Antibody labeling with copper-67 using the bifunctional macrocycle 4

Nov 1, 1991 - ... imaging with copper-64 in the drug discovery and development arena ... Mahesh K. Bhalgat , Jeanette C. Roberts , Janet A. Mercer-Smi...
0 downloads 0 Views 2MB Size
Bioconjugate Chem. 1991, 2, 415-421

415

Antibody Labeling with Copper-67 Using the Bifunctional Macrocycle 4-[ (1,4,8,11-Tetraazacyclotetradec-l-yl)methyl]benzoic Acid Peter M. Smith-Jones,? Raimund Fridrich,t Thomas A. Kaden,s Ilse Novak-Hofer,? Karl Siebold,t Daniel Tschudin,* and Helmut R. Maecke’yt Radiopharmacy Division, Paul Schemer Institut, Wuerenlingen, Switzerland, Institut fur Anorganische Chemie der Universitat Basel, Basel, Switzerland, and Department of Nuclear Medicine, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland. Received July 24, 1991

The high kinetic stability of the Cu2+complex of the chelator 4-[ (1,4,8,ll-tetraazacyclotetradec-l-y1)methyllbenzoic acid was demonstrated at physiological pH as well as under acidic conditions. The chelating agent was conjugated to AB35, a monoclonalantibody directed against CEA, without a significant loss of immunoreactivity. The conjugate could, under optimal labeling conditions, be labeled with W u in acetate buffer with a full occupancy of ligands within 20 min. This radiolabeled conjugate showed no transfer of radiocopper to serum proteins in human serum over 7 days. The biodistribution in tumor-bearing mice was measured and compared to that of iodinated AB35. Tumor uptake was high with 15 f 3 9; ID (injected dose)/g after 24 h and 32 f 7% ID/g after 96 h for the e7Cu-labeledantibody and 13 f 4 % ID/g after 24 h and 14 f 2% ID/g after 96 h for the 1251-labeledantibody. Whereas radioactivity in normal organs decreased with time after 24 h, increased residence time was shown up to 4 days with the 67Cu-labeledAB35.

Radiolabeled monoclonal antibodies show promise as specific agents for cancer diagnosis and therapy (1-4). The most commonly used radioisotopes for antibody labeling are those of iodine (5). They have several drawbacks, e.g. logistical problems, the need of having labeling units for the iodination of the antibodies, etc. Furthermore, the in vivo deiodination process leads to the incorporation of iodine into both the thyroid and the stomach; it also appears to be responsible for short residence times of the radiolabel on or in the tumor cell because of antibody processing. Radiometals have the potential to overcome these problems if they can be stably bound to the antibody with preservation of its immunoreactivity (6, 7). Higher flexibility in terms of labeling is also possible because this may be performed by simply mixing of the radioisotope, in an appropriate buffer, with the antibody conjugated to a suitable bifunctional ligand (8). Among the metallic radionuclides for therapy, e7Cuis one of the more promising because of a 61.9-h half-life, which is ideal in terms of the residence time of antibodies on tumors and in terms of dose rate. It releases @-particlesas well as y-emissions of 93 and 184 keV which allow easy biodistribution studies and tumor imaging in patients. Furthermore, 64Cuis a positron emitter with the potential of accurate dose assessment, three dimensional-representation of tumors, and high resolution. The properties of these two copper radionuclides are summarized in Table I. In order to utilize 67Cu2+, a kinetically labile radiometal, one has to incorporate it into a bifunctional ligand which ensures that the metal ion remains bound to the carrier antibody under physiological conditions. Moreover, the chelator must be covalently bound to the protein. We have previously studied the covalent attachment of the Cu2+complex of the functionalized macrocyclicligand 4-[(1,4,8,11-tetraazacyclotetradec-l-yl)methyl] benzoic acid (CPTA, Figure 1) to simple amines as model compounds and to bovine serum albumin (9). Furthermore, we showed that the 64Cucomplex can be attached to the antibody + Paul

Table I. Properties of Copper Radionuclides properties w u 6’CU half-life, h 12.9 61.9 decay modes @ (39.6% ) B (100%) EC (45.0%) 8+(19.3%) 8-energies,keV B 573 (40%) 395 (51%) 484 (28%) 577 (20%) P+ 654 (19%) y-energies, keV 511 (39%) 91 (6%) 1346 (0.6%) 93 (35% ) 185 (45%) 300 (0.6%)

b-12 using the preconjugation labeling approach without significant loss of immunoreactivity (10). In continuation of these studies, we present new studies with the same ligand in the postconjugation labeling approach with the antibodyAB35, which is directed against CEA producing cells (11). Bifunctional macrocyclicligands have been used before as a means to synthesize ‘j7Cu- and/or 64Cu-labeled antibodies (12,13). Most advanced is the use of the excellent copper chelator 6-(pnitrobenzyl)-TETA,’ a bifunctional carboxymethylated cyclam-14 derivative (14). The high stability of the copper chelate under physiological conditions has been demonstrated and attributed to its high kinetic stability. The chelate has been tested bound to the LYM-1 antibody in tumor-bearing mice with good success. Macrocyclic amines are known to form Cu2+

Schemer Institute.

University Hospital Basel. 5 Institut fur Anorganische Chemie der Universitat Basel.

8

U Figure 1. Structure of 4-[(1,4,8,11-tetraazacyclotetradec-l-yl)methyl]benzoic acid (CTPA).

lTETA, 1,4,8,ll-tetraazacyclotetradecane-N,N’,N’’,N”’-teb raacetic acid. 0 199 1 American Chemical Society

416

Bioconjugate Chem., Vol. 2, No. 6, 1991

SmithJones et al.

complexes which are kinetically more stable than polyamino polycarboxylates (15). Two bifunctional derivatives of cyclam-14 conjugated to antibodies have also been tested by two different groups (12, 16). The aim of this study was to demonstrate the feasibility of a kit formulation with our simple bifunctional macrocycle and to compare this radiopharmaceutical with the corresponding iodinated antibody in tumor-bearing mice. EXPERIMENTAL PROCEDURES Materials a n d Methods. All reagents and solvents were obtained from commercial sources and were used without further purification. Copper stock solutions were prepared from CuSOc5HzO. The concentration was determined by volumetric titration with NazEDTA and murexid as indicator (17). Sephadex G-50 was used for gel filtration, and HPLC on the labeled antibody and antibody chelator conjugates was done on a HewlettPackard HPLC system (Model 1050) and a Waters 30-cm TSK 3000 column with a 5-cm TSK 125 precolumn. UV studies were performed with a Perkin-Elmer Lambda 2 UVspectrophotometer coupled to a Compaq 386 computer and PEXIS software (Perkin-Elmer). Radioactive fractions were measured with an automated Packard well counter and a manual Picker ionization chamber depending on the activities involved. Both detectors were calibrated with e7Cu sources whose activities were determined by y-ray spectroscopy with a HPGe detector. Radiochromatograms of either TLC or gel chromatography were analyzed using a Berthold TLC linear analyzer coupled to a IBM PC-XT, using a Chroma software (Berthold). Polyacrylamide SDS gel electrophoresis was done according to Lammli (18) on a LKB 2001 electrophoresis chamber and a Biowerk power supply with Tris buffer (0.375 M, pH 8.3) at 300 V. The protein bands were visualized using Coomassie Brilliant Blue staining. Ultra filtration was done using 10 kDa Omegacell filters (Filtron Technology Corp.) and/or 30 kDa Mini Cent filters (Bio-Rad Laboratories). The bifunctional macrocycle 4-[(1,4,8,11-tetraazacyclotetradec-1-y1)methyllbenzoicacid tetrahydrochloride (CPTA) (Figure 1) was synthesized as described earlier (19). Three batches of G7Cuwere obtained, two from Los Alamos Scientific Laboratory (Los Alamos, NM) (39 pCi 67Cu/pgCu and 1.6 mCi/pg of Cu in 2 M HCl) and another from Brookhaven National Laboratory (Upton, NY) (0.77 mCi 67Cu/pg of Cu in 0.1 M HCl). Zn2+impurities were removed by using cation-exchange chromatography (1.5mL column AGl-X8, 100-200 mesh). The monoclonal antibody used in this study was the IgGl AB35, which is a murine anti-CEA antibody and was originally prepared as described by Haskell et al. (11). Stability of Cu-CPTA in Aqueous Solution a n d Human Serum. The Cu2+ complex of CPTA was synthesized as described earlier (9)and isolated as a perchlorate salt. A 5 mM stock solution ( I = 1.0, KN03) was prepared. The reaction was started by mixing this solution with 1.0 M HN03, 0.1 M HN03, or 5 X M EDTA (pH 7.3,I = 1.0,KN03). The decay of the complex was followed spectrophotometrically at 37 "C. Spectra were taken at different time intervals up to 7 days. The stability in human serum was determined by mixing 7.5 pL of a 6.6 mM solution of CPTA at pH 5 (0.1 M NaOAc) with 0.5 pL of a 6 pM 67CuC1~ stock solution (610 pCi/mL, 1.6 mCi/pg), and the mixture diluted to 38 pL (0.1 M NaOAc, pH 5). After 2 h at 37 "C the mixture was mixed into 1.3 mL of sterile human serum and incubated up to 8 days at 37 "C in a chamber maintained a t 5% COZ and 95% air.

Figure 2. G-50gel filtration elution curve of EDTA removal of nonspecifically boundWu with both the CPTA-AB35 conjugate and unmodified AB35. The protein peak has an elution maximum at 6 mL and 67C~(EDTA)2has an elution maximum at 12 mL.

After appropriate intervals, 100-pLsamples were taken and subjected to gel filtration. The relative amounts of both CPTA and protein-bound radiocopper were determined by gel-filtration chromatography on a Sephadex G-50 column (19 X 1cm) using Tris buffer (0.1 M, pH 7). Conjugation Chemistry. A mixture of 30.6 mg of ligand, 75.3 mg of N-ethyl-N'-[3-(diethylamino)propyl]carbodiimide dihydrochloride and 16.2 mg of N-hydroxysuccinimide was dissolved in 2.2 mL of phosphate buffer (0.1 M, pH 7). The reaction was allowed to proceed for 1 h a t room temperature before being further diluted to 4.4 mL with phosphate buffer. Three portions of 550 p L of AB35 (3.33 mg/mL, 0.1 M Tris buffer, pH 7) were reacted with 8.2, 24.6, and 73.8 pL of the above mixture. After 2 h a t room temperature the protein was recovered by ultrafiltration and buffer changed to 0.1 M sodium acetate (pH 5.5). The protein fractions were then filtered using a 0.22-pm filter, and an aliquot was assayed using the absorbance at 276 nm. The absorption a t this wavelength was determined to be 1 AU for 1 mg/mL. In a second experiment three portions of 500pL of AB35 (3.33 mg/mL, 0.1 M phosphate buffer, pH 7.0) were reacted with 20,40, and 80 pL of the above mixture. A further two portions of 250 pL of AB35 solution were reacted with 80 and 160 pL of the ligand/activated ester mixture. After 1 h at room temperature the ligand-conjugated protein was recovered by ultrafiltration and the buffer changed to 0.1 M acetate (pH 5.5). Quantitating the Number of Copper Binding Sites. Ten- or 100-pg portions of modified antibody were incubated with a known excess of a standardized solution of CuS04 (0.1 M NaOAc, pH 5.5), containing tracer amounts of 67Cu, at 37 "C. Incubation times varied between 20 min and 24 h. Following incubation, the excess copper was complexed with EDTA (50 mM, pH 5.5) a t room temperature for 10min. The mixtures obtained were then separated on a G-50 column (19 X 1cm) or on a TSK 3000 column (30 cm), using either 0.1 M PBS (pH 7.0 with 0.5% BSA) or 0.1 M PBS (pH 7.0), depending on the concentration of antibody. One-milliliter fractions of the eluate were collected and counted in an automatic well counter. A typical elution profile is shown in Figure 2. The number of macrocyclic ligands attached to the antibody was determined by relative amount of activity recorded in protein fractions and calculated as follows: number of ligands = activity in protein fraction x moles of Cu total activity X moles of AB 35 As we found slight differences in the calculated number of copper binding sites depending on the separation

Bioconlugate Chem., Vol. 2, No. 6, 1991 417

Antibody Labeling with Copper-67

method (G-50 or TSK 3000 column), all further experiments were done on HPLC-purified labeled immunoconjugates. Several ligand binding site assays were performed using a 50 times excess of zinc with respect to copper to ascertain if there is a need to purify 67Cu from any zinc present in commercial supplies. Radiolabeling of CPTA-AB35 Conjugate. Five hundred microliters of CPTA-AB35 conjugate (1.36 binding sites, 1.62 mg/mL in 0.1 M NaOAc, pH 5.5) was mixed with 60 pL of 67Custock solution and incubated for 1 h at room temperature. Following complexation, 100 pL of 50 mM NazEDTA (in 0.1 M NaOAc, pH 5.5) was added to chelate unspecifically bound 67Cu. After 5 min the reaction mixture was loaded onto a TSK 3000 column (8 X 300 mm, 8 X 50 mm precolumn). The radiolabeled conjugate was eluted with 0.1 M PBS (pH 7.3) at 1 mL/ min. The first peak corresponding to a 300-400 kDa molecular mass was rejected and only part of the main peak collected. In total, 284 pg of labeled AB35, with an activity of 71.3 pCi, was recovered in a 1.2-mL volume. The second coupling experiment afforded five fractions of modified AB35. They were incubated with 3-5 times the amount of e7Cu solution relative to the number of available CPTA binding sites in a 0.1 M acetate buffer (pH 5.5) for 30 min at room temperature. The same purification procedure was followed as described above to yield 41% of protein. Iodination of AB35. One hundred microgramsof AB35 was labeled with 1251 (0.5 mCi/18.5 MBq) using 20 pg of Iodogen (Pierce) in 300 pL of PBS for 15 min on ice according to the manufacturer's protocol. Labeled mAb was purified over a Sephadex G10 column (PD 10, Pharmacia). In Vitro Cell-Binding Tests. Titer curves of AB35CPTA conjugates were compared with underivatized AB35 by binding various concentrations of mAb's to a fixed number of cells (0.5 X 106/well) for 16 h at 37 "C. After removal of unbound mAb, bound mAb was detected by a second incubation of 1251-labeledgoat anti-mouse antibody ('ZbI-GAM, Bio-Rad) for 2 h. After incubation cells were washed with PBS/HSA, dissolved in 0.5 mL of 1N NaOH, and counted in a y-counter. Immunoreactivity of radiolabeled AB35 was determined by cell-binding assays with the CEA-producing LoVo (human colon adeno carcinoma) cell line (American Type Culture Collection). Cells were seeded in 24-well microtiter plates (maximally 2 X lo6 cells/well) and allowed to attach for 1 h at 37 "C. The medium (Minimal Eagles Medium containing 10% fetal calf serum) was then taken off; plates were dried for 3 days at room temperature in a laminar-flow hood and could then be stored at 4 "C for at least 6 months. Before use, plates were presaturated for 60 min at 37 "C with 0.5 mL/well of PBS containing 0.5% HSA. Binding of 1251-or e7Cu-labeled AB35 was measured by incubating lo5cpm of labeled mAb in 0.5 mL of PBS-HSA for 16 h at 37 "C. Plates were then washed three times with 0.5 mL of ice-cold PBS-HSA. Cells were dissolved in 0.5 mL of 1 N NaOH and radioactivity was counted in a y-counter. Nonspecific binding was determined in separate wells by addition of 10 pg of unlabeled mAb to the labeled preparations and was subtracted. Immunoreactivity of radiolabeled mAb was determined by binding to increasing numbers of cells and analysis of saturation curves by the method of Lindmo (20). Stability Studies of 67Cu-CPTA-AB35 in PBS Buffer and Human Serum. The stability of the W u labeled AB35-CPTA conjugate was determined by copper

Table 11. Stability of Cu-CPTA Complex from Spectrophotometric Data % dissociation of Cu-CPTA complex time, days 50 mM EDTAD 50 mM HN03 500 mM HN03 0

0

0.07 0.28 0.69 2.78 6.94

0 0

0 0 0

0.7 0.7

2.8 6.9

0

1.5 14.7 36.7

a The optical density change of measurements in EDTA is within experimental error.

exchange between the immunoconjugate and free Cu2+. The 67Cu-CPTA-AB35 conjugate (143 pM in PBS, pH 7.3,4.5 Cu atoms per AB35) was treated with a 50 times excess of cold Cu2+. The solution was maintained at 37 "C, and 500-pL aliquots were removed at intervals over a 5-day period. The removed solutions were immediately treated with 500 pL of a 50 mM EDTA solution allowed to stand for 5 min and subject to analysis on a G-50 column (19 X 1 cm) using 0.1 M PBS (pH 7.3). The stability of the radiocopper-labeled conjugate was studied in vitro by using human serum at 37 "C and in vivo by analyzing blood samples reserved from sacrificed mice by size-exclusion HPLC and SDS gel electrophoresis. In the former case samples were prepared by mixing 10 pL of 67Cu-labeledAB35 with 1 mL of human serum and incubating the mixture in a 5% C02, 95% air environment for up to 7 days. In the latter case samples of murine serum was recovered from the animals and 100pL samples were analyzed by SDS gel electrophoresis. Animal Biodistribution Studies. Nude mice (strain ZH ICR nu/nu, Institut fur Versuchstierforschung) were subcutaneouslyinjected with 5 X 106LoVo cells (suspended in Minimal Eagles Medium containing 10% fetal calf serum) into the right and left flank. After 1week, tumors of 20-300 mg were produced. The mice were then injected either with 10 pg of iodinated AB35 (5 mCi of lz51/mg)or 20 pg of 67Cu-labeledAB35 (250-500 pCi/mg) by means of an intravenous tail injection. The animals were counted in a well counter at regular time intervals after injection until sacrifice. The animals were sacrificed between 1 and 4 days after injection and were dissected directly after sacrifice. The various organs were removed, rinsed in saline, blotted, weighed, and counted in an automated well counter. The uptake of radioactivity by organs was calculated as percent injected dose per gram tissue. At least three animals were used for every time point. Imaging Studies. Prior to sacrifice, the mice, previously injected with the 67Cu-labeledAb35, were anesthesized for imaging studies by intraperitoneal injection of Nembutal. They were subsequently imaged, at several time points, with a Picker y-camera interfaced to a VAX computer. A low-energycollimator was used and the 20 % window centered at 180 keV; 30 000 counts were collected. RESULTS

Stability of Cu-CPTA in Aqueous Solution and Human Serum. The stability of Cu-CPTA was studied spectrophotometrically at pH 7.3 and 37 "C with excess EDTA. Within 7 days no spectral change could be observed, indicating extremely high kinetic stability of the complex (Table 11). Even at low pH in 0.05 M HN03 only 7% of the complex dissociated within 7 days. Significant dissociation only occurs in 0.5 M HNO3 with around 37% after 3 days. The same high kinetic stability of the complex is found in human serum with less than 0.1 % of radioactive W u 2 +

418

SmWones et 81.

Bioconlugete Chem., Vol. 2, No. 6, 1991

Scheme I 0

P N - O H

Y

cE1

a

:

/- CHI

H

u HL2

c v

20

0

40

60

80

100

120

Maor of AB36(ng) Ab

Figure 3. Typical immunoreactivity binding curves for CPTAAB35 conjugate and unmodified AB35 The CPTA-AB35 conjugate has an average of 4.5 chelators per antibody.

*

u

0

Ab

W

Table 111. Properties of mAb3SCPTA Conjugate reaction % immuno- ratio % immuno'3'Cu ratio ratio reactivity of CPTA reactivity of heavy:light ligand:AB35 CPTA-AB350 AB35 Wu-CPTA-AB356 chain 80-100 2.14 51:l >90 4.5 80-100 1.95 102:l >90 4.5 80-100 1.96 204 1 >90 5.4 80-100 1.89 408 1 >90 5.6

loo I

101

QQ

Q6

0

I

1

2

3 4 Tim (Dam)

6

8

Immunoreactivity as determined by the Sandwich assay. Immunoreactivity as determined by the Lindmo procedure (20).

Figure 4. Exchange studies of s7Cu-CTPA-AB35with a 50-fold excess of cold copper in phosphate buffer, pH 7.0. The pseudofirst-order rate constant is estimated to be 6.7 X lo4 day1.

transferred to serum proteins within the time period of 8 days (data not shown). Conjugation Chemistry. The reaction scheme for conjugation of the ligand to AB35 is outlined in Scheme I. As shown before activation of the carboxylate group of CPTA can be done with a mixture of N-ethyl-N'-[3-(diethylamino)propyl]carbodiimide and N-hydroxysuccinimide (9). This affords a conjugate with 1.36 chelators per antibody when conjugation is done in Tris buffer under the conditions described. This conjugate showed preserved immunoreactivity and was used in the first of two in vivo experiments in tumor-bearing nude mice. Buffer exchange to PBS afforded chelator to antibody ratios of up to 5.6, depending on reagent ratios (Table 111). Radiolabeling and Quantitation of Binding Sites. Since all unlabeled immunoconjugates exhibited a high immunoreactivity,they were used for further experiments without further purification. For critical applications (a secondary immunoreactivity test and the animal experiment) the radiolabeled conjugate was purified by HPLC. The number of binding sites per antibody was determined by excessive reaction with a known concentration of Cu2+tracered with radiocopper. Copper complexation to the AB35-CPTA conjugate was done in acetate buffer (pH 5.5). It is fast and complete within 20 min. No further increase of protein-bound copper was observed after this time and up to 24 h. The labeling experiment in the presence of a 50 times excess of Zn2+ showed a slight reduction of copper incorporation to 80% of the theoretical labeling yield within 30 min, indicating rather high selectivity of the bifunctional macrocyclefor Cu2+. Thus it may be possible to avoid time-consuming purification of the radiocopper solutions. In several blank runs using unmodified AB35, the unspecific binding of copper was low, demonstrating the

effectiveness of the EDTA complexation of excess copper. Figure 2 shows typical G50 separation runs for the s7CuCPTA-AB35 conjugate and AB35 incubated with s7Cu2+. SDS gel electrophoresisof the HPLC-purified antibody showed approximately a 2:l ratio for the distribution of activity of the heavy and light chains, respectively. There was a slight trend for light chain labeling at the higher reaction ratios (Table 111). Iodination. Yield for iodination were typically between 90 and 100%. Specific activity of the iodinated antibody was 5 mCi/mg. In Vitro Cell-Binding Test. The unlabeled immunoconjugates were tested for immunoreactivity with the described Sandwich assay and compared to unmodified antibody. A typical binding curve can be seen from Figure 3. The immunoreactivity of different conjugates is shown in Table 111. Conjugation of up to 5.6 ligands per antibody did not impair the biological integrity to an appreciable extent. The binding assay showed binding of ca. 90% of modified AB35 to the plated LoVo cells if compared to the unmodified antibody. The immunoreactivity of the radiolabeled antibody measured according to Lindmo (20) was 85 f 15% (Table 111) and in accordance with the unlabeled conjugates. The immunoreactivity of the radioiodinated antibody was 90 f 10%. Stability Studies of 67Cu-CPTA-AB35 in PBS Buffer and Human Serum. Exchange measurement in phosphate buffer (pH 7) of the radiocopper-labeled conjugate and excess cold copper gave an apparent halflife of copper exchange of >lo00 days (Figure 4). This is consistent with the in vitro stability of Cu-CPTA toward excessEDTA and in human serum toward serum albumin. The labeled conjugate proved also to be unchanged after incubation in human serum (data not shown). Animal Biodistribution Studies. Animal experiments were done with a conjugate having 1.36 chelators

a

Antibody Labeling wRh Copper-67

Bloconjugate Chem., Vol. 2, No. 6, 1991

419

-" 16

P 610 \ D

e

s 5

"

81

He

Th

Ki

Sp

In

St

Mu

Li

0

Tu

81

He

Th

Ki

Sp

Oroan

In

St

Li

Mu

Tu

Orom

Figure 5. Biodistribution data for 67Cu-CPTA-AB35 in 12 tumor-bearing nude mice. Mice were injected with 20 pg of 'Wu-CPTA-AB35 containing either 1.3 or 4.5 CPTA groups per antibody. Data are means f SD of three animals. No significant

variations were observed between the two different radiolabeled conjugates. B1= blood, He = heart, Th = thyroid, Sp = spleen, Ki = kidney, St = stomach, In = intestines, Li = liver, Mu = muscle, Tu = tumor. "

Figure 7. Biodistribution data for radiolabeled AB35 in tumorbearing nude mice 1 day postinjection. Mice were injected with either 10 pg of T-AB35 (5 mCi/mg) or 20 pg of 6'Cu-CPTA-

AB35 (250-500 pCi/mg). See Figure 5 for legend. 50 I

I

401

1

l L q

CU-67

1-125

1

$ 30 e

20

rp

10 0

81

" 0

1

2

3

4

5

Time (Days)

Th

Ki St Wan

Sp

In

Li

Mu

Tu

Figure 8. Biodistribution data for radiolabeled AB35 in tumorbearing nude mice 4 days postinjection. Injection was as in Figure 7. See Figure 5 for legend.

Figure 6. Tumor to blood accumulation rates for lZbI-AB35and 'Tu-CPTA-AB35 in tumor-bearingnude mice. Data are means f SD of two animals (lSI-AB35) and of three animals (67CuAB35).

per mAb35 and a second experiment with 4.5 chelators per antibody. The biodistribution in tumor-bearing mice with 67CuCPTA-AB35 was followed up to 4 days. Figure 5 shows the organ distribution of the combined experiments with conjugates having 1.36 and 4.5 chelators per antibody. There was no significant biological variation in the in vitro or in vivo behavior of the two conjugates. The amount of radioactivity increases in tumors whereas the activity in normal organs decreases with time after 24 h. The liver activity is surprisingly low with 7.3% ID/g after 24 h and is decreasing with a time constant similar to that of the blood activity, indicating that most of the liver activity corresponds to blood-pool activity. Liver activity in two normal mice measured 2 h postinjection was found to be 12% ID/g liver. The tumor to blood ratio is given in Figure 6. The comparative biodistribution of 1251-labeledand 6'Culabeled antibody is shown in Figure 7 and 8. The data are given for days 1 and 4. The whole-body clearance of both the iodinated and radiocopper-labeled antibody is shown in Figure 9. Assuming a first-order excretion, the two data sets give respective half-lives of 120 and 210 h. Serum studies using SDS gel electrophoresis showed that the circulating Cu-CPTA-AB35 conjugates remain intact with time in vivo (data not shown). Imaging Studies. Due to the favorable y-energy of the isotope, good y-camera images could be obtained. A typical image is shown in Figure 10. The picture was taken 100 h postinjection (animal A) and 102 h postinjection

He

1201

I

I

CU-67

-

*

1

1

1-125

C

rp

60-

71

0

1

2 3 Time (Dayel

4

Figure 9. Whole-body retention of both 1261-AB35and 67CuCPTA-AB35 in tumor-bearing nude mice. Apparent biological half-lives of 120 and 210 h were observed for the respective lSI-

and 67Cu-labeledantibodies.

(animal B). The tumor masses of A were ca. 34 and 150 mg. In the smaller tumor ca. 1500 cpm/mg were measured after excision, whereas the larger tumor measured about 1400 cpm/mg. The tumors of animal B weighed ca. 215 mg (left) and 100 mg (right) with around 1650 and 1700 cpm/mg. The figure represents photographs of unprocessed images. DISCUSSION Copper-67 is one of the attractive radiometals for radioimmunotherapy; it has abundant 0-particles, and the y-energy of 184 keV is suitable for imaging studies (Table I). The maximum 0-energy of 0.57 MeV is close to the corresponding energy of 13*1and appears to be suitable for the treatment of small tumors and metastases. Because of the similarity in @-energy,a comparative biodistribution study with an iodinated and copper-labeled immunoconjugate appeared especially interesting to us. The

SmithJones et al.

Bioconjugate Chem., Vol. 2, No. 6, 1991

420

Figure 10. Anterior y-camera images of two nude mice bearing LoVo tumors 4 days postinjection of 67Cu-CTPA-AB35 with two tumors each (arrows) ranging between 34 and 215 mg. Mice were injected with 20 jig of 67Cu-labeledAB35 (500 pCi/mg).

use of the metallic radionuclide allows the preparation of an instant kit and avoids complicated in-house labor with the iodination. An instant kit procedure affords the use of bifunctional chelators which can be attached to the protein in advance; the conjugate can be purified and stored. Thirteen- and 14-membered tetraazamacrocycles are known to form kinetically inert, well-defined complexes with Cu2+(21,22). Our choice was 4-[(1,4,8,11-tetraazacyclotetradec-1-yl)methyl]benzoic acid as bifunctional ligand. The synthesis is done very simply in one step and can be done in 1day, yielding gram amounts. We reported in a previous publication that the ligand can be conjugated to simple amines, bovine serum albumin, and a monoclonal antibody against a mucinelike carcinoma-associated antigen using activated ester methods and the prelabeling approach (9, 10). In these experiments the Cu2+ protects the amino groups on the macrocycle to prevent dimerization or polymerization of the ligand and allows exclusive reaction with the exogeneous amines. In the uncomplexed ligand some degree of intermolecular reaction is expected in addition to the reaction of the activated ester with the €-amino group of the lysine side chains. However since the macrocycle carries two protons around pH 7 (19),this may provide efficient enough protection to avoid extensive polymerization of the ligand. This appears indeed to be the case, conjugation occurs mildly and the biological integrity of the antibody is preserved. The fast rate of copper complexation by both the CPTA ligand and the protein-CPTA conjugate may be explained by disturbed symmetry of the tetraazamacrocycle due to alkylation of one nitrogen. This same activation of ligand toward complexation was found with the very rigid porphyrins if one nitrogen is substituted with a benzyl group (23) The fast rate of complexation is of particular importance since this can lead to reliable cold kit preparations of a predefined quality, i.e. immunoreactivity and number of binding sites. Excess copper can be removed easily after labeling by use of a small gel-filtration column. Despite the N-functionalization of the ligand, the Cu2+ complex exhibits high stability. It is totally inert in human serum and toward excess EDTA at pH 7.3. Only in strong acid has significant decay been shown. Moreover, the ligand appears to have remarkable selectivity for Cu2+over Zn2+. The presence of large excess of Zn2+had a minor effect on the labeling yield. This fact may facilitate the radiochemical manipulation of radio9

copper solution and render time-consuming removal of Zn2+ contaminations unnecessary. Any chemical separation performed on the initial67Cusolution will ultimately add more carrier copper, the amount depending on the purity of reagents. Comparative kinetic studies of the respective metal complexation rates are in progress. The low pH used for labeling avoids Cu(OH)2 colloid formation, moreover it removes the risk of high unspecific labeling of the protein due to efficient protonation of the amino groups of the antibody. The distribution of the macrocycle at the antibody is random and the biological integrity is preserved after conjugation and labeling. Similar data were reported with somewhat analogous systems by other groups. Franz et al. (16)report antibody labeling with a 1-(3-aminopropyl)4-methyl derivative of cyclam-14. They report preserved immunoreactivity but significant levels of nonspecifically bound copper even after chasing of copper with 1 mM EDTA. It is our impression that this is only due to higher pH than necessary for the labeling and chasing step. Parker’s group (12,13) uses a C-functionalized cyclam14 macrocycle to label the B 72.3 antibody. With only about 0.25 chelators per antibody high immunoreactivity was shown, but labelingefficiencyis only 33% . The labeled conjugate was tested in healthy mice and showed reasonable biodistribution and high in vivo stability. Most advanced as to chemical characterization of a 67Cuchelate and in vivo studies is the work of Meares’ group (14)with the bifunctional chelator 6-(p-nitrobenzyl)TETA, a C-functionalized tetra-N-carboxymethylated derivative of cyclam-14. The Cu2+complex was characterized by X-ray crystal structure (24) and high kinetic stability of the chelate in human serum was demonstrated. The macrocyclic tetra-N-acetate was clearly superior to the usual chelators EDTA and DTPA (15). Thep-(bromoacetamido)benzyl-TETA was conjugated to Lym-1 and stably labeled with e7Cuunder preservation of immunoreactivity. The radiopharmaceutical showed excellent biodistribution in RAJI tumor-bearing mice (14) and is now used in patients (25). The biodistribution data of our copper-labeled conjugates clearly show high tumor uptake and decreasing activity in all organs with time after 24 h, except in tumors. Most important is the extended residence time of radioactivity on the tumor if compared to radioiodinated antibody. Moreover, the liver activity is low and its time course is favorable and decreases with time at approximately the same extent as it disappears from blood. This indicates that processing and handling of the radionuclide and the chelate is more favorable than in the case of antibodies labeled with lllIn and using bifuhctional polyamino polycarboxylates like DTPA and EDTA. This may be due to the extreme kinetic stability of the 6 7 C ~ / C P T A complex and/or the ease of hydrolysis of the peptide bond formed in the conjugation process. ACKNOWLEDGMENT

This work was supported by the Swiss National Science Foundation (Grant No. 31-27739.89),the “Regionale Krebsliga beider Basel”, and Roche Research Foundation. The expert technical help of R. Stolz, D. Tschudin, N. Ulmer, and L. Siegfried-Hertli is gratefully acknowledged. LITERATURE CITED (1) Rainsburry, R. M., Westwood, J. H., Coombes, R. C., Nev-

ille, A. M., Ott, R. J., Kalirai, T. S., McCready, V. R., and Gazet, J. C. (1983) Location of Metastatic Breast Carcinoma

Antibody Labeling wRh Copper-67 by a Monoclonal Antibody Chelate Labeled with Indium-111. Lancet 934-938. (2) Carrasquillo, J. A., Bunn, P. A., Keenan, A. M., Reynolds, J. C., Schroff, R. W, Foon, K. A., Ming-Hus, S., Gazdar, A. F., Mushine, J. L., Oldham, R. K., Perentesis, P., Horowitz, M., Eddy, J., James, P., and Larson, S. (1986) Radioimmunodetection of Cutaneous T-cell Lymphoma with 111-In-Labeled T 101 Monoclonal Antibody. N . Engl. J. Med. 315,673-680. (3) Macklis, R. M., Kinsey, B. M., Kassis,A. I., Ferrara, J. L. M., Atcher, R. W., Hines, J. J., Coleman, C. N., Adelstein, S. J., Burakoff, S. J. (1988) Radioimmunotherapy with AlphaParticle-Emitting Immunoconjugates. Science 240,1024-1026. (4) Baldwin, R. W., and Byers, V. S. (1986) Monoclonal Antibodies in Cancer Treatment. Lancet 603-605. (5) Mach, J. P., Crarrel, S., Forni, M., Ritschard, J., Donath, A., and Alberto, P. (1980) Tumor Localization of Radiolabelled Antibodies against Carcinoembryonic Antigen in Patients with Carcinoma. N . Engl. J . Med. 303, 5-9. (6) Brechbiel, M. W., Gansow, 0. A., Atcher, R. W., Schlom, J., Esteban, J., Simpson, D. E., and Colcher, D. (1986) Synthesis of 1-@-isothiocyanatobenzyl) derivatives of DTPA and EDTA. Antibody labeling and tumor imaging studies. Znorg. Chem. 25, 2772-2781. (7) Meares, C. F., McCall, M. J., Reardan, D. T., Goodwin, D. A., Diamanti, C. I., and McTigue, M. (1984) Conjugation of Antibodies with Bifunctional Chelating Agents: Isothiocyanate and Bromoacetamide Reagents, Methods of Analysis, and Subsequent Addition of Metal Ions. Anal. Biochem. 142, 68-78. (8) Hnatowich, D. J., Layne, U. U., Childs, R. I., Lateigne, D., Davis, M. A., Griffin, T. W., and Doherty, P. W. (1983) Radioactive labeling of antibody: a simple and efficient method. Science 220, 613-615. (9) Studer, M., Kaden, Th. A., and Maecke, H. R. (1990) Reactivity Studies of the Pendant Carboxylic Group in a Macrocyclic Cu2+Complex Towards Amide Formation and Its Use as a Protein-Labelling Agent. Helu. Chim. Acta 73,149-153. (10) Maecke, H. R., Kaden, T., Riesen, A., Ritter, W., and Studer, M. (1989) Some aspects of Antibody Labeling with Metallic Radionuclides. Znt. J. Cancer Suppl. 2, 59, 307. 1) Haskell, C. M., Buchegger, F., Schreyer, M., Carrel, S., and Mach, J. P. (1983) Monoclonal Antibodies to Carcinoembryonic Antigen: Ionic Strength as a Factor in the Selection of Antibodies for Immunoscintigraphy. Cancer Res. 43, 38573864. 2) Morphy, J. R., Parker, D., Alexander, R., Bains, A,, Carne, A. F., Eaton, M. A. W., Harrison, A., Millican, A., Phipps, A., Rhind, S. K., Titmas, R., and Weatherby, D. (1988) Antibody Labelling with Functionalised Cyclam Macrocycles. J. Chem. SOC.,Chem. Commun. 156-157. (13) Morphy, J. R., Parker, D., Kataky, R., Eaton, M. A. W., Harrison, A., Millican, A., Phipps, A., and Walker, C. (1989)

Bioconjugate Chem., Vol. 2, No. 6, 1991 421

Towards Tumor Targeting with Copper-radiolabelled Macrocycle-Antibody Conjugates. J . Chem. SOC.,Chem. Commun. 792-794. (14) Desphande, S. V., DeNardo, S. J., Meares, C. F., McCall, M. J., Adams, G. P., Moi, M. K., and DeNardo, G. L. (1988) Copper-67-labeled Monoclonal Antibody Lym-1, a Potential Radiopharmaceutical for Cancer Therapy: Labeling and Biodistribution in RAJI Tumored Mice. J. Nucl. Med. 29,217225. (15) Cole, W. C., DeNardo, S. J., Meares, C. F., McCall, M. J., DeNardo, G. L., Epstein, A. L., O’Brien, H. A., and Moi, M. K. (1987) Comparative Serum Stability of Radiochelates for Antibody Radiopharmaceuticals. J. Nucl. Med. 28, 83-90. (16) Franz, J., Freeman, G. M., Barefield, E. K., Volkert, W. A., Ehrhardt, G. J., and Holmes, R. A. (1987) Labelling of Antibodies with W u Using a Conjugate Containing a Macrocyclic Amine Chelating Agent. Znt. J. Radiat. Appl. Instrum. Part B 14, No. 5, 479-484. (17) Schwarzenbach, G. (1957) Die komplexometrische Titration, Verlag F. Enke, Stuttgart. (18) L b m l i , U. (1970) Cleavage of Structural Proteins During the Assembly of the Head of Bacteriophage T4. Nature 227, 680-681. (19) Studer, M., and Kaden, Th. A. (1986) One-Step Synthesis of Mono-N-substituted Azamacrocycles with a Carboxylic Group in the Side-Chain, and Their Complexes with Cu2+and Ni2+. Helv. Chim. Acta 69, 2081-2086. (20) Lindmo, T., Boven, E., Cuttitta, F., Fedorko, J., and Bunn, P. A., Jr. (1984) Determination of the immunoreactive fraction of radiolabelled monoclonal antibodies by linear extrapolation to binding at infinite antigen excess. J . Zmmunol. Methods 72,77-89. (21) Leugger, A. P., Hertli, L., and Kaden, Th. A. (1978) Metal Complexes with Macrocyclic Ligands. XI), Ring Size Effect on the Complexation Rates with Transition Metal Ions. Helv. Chim. Acta 61, 2296-2306. (22) Lindoy, L. F. (1989) The Chemistry of Macrocyclic Ligand Complexes, Cambridge University Press, Cambridge. (23) Mercer-Smith, J. A., Roberta, J. C., Figard, S. D., and Lavellee, D. K. (1988) The development of copper-67-labeled porphyrin-antibody conjugates. Targeted Diagnosis and Therapy, Vol. 1, pp 317-323, Marcel Dekker, New York. (24) Moi, M. K., Yanuck, M., Desphande, S. V., Hope, H., DeNardo, S. J., and Meares, C. F. (1987) X-ray Crystal Structure of a Macrocyclic Copper Chelate Stable Enough for Use in Living Systems: Copper(I1) Dihydrogen 6-@-Nitrobenzyl)1,4,8,11-tetraazacyclotetradecane-1,4,8,11-tetraacetate. Znorg. Chem. 26,3458-3463. (25) DeNardo, G., De Nardo, S. J., Kukis, D., Diril, H., Suey, C., and Meares, C. F. (1991) Strategies for Enhancement of Radioimmunotherapy. Nucl. Med. Biol. 13, 359-362.