Further Studies on the Protein Conjugation of ... - ACS Publications

Departments of Radiation Oncology, Medicine, and Pharmacology, and Comprehensive Cancer Center,. University of Alabama at Birmingham, Birmingham, ...
0 downloads 0 Views 87KB Size
18

Bioconjugate Chem. 1999, 10, 18−23

Further Studies on the Protein Conjugation of Hydroxamic Acid Bifunctional Chelating Agents: Group-Specific Conjugation at Two Different Loci Ahmad Safavy,*,† M. B. Khazaeli,‡ Marion Kirk,§ Lori Coward,§ and Donald J. Buchsbaum† Departments of Radiation Oncology, Medicine, and Pharmacology, and Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham, Alabama 35294. Received April 30, 1998; Revised Manuscript Received October 9, 1998

A procedure utilizing an activated ester approach for conjugation of unprotected hydroxamic acids to antibodies and peptides was recently reported. Here, an alternative method with advantages over the activated ester strategy is described. This protocol utilizes the hydrazone formation between a hydrazide derivative of the trihydroxamate ligand trisuccin and either a ketone derivative of antibody or the aldehyde groups, generated by oxidation of the carbohyrate residues. Thus, the trisuccin carboxylic acid (1) was derivatized with tert-butyl carbazate to the hydrazide 2, and the protecting groups were removed by catalytic hydrogenation and acidolysis with TFA to afford the hydroxamate hydrazide 4. Conjugation of 4 to monoclonal antibody CC49 was effected by two approaches: attachment through the amine (e.g., lysine) residues of the antibody or oxidation of the carbohydrate residues. The extent of conjugations were monitored by MALDI, through evaluation of the increases in molecular weights of the conjugates compared to the unconjugated antibody. The first approach utilizing a ketone linker (6-oxoheptanoic acid, OHA) which served as a hydrazide anchor, is being introduced in this report as a new technique for conjugation of hydrazide derivatives to proteins. The OHA approach proved to be a superior strategy over the aldehyde approach in the ease of the procedure and yield of protein recovery. It also had the advantage of yielding more control in adjusting the ligand-to-protein ratio and was therefore selected for protocol optimization. All conjugates resulting from both approaches were radiolabeled with 125I and screened for their immunoreactivity. Furthermore, the conjugates prepared through the optimized OHA protocol were radiolabeled with both 99m Tc and 125I for which the radiolabeling yields and immunoreactivities are reported.

INTRODUCTION

Radiometal labeling of monoclonal antibodies (MAbs),1 antibody fragments, and peptides has found a wide range of applicability in biology and medicine (1-6). A frequently used technique is the use of a heterobifunctional chelating agent which serves as a bridge between the metallic cation and the protein. Two requirements for such a linker to qualify as a potential labeling agent are (1) it should form metal chelation with high in vitro and in vivo stability, and (2) the linker as a whole should be biologically stable. Therefore, in addition to the metal complex stability, the overall usefulness of the labeled conjugate is dependent on the stability of the linker* To whom correspondence should be addressed at the Department of Radiation Oncology, 1824 sixth Ave. S., WTI 674, Birmingham, AL 35294. † Department of Radiation Oncology. ‡ Department of Medicine. § Department of Pharmacology. 1 Abbreviations: BCA, bifunctional chelating agent; cpm, counts per minute; DCC, dicyclohexyl carbodiimide; DMAP, N,N-(dimethylamino)pyridine; DPBS, Dulbecco’s PBS; HOBT, N-hydroxy benzotriazole; HSA, human serum albumin; ITLC, instant thin-layer chromatography; L:MAb, ligand-to-antibody ratio; MAb, monoclonal antibody; MALDI, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; OHANS, 6-oxoheptanoic acid, N-succinimidyl ester; ONP, o-nitrophenyl; PBS, phosphate-buffered saline; RY, radiolabeling yield; SD, standard deviation; SEC, size-exclusion chromatography; SOX, sugar-oxidation.

protein bond (7-9). Premature cleavage of this bond may result in the release of the radionuclide before the construct reaches its intended target and will result in an insufficient target radioactivity dose for imaging or therapy, and an increase in background uptake in normal tissues. In an effort to design protein radiolabeling reagents capable of meeting these two requirements, we reported on the application of hydroxamate-derived bifunctional chelating agents (BCAs) to radiometal labeling of antibodies and peptides (1, 10). Although the metal chelation of the hydroxamic acids had been well established, their utilization as BCAs was a new concept which turned out to be a challenging problem, due to the interference of the hydroxamate functionalities with the conjugation chemistries used for their attachment to a protein molecule. Our original procedure connected an O-benzyl-protected trihydroxamate, trisuccin, to MAbs, followed by removal of the benzyl groups by hydrogenation of the conjugate (10). Recently, we reported on an improved technique using the o-nitrophenol (ONP) activated ester of unprotected trisuccin (11). In this procedure, the fully deprotected hydroxamate was activated by ONP in the presence of a carbodiimide and used for the antibody conjugation without isolation. Due to the importance of the conjugation step and its effect on the subsequent application of the constructs, we continued our search for still better techniques. Our ultimate goal was a protocol in which stable derivatives of hydroxamic acids are conveniently and efficiently conjugated to the protein via a covalent bond which is stable under most

10.1021/bc980045d CCC: $18.00 © 1999 American Chemical Society Published on Web 12/15/1998

Hydroxamate Hydrazide Protein/Peptide Conjugation

chemical and physiological conditions. Also of importance are the selectivity and site-targeting involved in the conjugation strategy. Electrophilic covalent coupling of BCAs to antibodies has traditionally been targeted at the amino terminal ends of polypeptide chains and -amino functions of the lysine residues, with the latter probably being the major site of reaction (12). A mechanistically different procedure for antibody conjugation of BCAs is utilization of the oxidized carbohydrate residues through Schiff base formation (13, 14). It has been postulated that derivatization of antibodies through this route would result in formation of conjugates with more preserved biological activities (15). The latter procedure was originally used for the attachment of horseradish peroxidase to an immunoglobulin (16) and was later extended to the conjugation of monoclonal antibodies (12, 15). We report here two new protocols for the attachment of unprotected hydroxamate hydrazides targeted at two different sites of the MAb CC49, the lysine residues residing across the immunoglobulin molecule and the carbohydrate moieties of the hinge region. Although, in the case of this antibody, our results favor the former method, the latter protocol may also be useful for other glycoproteins and molecules containing oxidizable carbohydrate residues or suitable carbonyl functions. This, in turn, will have the benefit of reaction manipulation in tailoring a desired conjugation and more control on the reaction outcome. EXPERIMENTAL PROCEDURES

General. The benzyl-protected trisuccin, 1, was prepared as described previously (10). HPLC analyses were carried out on a Bio-Rad model 5000 Titanium system (Bio-Rad, Richmond, CA) equipped with model 1806 UVvis detector and Beckman model 170 radioisotope detector, operated by ValueChrome software (Bio-Rad). For analytical SEC, a 7.5 mm × 25 cm G3000SW column (TosoHaas, Montgomeryville, PA) was used. PBS (10 mM) containing 10 mM Na2SO4 at pH 6.7 was used as solvent. Preparative SEC was carried out using sephadex G-25 columns (PD-10, Pharmacia Biotech AB, Uppsala, Sweden) eluted with DPBS. Mass spectra were run on an API III triple quadrupole mass spectrometer (PESciex, Toronto, Ontario) in electrospray mode and PerSeptive Biosystems (Framingham, MA) Voyager Elite MALDI instruments. Microanalyses were performed by Atlantic Microlab (Atlanta, GA). N-Succinimidyl 6-oxoheptanoate was provided by Biolinx, Inc. ([email protected]). Metal-free purified water was obtained from a Milli-QF system (Millipore, Bedford, MA). Sodium [99mTc]-pertechnetate was purchased from Syncor (Birmingham, AL). ITLC was carried out on silica gel-impregnated glass fiber slides (Gelman Sciences, Ann Arbor, MI). N-[Tris[2-[[N-(benzyloxy)amino]carbonyl]ethyl]methyl]succinic N′-(tert-butyloxycarbonyl)hydrazide, 2. To the solution of the trisuccin carboxylic acid 1 (500 mg, 0.75 mmol) in 10 mL of dry THF was added HOBT (102 mg, 0.75 mmol) and tert-butyl carbazate (Aldrich, 100 mg, 0.75 mmol). The mixture was cooled in an ice bath under argon and a solution of DCC (171 mg, 0.83 mmol) was added in small portions. The reaction was allowed to reach room temperature with stirring for 18 h. The solid was separated by filtration and the filtrate was concentrated in vacuo to an oily residue. The oil was redissolved in 50 mL of ethyl acetate, and the solution was washed with 10% citric acid, water, 10% NaHCO3, and water. After drying over Na2SO4, the solvent was distilled in vacuo and the residue was

Bioconjugate Chem., Vol. 10, No. 1, 1999 19

charged into a silica gel column and eluted with chloroform and a 0 to 5% methanol gradient. Pooled fractions containing the product were distilled in vacuo and triturated with ether to afford 535 mg (92%) of a white powder. An analytical sample was prepared by crystallization from toluene/2-propanol (crystalline needles): mp 99.8-101.2 °C; 1H NMR (DMSO-d6) δ 1.38 (s, 9H), 1.8 (m, 12H), 2.30 (m, 4H), 4.75 (s, 6H), 7.23 (bs, 1H), 7.35 (m, 15H), 8.65 (bs, 1H), 9.50 (bs, 1H), 10.95 (bs, 3H). Anal. calcd for C40H52N6O10: C, 61.84%; H, 6.75%; N, 10.82%. Found: C, 61.80%; H, 6.77%; N, 10.85%; ESMS (M + H) 777. N-[Tris[2-[(N-hydroxyamino)carbonyl]ethyl]methyl]succinic N′-(tert-butyloxycarbonyl)hydrazide, 3. The hydrogenation procedure reported previously was used (11). Briefly, the protected hydroxamate 2 (500 mg, 0.71 mmol) was dissolved in 100 mL of HPLC-grade methanol and a slurry of 10% Pd/C (41 mg, 0.039 mmol) in methanol was added. The sealed flask was evacuated for 10 min with stirring and was charged with hydrogen gas at 10 psi, and the mixture was shaken for 65 h. The solvent was partially removed in vacuo to 10% of its original volume, the catalyst was removed by an Acrodisc syringe-tip 0.2 µm filter (Gelman Sciences, Ann Arbor, MI) and the solvent was distilled in vacuo. The light yellow glass was recrystallized from 2-propanol to obtain 310 mg (95%) of a white powder. mp 91.2-91.9 °C (dec). 1H NMR (DMSO-d ) δ 1.40 (s, 9H), 1.82 (m, 12H), 2.26 6 (m, 4H), 7.74 (bs, 1H), 8.64 (bs, 3H), 9.50 (bs, 1H), 10.30 (bs, 1H). Anal. Calcd for C19H34N6O10: C, 45.05%; H, 6.77%; N, 16.59%. Found: C, 45.15%; H, 6.79%; N, 16.60%. ESMS (M + H) 507. N-[Tris[2-[[N-(hydroxy)amino]carbonyl]ethyl]methyl]succinic hydrazide, 4. Five milliliters of a mixture of trifluoroacetic acid, thioanisole, 1,2-ethanedithiol, and anisole (90:5:3:2, v/v) was added to the blocked hydrazide 3 (300 mg, 59 mmol) under argon and the resulting solution was allowed to stand at room temperature for 1 h. The solvent was distilled in vacuo and the solid residue was washed with hot 2-propanol to afford an off-white crystalline solid. The solid was dried in a vacuum over P2O5 to obtain 222 mg (92.5%) of the product. mp 79.3-82.6 °C (dec). 1HNMR (DMSO-d6) δ 1.78 (m, 12H), 2.32 (m, 4H), 7.28 (bs, 1H), 8.75 (bs, 2H), 10.30 (m, 4H). ESMS (M + H) 406. Antibody Conjugation. Method A. The solution of the MAb CC49, reactive with the TAG-72 antigen expressed in carcinomas (17) (50 µL, 7 mg/mL), in DPBS was added to 300 µL of a 50 mM PBS buffer at pH 8.1. The resulting solution was stirred at 0 °C and the solution of OHA-NS (5.7 µg, 23.5 nmol) in 50 µL of DMF was added. After 90 min at 0 °C, the reaction was quenched by addition of 20 µL of 2 M glycine in the PBS buffer and the mixture was purified on a PD-10 column eluted with a 100 mM acetate buffer, pH 5.5. The elution was monitored by a UV flow-through detector, and the single MAb peak was collected in a 2 mL fraction. This was concentrated to 300 µL in a Centricon-50 concentrator (Amicon, Inc., Beverly, MA). To this solution, at room temperature and with stirring, was added the solution of the hydrazide 4 (0.12-1.2 µmol) in water (5-20 µL). The reaction mixture was stirred at room temperature for 90 min at which time 20 µL of NaCNBH3 solution in water was added to a final concentration of 100 mM in NaCNBH3. The resulting mixture was stirred gently at room temperature for 18 h followed by purification of the conjugate on a PD-10 column eluted with DPBS. The fraction containing the conjugate was concentrated to 300 µL as described above.

20 Bioconjugate Chem., Vol. 10, No. 1, 1999

Method B. The solution of the MAb CC49 (50 µL, 7 mg/mL) in DPBS was added to 400 µL of 100 mM acetate buffer at pH 4.2 containing 0.3% EDTA. To this solution was added a solution of sodium periodate in water to a final molar ratio of 1000 with respect to the MAb (32 mM in NaIO4). The mixture was stirred at 5 °C in the dark for 1 h followed by purification of the oxidized MAb on a PD-10 column eluted with 50 mM acetate buffer, pH 4.2. The MAb-containing fraction was concentrated to 200 µL in a Centricon-50 concentrator. A solution of the hydrazide 4 (0.12-1.2 µmol) in 200 mM acetate buffer at pH 5.5 was added to the MAb solution at room temperature with stirring, followed after 90 min by the addition of 15 µL of NaCNBH3 in water to a final concentration of 100 mM in NaCNBH3. The solution was stirred at room temperature for 18 h, and the conjugate was purified by gel filtration on a PD-10 column eluted with DPBS. MALDI Analyses. Samples were analyzed in the positive mode on a Voyager Elite mass spectrometer with delayed extraction technology (PerSeptive Biosystems, Framingham, MA). The acceleration voltage was set at 25 kV, and 50-100 laser shots were summed. Sinapinic acid (Aldrich Chemical Co., Milwaukee, WI), dissolved in a 1:1 (v/v) mixture of acetonitrile:0.1% TFA, was used as the matrix. Bovine serum albumin (1 pmol/µL) was added as an internal standard. Equal volumes of sample and matrix were mixed on a smooth plate. The average of three measurements was used. For each reaction, the intact MAb was scanned as the control for evaluation of the following conjugation. In the OHA method, the molecular weight increase of the OHA-MAb intermediate was calculated using the unconjugated control. The molecular weight of this intermediate then served as the control to measure the increase in the molecular weight of the final product conjugate. Each difference in molecular weight was divided by the molecular weight of the corresponding linker to measure the linker:MAb ratio. For the SOX method, with no intermediate conjugate, the molecular weight difference between the final conjugate and the starting intact MAb was used. Technetium-99m Radiolabeling. Sodium [99mTc]pertechnetate (5 mCi in saline) was added to the solution of the MAb conjugate in DPBS. Under oxygen-free conditions, 25 µL of a 2 mg/mL solution of stannous chloride, containing 40 mg/mL of 2-hydroxy-2-methyl propionic acid, was added, and the mixture was incubated at 35 °C for 45 min. The labeled conjugate was then purified on a PD-10 column eluted with DPBS containing 0.1% HSA. Iodine-125 Radiolabeling. A standard iodogen radiolabeling method was used (18). Briefly, 10 mg of iodogen was dissolved in 1 mL of chloroform, and 5 µL of this solution was dispensed into a vial containing a polystyrene bead. The bead was then dried under an argon flow. Sodium [125I]-periodate (5 mCi/bead) was added, followed by 5× (v/v) of a pH 7.6 phosphate buffer. The MAb was added, the mixture was incubated for 15 min, and the labeled MAb was purified on a PD-10 column eluted with 0.1% HSA in DPBS. Immunoreactivity Screening. Polystyrene beads (6.4 mm, specular finish) were coated with mucin (Sigma) in PBS (4.4 mg/100 mL) overnight and were washed and blocked with 0.1% HSA. For each assay, a control test tube containing no bead, as a counting standard for [99mTc]- or [125I]-CC49, and a nonspecific binding tube containing the bead and 5 µL of a 5 mg/mL solution of the unlabeled MAb as a competitor of the labeled CC49, were used. Other tubes contained decreasing concentra-

Safavy et al. Scheme 1a

a (a) BocNHNH , HOBT, DCC, THF; (b) H , Pd/C (10%), 2 2 MeOH; (c) TFA:thioanisole:ethane-dithiol:anisole, 90:5:3:2, v/v.

tions (1:2 to 1:200) of the unlabeled MAb. To all tubes were added 100 µL of the labeled MAb CC49 (1200020000 cpm) followed by vortexing at room temperature for 1 h. All tubes, except the counting standard, were washed with DPBS, the beads were transferred into clean test tubes and were counted in a γ counter for 1 min. The percent of the radioactivity bound to the mucincoated beads were calculated. The inverse of the percent binding values was plotted against the cold MAb concentrations. Extrapolation to infinite MAb concentration determined the percent immunoreactivity2 of the labeled MAb CC49. RESULTS AND DISCUSSION

Since the introduction of our first hydroxamate BCA, trisuccin (10), we have devoted considerable effort to develop a flexible and efficient conjugation protocol for the covalent attachment of unprotected hydroxamate ligands to proteins and peptides (1, 11). Our ultimate goal was establishing a protocol which uses a stable derivative of hydroxamic acids to target a specific function on the protein, generating a stable hydroxamate-protein bond with retention of both the biological activity of the protein or peptide moiety and the metal-chelating property of the hydroxamate ligand. The stability and shelf life duration of the BCA would be of importance especially for the development of a radiolabeling kit. Recently, we reported a procedure in which an o-nitrophenyl activated ester of the hydroxamic acid was used for conjugation to the amine residues of the MAb CC49, as well as a synthetic peptide, used as model proteins (11). In this report, we introduce a newer technique which utilizes a suitably functionalized yet stable derivative of trisuccin which is capable of conjugation to the same model protein under mild reaction conditions. Synthesis. The hydrazide derivative of trisuccin was prepared in three steps from the protected carboxylic hydroxamate 1, as shown in Scheme 1. Thus, compound 1 was coupled to tert-butyl carbazate by a standard DCC/ HOBT reaction and the product 2 was hydrogenated to remove the benzyl protecting groups. This hydrogenation protocol had been developed previously (11) to minimize or eliminate the overreduction of the hydroxamate functions (19). All mass spectroscopic and NMR data indicated hydrogenation only to the desired extent with no overreduced products formed. Compound 3 was further deprotected by acidolysis to afford the final product 4. All intermediates and the final product were analyzed 2 The percent immunoreactivity refers to the fraction of the immunoreactive radiolabeled antibody as determined by its binding at infinite antigen excess.

Hydroxamate Hydrazide Protein/Peptide Conjugation Scheme 2a

Bioconjugate Chem., Vol. 10, No. 1, 1999 21 Table 1. Conjugation Efficiencies of 4-CC49 Formation as a Function of OHA:MAb Mole Ratiosa,b conjugate A B C D

OHA:MAbc 5 10 20 50

OHA:MAbd (0.7)e

2 4 (0.7) 8 (1.0) 17 (1.6)

4:Mabd 0.4 (0.1)e 0.4 (0.1) 0.6 (0.1) 0.8 (0.2)

a The OHA-MAb coupling was carried out at pH 7.5. b A theoretical 4:MAb ratio of 10 was used for the second step of each reaction. c Theoretical. d Experimental (MALDI). e (SD (n ) 3).

a

(a) PBS, pH 8.1; (b) 4, acetate buffer, pH 5.5, 2-propanol, NaCNBH3.

by standard techniques which identified the expected structures. The overall yield of the synthesis was 72%. Protein Conjugation. Two methods for the conjugation of 4 to the MAb CC49 were used: a linker-mediated technique (the OHA method, A), targeting the lysine residues of the protein, and an oxidation-mediated technique (the SOX method, B), targeting the carbohydrate residues of the antibody. The OHA Method. The linker-mediated technique is a two-step process in which a hydrazide-binding molecule is first introduced into the protein to serve as the anchor for the incoming BCA. In the second step, a molar excess of the hydrazide-containing BCA is added to target the ketone function of the anchor. The rationale for this approach is based on two considerations: (1) the higher rate of condensation of the of hydrazide groups with ketones at the pH range of 5 < pH < 6, and (2) the use of a large excess of the hydrazide with respect to the available protein -NH2 groups. To optimize this procedure, we first carried out a number of conjugations to assess the general trend of the reaction and then used these results to optimize the protocol. Since the selected lysine residues of MAbs are readily available for conjugation and their modification, carried out under appropriate conditions, does not result in the loss of biological activity, we selected this residue to build the hydrazide-binding scaffold on the protein as shown in Scheme 2. Thus, N-succinimidyl 6-oxoheptanoate (OHA-NS, 5) was selected as the bifunctional linker due to its ability to react with the lysine -amino groups through the activated ester moiety on one end and to a hydrazide through its ketone function, on the other. The OHA-NS was conjugated to MAb CC49 through standard N-hydroxysuccinimide active ester procedures. The reactions were generally complete within an hour or less and showed high coupling efficiencies, i.e., the experimentally measured number of linkers per MAb as compared to their theoretical values. It should be noted that while the kinetics of this reaction was not quantified, we observed that the number of linkers attached to the antibody, as measured by MALDI, did not increase any further by an increase in the incubation time. These data, for different OHANHS:CC49 mole ratios, are shown in Table 1. In the second step of the process, the hydroxamate hydrazide 4 was reductively coupled to the OHA-protein conjugate 6 in a low-pH medium to afford the final product conjugate 7. The low pH of this reaction provides the advantage of a selective coupling between the hydrazide function of the hydroxamic BCA and the implanted ketone group of the OHA handle. While amines undergo only very slow reactions with ketones at pH 5.0-5.5, hydrazides condense quite readily with ketones at this pH range to produce the corresponding hydrazone (20). This selectivity is an important feature of the hydrazide functionality, which prevents the competing reaction of

Table 2. Conjugation Efficiencies of 4-CC49 Formation as a Function of both OHA:MAb and 4:MAb Mole Ratiosa conjugate

OHA:Mabb

4:MAbc

4:Mabd

E F G H J K

9e/10

5 10 20 50 500 1000

0.5 (0.1)g 0.6 (0.1) 0.8 (0.1) 1.4 (0.4) 1.7 (0.3) 2.3 (0.3)

9e/10 9e/10 9e/10 3f/5 3f/5

a The OHA-MAb coupling was carried out at pH 8.1. b Experimental/theoretical. c Theoretical. d Experimental. e (SD ) 0.9 (n ) 3). f (SD ) 0.5 (n ) 3). g (SD (n ) 3).

the endogenous amine functions of the protein.3 These competing reactions may complicate the outcome of the conjugation process in the form of inter- or intramolecular protein condensations. To further stabilize the conjugate, the hydrazone bond was reduced under mild conditions with sodium cyanoborohydride, to the corresponding substituted hydrazide (21). The final product conjugates were analyzed by MALDI to evaluate the number of ligands incorporated into the protein molecule. These results for the conjugation of trisuccin hydrazide 4 to CC49 are shown in Table 2. Under the unoptimized conditions of Table 2, even at a high OHA:MAb ratio of 9, and at 4:MAb ratios of 20 and below, the extent of conjugation was negligible. A larger difference in the molecular weights appeared at 4:MAb ratios of >50, where the increase continued even at a lower OHA:MAb of 3. The SOX Method. The in situ conversion of the carbohydrate moieties of an antibody to aldehyde groups by oxidizing agents, followed by reductive amination by an NH2-containing molecule, has been reported (14, 16). In our earlier approaches, we attempted without much success to apply this technique to the conjugation of an aminohydroxamate to MAbs (11). We ascribed this lack of reaction to the low carbohydrate content of the IgGs (22) and did not further pursue the strategy. Interested in the development of conjugates with highly preserved immunoreactivities, we decided to investigate the oxidation method again, this time with a hydrazide (4) instead of an amine. This approach is shown in Scheme 3. Thus, the Mab CC49 was oxidized with sodium periodate in a low-pH acetate buffer to prevent self-condensation of the protein. After purification of the oxidized antibody, the trisuccin hydrazide 4 was added and the pH was raised to activate the hydrazide. Several reactions at different 4:MAb ratios were carried out, and the conjugate product (9, Scheme 3) was analyzed by MALDI. In contrast to 3 Safavy, A., unpublished results. The selective hydrazideketone coupling reported here has also been confirmed by our work on the synthesis of a conjugate of trisuccinn hydrazide with an 8-amino acid peptide, containing an unprotected lysine residue and the OHA linker on the N-terminal. No detectable coupling of the hydrazide with the side-chain amine of lysine was observed. These results will be submitted for publication shortly.

22 Bioconjugate Chem., Vol. 10, No. 1, 1999 Scheme 3a

Safavy et al. Table 3. Optimized Conjugation Efficiencies of 4:CC49 Formation as a Function of both OHA:Mab and 4:Maba conjugate

OHA:Mabb

4:MAbc

4:Mabd

L M N P

9.6e/10

50 100 250 500

2.1 (0.3)f 2.5 (0.3) 3.2 (0.5) 3.5 (0.5)

9.6e/10 9.6e/10 9.6e/10

a All 4-CC49 couplings carried out in acetate buffer/2-propanol mixture. b Experimental/theoretical. c Theoretical. d Experimental. e (SD ) 0.6 (n ) 3). f (SD (n ) 3).

Table 4. Radiolabeling Yields (RY) and Immunoreactivities (IR) of the 99mTc- and 4-CC49 Conjugatesa 99mTc-RY

a (a) NaIO , acetate buffer, pH 4.2; (b) 4, acetate buffer, pH 4 5.5, NaCNBH3.

the OHA method, the oxidation reactions generally showed lower yields of final product recovery. Closer analyses of these reactions revealed that a considerable portion of the protein was lost due to self-condensation as evidenced by the formation of an insoluble precipitate. This loss of product was found to be pH dependent: to activate the amino groups of a BCA for condensation with an aldehyde, the pH of the solution has to be raised from the pH at which the antibody was oxidized. At this stage, the native amine residues of the protein may start condensing with the newly formed aldehydes, resulting in higher molecular weight aggregates. This may explain why our earlier attempts with an amino-trisuccin (11) were not successful. With the hydrazide 4, however, it was possible to use a lower pH for the condensation reaction, and we were able to isolate the desired product, albeit at a low yield. Furthermore, it was found that the oxidation step of this approach was quite sensitive toward concentration of the oxidizing agent. At NaIO4 concentrations above 50 mM, the MALDI showed considerable amounts of fragmentation of the MAb. An optimized protocol for conjugation of 4 to CC49 was, however, worked out, as shown in the Experimental Procedures, which resulted in a 4:CC49 ratio of 2.5 and an immunoreactivity of 40% for the 125I-labeled conjugate. Optimization of the Conjugation Protocol. Because no special immunoreactivity advantages were realized from the oxidation method, we chose the OHA method for optimization and evaluated further modifications accordingly. The results of Table 1 (column 3, the OHA:MAb MALDI values) show that the average efficiency4 of the first step, conjugation of OHA to the MAb, was 38.5 ( 2.8% at a buffer pH of 7.5. Increasing the pH to 8.1 increased this yield to 80 ( 20% as shown by the OHA:MAb ratios of Table 2 (column 2). The latter pH was therefore used for all OHA couplings which consistently showed coupling yields within the same range or higher (Table 3), according to the MALDI analyses. We further considered that the 7-carbon structure of OHA may impart some degree of lipophilicity which may pull the linker deep inside of the antibody molecule due to hydrophobic forces. This hydrophobic repulsion may, in turn, “hide” the ketone residue from a reactive collision 4 The average efficiencies (AEs) were calculated from the following equation: AE ) 100{∑(n, i)1)[(OHA:MAb)E/(OHA: MAb)T]/n}, where n is the number of conjugates in a series, [OHA:MAb]E and [OHA:MAb]T are the experimental and theoretical linker-to-MAb ratios, respectively.

125I-Labeled

SAb (mCi/mg)

125I-IR

(%)

99mTc-IR (%)

125I-RY

conjugate

(%)

(%)

L M N P

72 81 81 87

56 51 50 45

9 9 13.5 15.4

50 50 50 50

63 59 67 50

a Averages of two runs. All conjugates prepared under optimized conditions. b Specific activities for 99mTc-CC49 conjugates.

with the incoming hydrophilic hydrazide. To overcome this problem, we increased the organic character of the reaction medium by lowering the ionization power of the reaction buffer and introducing 2-propanol as the cosolvent. Indeed, these modifications did increase the conjugation efficiency, as shown in Table 3. The measured ratios for conjugates 50:1 and 100:1 (L and M of Table 3, respectively), for instance, are higher or comparable to those for 500:1 and 1000:1 conjugates obtained under unoptimized conditions (J and K of Table 2, respectively). Compounds L-P (Table 3) were radiolabeled with both 99mTc and 125I, for evaluation of their labeling efficiencies and immunoreactivities as shown in Table 4. The use of 125I instead of a radiometal has the advantage of excluding the effects of the metallic cation and associated labeling chemistries on the immunoreactivity, thus revealing the pure effect of the BCA conjugation. These would constitute reference values against which the immunoreactivity of the radiometal-linked antibody could be compared. Immunoreactivities were estimated through a modified Lindmo assay (23). The RYs for 99mTc labeling of the conjugates and their specific activities increased with an increase in the L:MAb ratio while the RYs for iodinations did not significantly change indicating the lack of dependence of this labeling on the number of metal-chelating ligands. Interestingly, all immunoreactivities of the 99mTc-labeled antibodies were slightly lower than those for the 125I-labeled conjugates. This observation may show an expected effect of the metal on the overall structure of the protein. CONCLUSION

The work described in this report introduces two new strategies for the conjugation of unprotected hydroxamic acids to suitable proteins, peptides, and other molecules of biological and medicinal interest. The technique utilizes a new derivative of our trihydroxamate ligand, trisuccin, containing a hydrazide function and capable of forming a hydrazone with carbonyl-containing molecules. Two important advantages of this new method are (1) the use of a stable derivative of hydroxamic acids, as opposed to those with more limited stabilities (e.g., activated esters) (11), and (2) the chemical compatibility of the hydrazide and hydroxamate functionalities. Conjugation may be conducted through two different routes: in the ketolinker-mediated path, the hydroxamate hy-

Hydroxamate Hydrazide Protein/Peptide Conjugation

drazide is directed to a carbonyl function, anchored on the protein molecule during a previous step; in the sugar oxidation route, the hydroxamate is directed toward the in situ generated aldehydes. These studies indicated that the former method was a more efficient technique which better preserved the structural integrity of the protein. As a whole, this new strategy may constitute a useful option in the covalent attachment of unprotected hydroxamates to molecules with biological, industrial, and synthetic applications. ACKNOWLEDGMENT

This work was supported by NIH Grants CA62550, CA44173, and DOE DE-FG02-96ER62181. The authors acknowledge the assistance of Dr. Ramin Arani (Biostatistics Unit) for the calculation of standard deviations, the Comprehensive Cancer Center Nuclear Magnetic Resonance Shared Facilities for 1H NMR spectra, and Kamellia Safavy for protein determinations and radioimmunoassays. We thank Sally Lagan for manuscript preparation. LITERATURE CITED (1) Safavy, A., Khazaeli, M. B., and Buchsbaum, D. J. (1997) Synthesis of bombesin analogues for radiolabeling with Rhenium-188. Cancer 80 (Suppl.), 2354-2359. (2) Yarranton, G. (1997) Antibodies as carriers for drugs and radioisotopes. In Antibody Therapeutics (W. J. Harris and J. R. Adair, Eds.) pp 53-72, CRC Press, Boca Raton. (3) Govindan, S. N., Goldenberg, D. M., Grenbenau, R. C., Hansen, H. J., and Griffiths, G. L. (1996) Thiolations, 99mTc labelings, and animal in vivo biodistributions of divalent monoclonal antibody fragments. Bioconjugate Chem. 7, 290297. (4) Kotts, C. E., Su, F. M., Leddy, C., Dodd, T., Scates, S., Shalaby, M. R., Wirth, C. M., Giltinan, D., Schroff, R. W., Fritzberg, A. R., Shepard, H. M., Slamon, D. J., and Hutchins, B. M. (1996) 186Re-labeled antibodies to p185HER2 as HER2targeted radioimmunopharmaceutical agents: Comparison of physical and biological characteristics with 125I and 131Ilabeled counterparts. Cancer Biother. Radiopharm. 11, 133144. (5) Fischman, A. J., Babich, J. W., and Strauss, H. W. (1993) A ticket to ride: Peptide radiopharmaceuticals. J. Nucl. Med. 34, 2253-2263. (6) Liu, S., Edwards, D. S., and Barrett, J. A. (1997) 99mTc labeling of highly potent small peptides. Bioconjugate Chem. 8, 621-636. (7) Arano, Y., Wakisaka, K., Mukai, T., Uezono, T., Motonari, H., Akizawa, H., Kairiyama, C., Ohmomo, Y., Tanaka, C., Ishiyama, M., Sakahara, H., Konishi, J., and Yokoyama, A. (1996) Stability of a metabolizable ester bond in radioimmunoconjugates. Nucl. Med. Biol. 23, 129-136. (8) DeNardo, S. J., Zhong, G.-R., Salako, Q., Li, M., DeNardo, G. L., and Meares, C. F. (1995) Pharmacokinetics of chimeric L6 conjugated to indium-111- and yttrium-90-DOTA-peptide in tumor-bearing mice. J. Nucl. Med. 36, 829-836. (9) Ultee, M. E., Bridger, G. J., Abrams, M. J., Longley, C. B., Burton, C. A., Larsen, S. K., Henson, G. W., Padmanabhan, S., Gaul, F. E., and Schwartz, D. A. (1997) Tumor imaging with technetium-99m-labeled hydrazinonicotinamide-Fab′ conjugates. J. Nucl. Med. 38, 133-138.

Bioconjugate Chem., Vol. 10, No. 1, 1999 23 (10) Safavy, A., Buchsbaum, D. J., and Khazaeli, M. B. (1993) Synthesis of N-[tris[2-[[N-(benzyloxy)amino]carbonyl]ethyl]methyl]succinamic acid, trisuccin. Hydroxamic acid derivatives as a new class of bifunctional chelating agents. Bioconjugate Chem. 4, 194-198. (11) Safavy, A., Sanders, A., Qin, H., and Buchsbaum, D. J. (1997) Conjugation of unprotected trisuccin, N-[tris[2-[(Nhydroxyamino)carbonyl]ethyl]methyl]succinamic acid, to monoclonal antibody CC49 by an improved active ester protocol. Bioconjugate Chem. 8, 766-771. (12) Brinkley, M. (1993) A brief survey of methods for preparing protein conjugates with dyes, haptens, and cross-linking reagents. In Perspectives in Bioconjujate Chemistry (C. F. Meares, Ed.) pp 59-70, American Chemical Society, Washington, DC. (13) Tijsen, P., and Kurstack, E. (1984) Highly efficient and simple methods for the preparation of peroxidase and active peroxidase-antibody conjugates for immunoassays. Anal. Biochem. 136, 451-457. (14) Chamow, S. M., Kogan, T. P., Peers, D. H., Hastings, R. C., Byrn, R. A., and Ashkenazi, A. (1992) Conjugation of soluble CD4 without loss of biological activity via a novel carbohydrate-directed cross-linking reagent. J. Biol. Chem. 267, 15916-15922. (15) Rodwell, J. D., Alvarez, V. L., Lee, C., Lopes, A. D., Goers, J. W. F., King, H. D., Powsner, H. J., and McKern, T. J. (1986) Site specific covalent modification of monoclonal antibodies: In vitro and in vivo evaluations. Proc. Natl. Acad. Sci. U.S.A. 83, 2632-2636. (16) Nakane, P. K., and Kawaoi, A. (1974) Peroxidase-labeled antibody. A new method of conjugation. J. Histochem. Cytochem. 22, 1084-1091. (17) Muraro, R., Kuroki, M., Wunderlich, D., Poole, D. J., Colcher, D., Thor, A., Greiner, J. W., Simpson, J. F., Molinolo, A., Noguchi, P., and Schlom, J. (1988) Generation and characterization of B72.3 second generation monoclonal antibodies reactive with the tumor-associated glycoprotein 72 antigen. Cancer Res. 48, 4588-4596. (18) Fraker, P. J., and Speck, J. C. (1978) Protein and cell membrane iodinations with a sparingly soluble chloroamide, 1,3,4,6-tetrachloro-3R,6R-diphenylglycoluril. Biochem. Biophys. Res. Commun. 80, 849-857. (19) Bergeron, R. J., McManis, J. S., Phanstiel, O., IV, and Vinson, J. R. T. (1995) A versatile synthesis of deferrioxamine B. J. Org. Chem. 60, 109-114. (20) Shao, J., and Tam, J. P. (1995) Unprotected peptides as building blocks for the synthesis of peptide dendrimers with oxime, hydrazone, and thiazolidine linkages. J. Am. Chem. Soc. 117, 3893-3899. (21) Jentoft, N., and Dearborn, D. G. (1979) Labeling of proteins by reductive methylation using sodium cyanoborohydride. J. Biol. Chem. 254, 4359-4365. (22) Goding, J. W. (1986) Production and applications of monoclonal antibodies in cell biology, biochemistry and immunology. Monoclonal Antibodies: Principles and Practice, pp 5-58, Academic Press, Orlando. (23) Lindmo, T., Boven, E., Cuttitta, F., Fedorko, J., and Bunn, P. A., Jr. (1984) Determination of the immunoreactive fraction of radiolabeled monoclonal antibodies by linear extrapolation to binding at infinite antigen excess. J. Immunol. Methods. 72, 77-89.

BC980045D