylboronic acid (PITC) was synthesized in our laboratories by reaction of 3-aminophenylboronic acid and thiophosgene; details are reported elsewhere (32). Ammonium pertechnetate (NH4Tc04)was purchased from Oak Ridge National Laboratories and recrystallized from dilute aqueous hydrogen peroxide. The complexes Tc(dioxime)3(p-OH)SnCl3 (dioxime = DMG, CDO) were prepared as described previously (27). ["Tclpertechnetate (-TC04-) was obtained in a saline solution from a Squibb Minitec Generator. All other chemicals were of reagent grade and were used as received. All solvents were of either HPLC or reagent grade (except acetonitrile, which was pesticide grade) and were used as received. All water used in these studies was obtained from a Millipore Milli-Q water-purification unit. Most reagents and standards used in the SDS-polyacrylamide gel electrophoresis were purchased from BioRad. Phast gel blue, 10-15% gradient gel, and SDS buffer strips were purchased from Pharmacia. The following buffers were prepared: 0.1 M ammonium acetate (pH 4.6), phosphate-buffered saline [PBS; 0.01 M sodium phosphate buffer (pH 6.0) and 0.15 M of NaCl], and 0.1 M sodium phosphate buffer (pH 9.5). The 6 M guanidine was prepared in water and filtered through a 0.2-pm membrane. Buffers and solutions used in Phast gel electrophoresis were prepared according to the directions in the Phast system manual. A TAG-72-Reacti-Gel affinity column was prepared inhouse (33). Freeze-dried "kits" were prepared for the formation of -TcCl(DMG)3 from WmTcO4-. Each kit consisted of a sealed 5-mL siliconized serum vial containing DMG (12.9 pmol), NaCl (1.7 mmol), citric acid (52 pmol), diethylenetriaminepentaaceticacid (DTPA, 2.5 pmol), SnCl2 (0.22 pmol), and y-cyclodextrin (38.5pmol). Immediately prior to freeze-drying, the solution pH was adjusted to 2.0 with HC1, and after freeze-drying, the vials were filled with

Bioconjugate Chetn., Vol. 2, No. 3, lQQl

181

nitrogen gas and sealed. Kits in which CDO (10.5 pmol) was substituted for DMG and pH was adjusted to 3.75 were also prepared as described above. Analysis of the Labeling Reagents. Silica gel TLC plates (Bakerflex 1B) were developed with CH2C12 or CH3CN. HPLC separations were performed on a Spectra Physics Model SP8700XR HPLC system equipped with SP4270 integrator, spectrophotometer (Kratos, SF769Z), radiometric detector (Tennelec minibin), dual-pen recorder, and fraction collector. The outputs from the radiation detector and spectrophotometer (set at 280 nm) were integrated via the SP4270 integrator. The data were stored and reprocessed with LABNET software. 99Tc-and *Tc-labeling reagents were analyzed on a 30-cm Lichrosorb RP-18 column (E.Merck) using a mobile phase of acetonitrile/O.l M ammonium acetate pH 4.6 (75/25, v/v) a t a flow rate of 1mL/min. The radiochemical purities (RCP)of the *Tc complexes were determined as lOOX the radioactivity associated with the complex peak, divided by the total radioactivity eluted from the column. gsTc complexes were monitored by UV detection at 250 nm. Proteins were monitored at 280 nm. Infrared spectra were obtained on a Sirius 100 FT-IR spectrometer as KBr pellets. Proton NMR data were obtained on a JEOL GX 250 spectrometer. UV-visible spectra were measured on a Hewlett-Packard 8451A photodiode-array spectrophotometer. Fast-atom-bombardment mass spectra were obtained on a VG-ZAB-2F spectrometer, from a matrix of thioglycerol/CHsCN. Elemental analyses were performed in-house by Squibb Microanalytical Department. Synthesis and Characterizationof 99Tc Complexes. Synthesis of wTcCZ(CDO)#ITC from Tc(CDO)3(p-OH)SnCl3. A solution containing Tc(CDO)3(p-OH)SnCb3HzO (77 mg, 0.095 mmol), 3-isothiocyanatophenylboronicacid (22 mg, 0.12 mmol), EtOH (10 mL), and 1M HCl(2 mL) was boiled in a small beaker for 1h. During this time, the color of the solution changed from yellow-orange to deep orange-red and the volume reduced to 90% RCP by a two-step purification process involving initial purification on a G-25 column (which removed most of the low molecular weight byproducts) followed by TSK HPLC purification. In labeling B72.3, it was found that the G-25/TSK purification protocol failed to remove labeled high molecular weight protein aggregates. This problem was eliminated by substituting ISRP chromatography for G-25 as the initial purification step. Covalent attachment of the labeling reagent to B72.3 monoclonal antibody was demonstrated by Phast SDSPAGE. Under nonreducingconditions, 84% of the activity was associated with the antibody monomer. The stain/ destain protocol (used to visualize proteins on the gel) removed most of the unbound activity, giving an apparent RCP of 94%. Protein reduction with 2-mercaptoethanol was used to fragment the antibody into heavy and light chains, SDS-PAGE followingthis treatment indicated that the label was randomly distributed on the whole antibody, as the distribution of radioactivity between heavy and light chains is similar to the molecular weight ratio of these fragments.' Protein reduction, by harsh treatment with 2-mercaptoethanol, did result in sigificant loss (45% ) of the label from antibody, presumably due to transfer of technetium by a ligand-exchange process. However, under milder challenge conditions (incubation in serum), technetium remained bound to B72.3. The preliminary examination of the biodistributions of mTcCl(DMG)3PITC and its B72.3 conjugate in normal mice also provide some demonstration that a stable link exists between the label and the antibody. The biodistribution of the labeling reagent was very different from that of the radiolabeled antibody. The labeling reagent showed relatively low blood values at 1 h, with rapid clearance to kidneys and to the GI tract. In contrast, the radiolabeled antibody showed, as expected, prolonged retention in blood. In addition, low levels of radioactivity in liver and spleen suggested that very few high molecular protein aggregates had formed in the -Tc-labeled B72.3 preparation. Comparing our biodistribution data on -Tc-labeled B72.3 to that of 12Wabeled B72.3 and lllIn-labeled B72.3 reported by Sands et a1 (42), we observed that -TeB72.3 demonstrated lower spleen and liver uptake than 1261-B72.3. Liver uptake was distinctly lower with 99mTc-B72.3 than with Tn-B72.3. The immunoreactivity (77%) of B72.3 labeled with -TcCl(DMG)3PITC, measured with the TAG 72 column, is high, particularly when compared to 1261-labeledB72.3 ~~~

A distribution of radioactivity in proportion to chain molecularweight is seen in unstained gels only. A greater proportion of radioactivity is associated with the heavy chain after the stain/ destain process, suggesting preferential loss of label from the light chain during this process. The reason for the selective loss of radiolabel is unclear at this time.

(5540 76 1. This indicates that the antibody retains most of its immunoreactivity after the labeling and purififcation processes. These preliminary data suggest that the labeling reagent gsmT~C1(DMG)3PITC will form a stable conjugate with proteins, with little loss of immunoreactivity. However, the overall yield of the labeling process (starting from mT~04-)is low, and the process requires two purification steps. This, in part, can be attributed to the highlipophilicity, and some improvements in labeling efficiency and labeling convenience could be achieved by using analogues of lower lipophilicity. However, we have demonstrated that BATOs are suitable technetium chelates for the radiolabeling of monoclonal antibodies. ACKNOWLEDGMENT

We thank Lillian Belnavis, Maryann Homack, and Debbie Silva for technical assistance with the biodistribution studies. Supplementary Material Available: Tables listing final atomic positional parameters, thermal parameters, bond distances, and bond angles for TcCl(DMG)sPITC (7 pages); a table of structure factors (11pages). Ordering information is given on any current masthead page. LITERATURE CITED (1) Eckelman, W. C., and Paik, C. H. (1986) Comparison of Tc99m and In-111labeling of conjugated antibodies. Nucl. Med. Biol. 13, 335-343. (2) Paik, C. H., Eckelman, W. C., and Reba, R. C. (1986) Transchelation of ~ Tfrom c low affinity to high affinity sites of antibody. Nucl. Med. Biol. 13, 359-362. (3) Deutach, E., Libson, K., Jurisson, S.,andLindoy, L. F. (1983) Technetium Chemistry and Technetium Radiopharmaceuticals. B o g . Inorg. Chem. 30, 75-139. (4) Reno, J. M., and Bottino, B. J. (1987) Improved radionuclide antibody coupling. Eur. Pat. Appl. 87300426.1. (5) Bremer, K. H., Kuhlman, L., Schwartz, A,, and Steinstrasser, A. (1987)Studies to discover an organ specific substance labeled with technetium-99m. Eur. Pat. Appl. 87118142.6. (6) Jones, A. L., Kinz, A., Grebenau, R., et al. (1990) Radiolabeling of MonoclonalAntibody (MAb) Fragments with Technetium-99m (Tc-99m) Using a One Vial Kit. J. Nucl. Med. 31,905. (7) Pak, K. Y., Nedelman, N. A.,Tam, S. H., et al. (1990)Stability and Immunoreactivity of Technetium-99m Antibody Fragmenta by a Direct Labeling Method. J. Nucl. Med. 31,905. (8) Fritzberg, A. R., Abrams, P. G., Beaumier, P. L., et al. (1988) Specific and stable labeling of antibodies with technetium99m with a diamide dithiolate chelating agent. Roc. Natl. Acad. Sci. 85,4025-4029. (9) Franz, J., Volkert, W. A., Barefield, E. K., and Holmes, R. A. (1987) The production of Tc-99m labeled conjugated antibodies using a cyclam based bifunctional chelating agent. Nucl. Med. Biol. 26, 293-299. (10) Morphy, J. R., Parker, D., Alexander, R., et al. (1988) Antibody Labelling with Functionalised Cyclam Macrocycles. J. Chem. SOC. Chem. Commun. 156-158. (11) Arano,Y.,Yokoyama,A.,Magata,Y.,etal. (1986) Synthesis and Evaluation of a New Bifunctional Chelating Agent for h T c Labeling Proteins: p-Carboxyethylphenylglyoxaldi(Nmethylthiosemicarbazone). Nucl. Med. Biol. 12, 425-430. (12) Arano, Y., Yokoyama, A., Furukawa, H., et al. (1987) Technetium-99m-Labeled Monoclonal Antibodywith PreservedImmunoreactivity and High in Vivo Stability. J.Nucl. Med. 28, 1027-1033. (13) Nicolotti, R. A., and Dean, R.T. (1987) Coupling agenta for radiolabeling proteins. Eur. Pat. Appl. 87304716.1.

170 Bioconjugete Chem.,Vol. 2, No. 3, 1991

(14) Byrne, E. F., and Tolman, G. L. (1983) Technetium-99m bifunctional chelating agent-thiolactone coupling to biomolecules. N2Sz ligand for chelation to technetium. J.Nucl. Med. 24, P126. (15) Lever, S. Z., Baidoo, K. E., Kramer, A. V., et al. (1988) Synthesis of a novel bifunctional chelate designed for labeling proteins with technetium-99m. Tetrahedron Lett. 29,32193222. (16) Misra, H.K., Virzi, F., Hnatowich, D. J., and Wright, G. (1989)Synthesis of a Novel DiaminodithiolLigand for Labeling Proteins and Small Molecules with Technetium-99m. Tetrahedron Lett. 30, 1885-1888. (17) Baidoo, K. E., and Lever, S. Z. (1990) Synthesis of a Diaminedithiol Bifunctional Chelating Agent for Incorporation of Technetium-99m into Biomolecules. Bioconjugate Chem. 1, 132-137. (18) Morelock, M. M., and Tolman, G. L. (1987) Metallothionein: A Bifunctional Chelator for the Radiolabeling of Biologically Active Molecules. Erperientia 52 (Supplement: Metallothionein ZZ), 247-253. (19) Morelock, M. M., Cormier, T. A., and Tolman, G. L. (1988) Technetium Metallothioneins. Inorg. Chem. 27,3137-3140. (20) Brown,B. A.,Drozynski, C. A.,Dearborn, C.A., et al. (1988) Conjugation of Metallothionein to a Murine Monoclonal Antibody. Anal. Biochem. 172, 22-28. (21) Fritzberg, A. R. (1987) Advances in mTc-labeling of antibodies. Nuklearmeditin 26, 7-12. (22) Fritzberg, A. R., Kasina, S., Reno, J. M., et al. (1986) Radiolabeling of antibodies with -Tc using N2S2 ligands. J. Nucl. Med. 27,957-958. (23) Wilkening, D., Srinivasan, A., Kasina, S., et al. (1988) Tc99m Antibody Labeling with NaS and N2S2 Amide Mercaptides: Active Ester Complex Yield and Side Chain Length. J. Nucl. Med. 29, 815. (24) Parker, D., and Millican, T. (1987) Conjugate Compound. Int. Pat. Appl. PCT/GB87/00106. (25) Morphy, J. R., Parker, D., et al. (1988) Antibody labeling with Functionalised Cyclam Molecules. J.Chem. Soc. Chem. Commun. 156-158. (26) Treher, E. N., Francesconi, L. C., Gougoutas, J. Z., et al. (1989) Monocapped Tris(dioxime)Complexesof Technetium(111): Synthesis and Structural Characterization of TcX(dioxime)SBR (X = C1, Br; Dioxime = Dimethylglyoxime, Cyclohexanedione Dioxime; R = CHs, CdHe). Inorg. Chem. 28, 3411-3416. (27) Linder, K. E., Malley, M. F., Gougoutas, J. Z., et al. (1990) Neutral, Seven-Coordinate Dioxime Complexes of Technetium(II1): Synthesis and Characterization. Znorg. Chem. 29, 24262435. (28) Nunn, A. D., Feld, T., and Narra, R, K. (1990) Quantitative Structure Distribution Relationships (QSDRs)of BATOs. In

Llnder et ai.

Technetium and Rhenium in Chemistry and Nuclear Medicine (M. Nicholini, G. Bandoli, and U. Mazzi, Eds.) Vol. 3, pp 399-404, Cortina International, Verona, Italy. (29) Zielonka, J. S., Cannon, P., Johnson, L., et al. (1990) Multicenter Trial of mTc-Teboroxime (Cardiotec): A New Myocardial Perfusion Agent. Nuklear Medizine (Suppl.26) 211212. (30) Narra, R. K., Nunn, A. D., Kuczynski, B. L., et al. (1990) A Neutral Lipophilic Tc-99m Complex for Regional Cerebral Blood Flow Imaging. J. Nucl. Med. 31, 1370-1377. (31) Mirzadeh, S., Brechbiel, M. W., Atcher, R. W., and Gansow, 0. A. (1990) Radiometal Labeling of Immunoproteins: Covalent Linkage of 2- (4-Isothiocyanatobenzyl)diethylenetriaminepentaacetic Acid Ligands to Immunoglobulin. Bioconjugate Chem. 1 , 5 9 4 5 (and references therein). (32) Ramalingam, K., Nowotnik, D. P., and Nunn, A. D. The synthesis of some isothiocyanato phenylboronic acids. Org. Prep. Proc., submitted for publication. (33) Hagan, J., Croze, E., and Yost, F., unpublished data. (34) Walker, N., and Stuart, D. (1983) An empirical method for correcting diffractometer data for absorption effects. Acta Crystallogr., Sect. A 39, 158-166. (35) SDP. Structure Determination Package. A. Frenz and Associates, College Station, TX. (36) Little, R. J., and Donahue, M. A. (1968)Spectral properties of proteins and small molecules of immunological interest. Methods Immunol. Immunochem. 2,343-364. (37) Johnson, C. K. (1965) ORTEP. Report ORNL-3794. Oak Ridge National Laboratory, Oak Ridge, TN. (38) Linder, K. E., Jurisson, S., Francesconi, L., et al. (1990) Template Synthesis of Technetium and Rhenium Dioxime Complexes: The Mechanism of Formation of TcX(Dioxime)aBR from Pertechnetate. In Technetium and Rhenium in Chemistry and Nuclear Medicine (M. Nicholini, G. Bandoli, and U. Mazzi, Eds.) Vol. 3, pp 195-199. (39) Linder, K. E., Malley, M. F., Nowotnik, D. P., et al. An unexpected by-product obtained during the preparation of technetium(II1) boronic acid adducts of dioximes. The single crystal structure of TcCl(DMG)2(BDI)BEt (DMG = dimethylglyoxime,BDI = butane-2,3-dione imine-oxime). Znorg. Chim. Acta, submitted for publication. (40) Bergquist, N. R., and Nilsson, P. (1974) The conjugation of immunoglobulinswith tetramethylrhodamine isothiocyanats by utilization of dimethylsulfoxide (DMSO) as solvent. J. Zmmunol. Method 5,189-198. (41) Jurisson, S. S., Hirth, W., Linder, K. E., e t al. ChlorvHydroxy Substitution on Technetium BAT0 (TcCl(Dioxime)&R) Complexes. Nucl. Med. Biol., in press. (42) Sands, H., and Jones, P. L. (1987) Methods for the Study of the Metabolism of Radiolabeled MonoclonalAntibodies by Liver and Tumor. J. Nucl. Med 28, 390-398.