Abstract
The use of immuno-PET, the combination of PET with monoclonal antibodies (mAbs), is an attractive option to improve tumor detection and mAb quantification. The long-lived positron emitter 89Zr has ideal physical characteristics for immuno-PET, such as a half-life of 3.27 d, which is compatible with the time needed for intact mAbs to achieve optimal tumor-to-nontumor ratios. Thus far, a major limitation in the use of 89Zr has been the lack of suitable methods for its stable coupling to mAbs. In this article, practical protocols for reproducible isolation of highly pure 89Zr and the production of optimal-quality mAb-89Zr conjugates are provided. Methods: 89Zr was produced by a (p,n) reaction on natural yttrium (89Y), isolated with a hydroxamate column, and used for labeling of premodified mAbs. mAbs were premodified with a novel bifunctional derivative of the chelate desferrioxamine B (Df) via a new linker chemistry. To this end, Df was initially succinylated (N-sucDf), temporarily filled with Fe(III), esterified by use of tetrafluorophenol, and then directly coupled to mAbs. Chimeric mAb (cmAb) U36, directed against head and neck cancer, was used for in vitro and in vivo evaluation. The in vitro stability of cmAb U36-N-sucDf-89Zr was assessed in human serum, and its in vivo behavior was evaluated by biodistribution and PET imaging studies in tumor-bearing nude mice. A cmAb U36-Df-89Zr conjugate containing a previously described succinimide ring–thioether unit in the linker was used as a reference. Results: 89Zr was produced in large batches (6.5–13.5 GBq) and isolated with improved radionuclidic purity (>99.99%) and high yield (>94%). The Df-premodified mAbs gave 89Zr-labeling efficiencies of 80% within 30 min, resulting in conjugates with preserved integrity and immunoreactivity. With respect to stability, the novel cmAb U36-N-sucDf-89Zr conjugate appeared to be superior to the reference conjugate. In vivo, the novel conjugate demonstrated selective tumor targeting, and on PET images obtained at 24, 48, and 72 h after injection of this conjugate, small tumors in the range of 19–154 mg were readily visualized. Conclusion: Methods were developed for improved purification of the long-lived positron emitter 89Zr. Moreover, a novel bifunctional Df chelate was synthesized for the reproducible coupling of 89Zr to mAbs. The suitability of such conjugates to detect millimeter-sized tumors in xenograft-bearing nude mice was demonstrated.
Radiolabeled monoclonal antibodies (mAbs) have shown considerable potential for diagnosis and treatment of cancer (1,2). In recently performed clinical radioimmunoscintigraphy and radioimmunotherapy studies at our institute, the potential of the CD44 variant 6–specific mAb U36 for these purposes has been demonstrated (3,4). For the detection of head and neck squamous cell carcinoma (HNSCC), primary tumors as well as metastases, radioimmunoscintigraphy with 99mTc-labeled mAb U36 IgG was found to be as valuable as the conventional imaging techniques CT and MRI, but the detection of tumor deposits smaller than 1 cm appeared to be a problem (3). Introduction of PET might further improve tumor detection because of its high resolution. In addition, PET has potential for quantitative imaging. These features should enable PET to provide proof of principle of antibody targeting and dosimetric determinations before radioimmunotherapy. For this purpose, we are focusing on the coupling of positron emitters to mAbs and the use of these radiolabeled mAbs in immuno-PET.
For a positron emitter to be appropriate for immuno-PET, it has to fulfill several requirements. Its physical half-life has to be compatible with the time needed for a mAb to achieve optimal tumor-to-nontumor ratios. For intact mAbs used for targeting solid tumors, this time is generally 2–4 d. Therefore, commonly used positron emitters such as 11C and 18F (half-lives of 20 and 110 min, respectively) are not suitable for labeling of mAbs. With respect to decay characteristics, the positron-emitting isotope should by preference have no prompt γ-photons with an energy near 511 keV to optimize quantitative accuracy, and the β+-energy should be as low as possible to obtain a high resolution. Besides this, production of the positron emitter should be easy (by preference with a medium-to-small cyclotron), reproducible, and of low cost. Finally, procedures should be available for stable coupling of the positron emitter to the mAb, with maintenance of the in vivo biodistribution characteristics of the latter.
Regarding the required half-life of several days, the 2 positron emitters 124I (half-life, 4.18 d) and 89Zr (half-life, 3.27 d) are most suitable. The low natural abundance of the target material for the production of 124I, however, requires enrichment into 124Te for a (p,n) or a (d,2n) reaction or into 125Te for a (p,2n) reaction. At the current stage of development, production of 124I is too expensive to allow routine clinical application. With respect to the latter, 89Zr can be produced by a (p,n) reaction on 89Y, an element that does not require enrichment because of its natural abundance of 100%. Taking the aforementioned considerations into account, we regard 89Zr (β+ = 22.7%, maximum β+ energy = 0.897 MeV) (Table 1) to be a promising candidate for immuno-PET.
γ-Spectroscopic Analysis of 89Zr and Its Purification by Hydroxamate Column
The bifunctional chelate desferrioxamine B (Df) (Fig. 1) is the ideal chelate for binding of 89Zr because of the stable bond formation of 89Zr with the 3 hydroxamate groups of this chelate (6). For the coupling of this chelate to mAbs, Meijs et al. (7) modified the chelate into N-(S-acetyl)mercaptoacetyldesferrioxamine B (SATA-Df), and the lysine groups of the mAb were modified into maleimide groups. Upon reaction of these 2, a linker was formed containing a succinimide ring–thioether unit. At physiologic pH, the latter unit, however, might be responsible for the release of the chelate from the mAb as has, for example, been shown by Lewis and Shively (8) for their dodecanetetraacetic acid (DOTA)–conjugated mAbs. It can be anticipated that such instability will result in suboptimal delivery of 89Zr to the tumor.
Schematic representation of premodification and postlabeling of mAbs with 89Zr. Step 1 is synthesis of N-sucDf, as described in Materials and Methods. Step 2 is complexation of N-sucDf with Fe(III). N-sucDf (9 mg, 13.6 μmol) is dissolved in 3 mL of 0.9% NaCl, containing 60 μL of 0.1 mol/L Na2CO3 (final pH 6.5–7.0). To this solution, 300 μL (14.8 μmol) of FeCl3 solution (8 mg/mL in 0.1 mol/L HCl) is added. Step 3 is esterification of N-sucDf-Fe. After 10 min, to N-sucDf-Fe solution are added 300 μL (0.36 mmol) of TFP solution (200 mg/mL in MeCN) and 120 mg (0.63 mmol) of solid EDC (final pH 5.8–6.0). After 45 min of incubation, reaction mixture is loaded onto conditioned Sep-Pak C18 cartridge (Waters), followed by washing with 60 mL of sterile water for injection. TFP-N-sucDf-Fe is eluted from Sep-Pak cartridge with 1.5 mL of MeCN. Step 4 is conjugation of TFP-N-sucDf-Fe to mAb. To 1 mL (33 nmol) of mAb solution (5 mg/mL), pH 9.5–9.8 (adjusted with 0.1 mol/L Na2CO3), 20 μL (63 nmol) of TFP ester solution (2.5 mg/mL in MeCN) are added to obtain final chelate:mAb ratio of 1:1 (based on 54% reaction efficiency). After 30 min, 2 times 25 μL of gentisic acid solution (100 mg/mL in 0.32 mol/L Na2CO3) are added to reaction mixture and pH is adjusted to 4.3–4.5 with 4 times 6 μL of 0.25 mol/L H2SO4. Step 5 is removal of Fe(III) from mAb-N-sucDf-Fe. To reaction mixture, 50 μL (3.3 μmol) of an EDTA solution (25 mg/mL) is added and solution is incubated for 30 min at 35°C (final pH 4.3–4.5). After 30 min, EDTA, TFP, iron (as [Fe(III)EDTA]-), and unreacted hydrolyzed ester (N-sucDf) are removed by gel filtration using PD-10 column (eluent: 0.9% NaCl/gentisic acid [5 mg/mL], pH 5): First 2.6 mL (containing reaction volume and first 1.5 mL) are discarded, and modified mAb is collected in next 2 mL. Step 6 is labeling of mAb-N-sucDf with 89Zr. To 600 μL of 89Zr oxalic acid solution (1 mol/L oxalic acid), 130 μL of 0.9% NaCl, 270 μL of 2 mol/L Na2CO3, and 3 mL of 0.5 mol/L HEPES (pH 7.2–7.4) are added, followed by 2 mL (33 nmol) of modified mAb solution (2.5 mg/mL in 0.9% NaCl/gentisic acid [5 mg/mL], pH 5), final pH 7.2–7.4. Reaction volume can be varied provided amounts of oxalic acid, Na2CO3, and HEPES buffer are adjusted accordingly. After 30 min, reaction mixture (6 mL) is divided over 3 PD-10 columns (eluent: 0.9% NaCl/gentisic acid [5 mg/mL], pH 5): First 2.5 mL (2-mL sample volume and first 0.5 mL) are discarded, and radiolabeled mAb is collected in next 3 mL. Bold arches represent -(CH2)2CONH(CH2)5-. Desferal (Df) (Novartis) is term used instead of desferrioxamine B, and Df-Fe is term used to represent corresponding iron(III) complex (ferrioxamine).
In our group, good results have been obtained with a conjugation method based on the reaction of an active 2,3,5,6-tetrafluorophenol-chelate ester (TFP-chelate ester) with the lysine moieties of the mAb, resulting in a stable amide bond as the linker unit. This method gives optimal control over the number of groups conjugated to the mAb and has been used for the production of radioimmuno- and photoimmunoconjugates (9,10).
In the present report, we describe the use of this TFP ester approach for the improved purification of 89Zr as well as for the coupling of a novel bifunctional derivative of Df to mAbs. The in vitro stability of these 89Zr-conjugates and their biodistribution in xenograft-bearing nude mice were determined and compared with succinimide ring–thioether linked Df-89Zr conjugates. The feasibility of visualizing millimeter-sized tumors was demonstrated in HNSCC-bearing nude mice.
MATERIALS AND METHODS
mAbs
Selection, production, and characterization of chimeric mAb (cmAb) U36, as well as of the control murine mAbs (mmAbs) 425 and E48, have been described elsewhere (10).
89Zr Production and Purification
89Zr was produced by a (p,n) reaction on natural yttrium (89Y). For this purpose, an 89Y-target was bombarded with 14 MeV protons (11) for 2–3 h (65–80 μA) while the target support was cooled with water (AVF cyclotron; Philips). 89Y targets were prepared by sputtering an 89Y layer (35 μm; Highways International) on a copper support (Mallinckrodt Medical) as described by Meijs et al. (12). After irradiation, the 89Y layer was slowly dissolved in 4 successive 0.5-mL portions of 1 mol/L HCl (Sigma-Aldrich). Then, 89Zr was oxidized to the IV-oxidation state with 0.1 mL hydrogen peroxide (30% v/v; Mallinckrodt Baker), and 0.22 mL of 12 mol/L HCl was added to set the final HCl concentration at 2 mol/L. After 1 h, 89Y and radionuclidic impurities were removed using a hydroxamate column.
Hydroxamate column material for purification of 89Zr was prepared from Accell Plus CM cation exchange medium (300 Å, 0.35 mmol/g ligand density; Waters). To ensure a reproducible high level of hydroxamate function, a new 2-step ester-mediated method was developed. In the first reaction step, the carboxylic acid groups of the cation exchange medium were esterified using an excess of TFP (Acros Organics) and 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC; Acros Organics). For an optimal conversion of carboxylic acid groups into ester groups, the TFP and EDC were added in 2 successive portions. In short: To 1 g of Accell, suspended in 8 mL of water for injection (Baxter), 75 μL of 3 mol/L HCl, 1 mL of a TFP solution (200 mg/mL in MeCN, 1.2 mmol), and 384 mg of solid EDC (2 mmol) were added (final pH 5.7–6.0). The suspension was mixed end over end for 1 h. Afterward, 105 μL of 3 mol/L HCl were added, along with a new 1-mL portion of TFP and a new 384-mg portion of EDC (final pH 5.7–6.0). After 1 h of mixing, the material was washed with 30 mL of MeCN to remove EDC, its corresponding urea-reaction product, and unreacted TFP.
In the second reaction step, hydroxamate groups were introduced on the resin via the reaction of the ester groups with an excess of hydroxylamine hydrochloride (Aldrich). The hydroxylamine hydrochloride solution was prepared by adding 690 mg (10 mmol) of hydroxylamine hydrochloride to a mixture of 1 mL of 1 mol/L NaOH and 2 mL of MeOH, followed after 5 min by the addition of 1 mL of 1 mol/L NaOH to bring the pH to 5.3–5.4. This solution was added to the esterified resin (final pH 5.1–5.2) and mixed overnight at room temperature. The column material was washed thoroughly with 140 mL of water for injection and 70 mL of MeCN, respectively, and dried in vacuo (freeze-drying). The material can be stored for at least 4 mo without any decrease of 89Zr-binding capacity (extended storage periods are under investigation). For the preparation of a hydroxamate column, an Extract-Clean tube (1.5 mL; Alltech) with a frit placed at the bottom (pore size, 20 μm) was packed with a suspension of 100 mg of hydroxamate column material in 0.9% NaCl (Baxter). After application of the 89Zr-target solution and the eluting solvents, the flow of solvents was initiated by connecting vacuum tubes (Vacutainers without additives; Becton Dickinson) to the column with a needle (0.6 mm × 25 mm).
Before use, a hydroxamate column was equilibrated with 5 mL of MeCN, followed by 10 mL of 0.9% NaCl and, finally, 2 mL of 2 mol/L HCl. After loading of the 89Zr-target solution onto the column, the column was rinsed with 6 mL of 2 mol/L HCl and 6 mL of sterile water for injection, respectively. Under these conditions, 89Zr and the trace amount of 88Zr remained bound to the resin, whereas 89Y and the radionuclidic metal impurities were eluted. The zirconium isotopes were eluted with 5 successive portions of 0.5 mL of 1 mol/L oxalic acid (Aldrich). In general, the oxalic acid fractions contained successively 40%, 40%, 10%, 5%, and 2% of the applied radioactive zirconium.
Preparation of mAb-N-Succinyldesferrioxamine B-89Zr
mAbs were premodified with a novel bifunctional derivative of the chelate Df (N-sucDf) via an amide linkage and subsequently labeled with 89Zr (Fig. 1).
The chelate Df was converted into N-succinyldesferrioxamine B (N-sucDf) in step 1 according to a modified procedure of Herscheid et al. (13). In short: 1.7 g of succinic anhydride (17 mmol; Baker Chemicals) were added to 7.5 mL of pyridine (Sigma-Aldrich Chemie) containing 0.5 g of Df (0.76 mmol; Novartis). The solution was stirred for 24 h at room temperature and added to 120 mL of 0.15 mol/L NaOH. After additional stirring for 16 h at room temperature, the pH was adjusted to 2 with 12 mol/L HCl and cooled for 2 h at 4°C. The precipitate was thoroughly washed with 500 mL of 0.01 mol/L HCl, and the white product was dried in vacuo (freeze-drying).
Coupling of N-sucDf to mAbs and labeling with 89Zr is schematically represented in steps 2 through 6 of Figure 1. In short, the hydroxamate groups of N-sucDf were temporarily blocked with Fe(III) in step 2, N-sucDf-Fe was esterified with TFP in step 3, and TFP-N-sucDf-Fe was coupled to mAbs in step 4. Thereafter, Fe(III) was removed by transchelation to ethylenediaminetetraacetic acid (EDTA) (formation of [Fe(III)EDTA]- (14)) in step 5, and mAb-N-sucDf was labeled with 89Zr in step 6.
Preparation of mAb-SMCC-SATA-Df-89Zr
As reference to the new method for 89Zr labeling (as depicted in Fig. 1), Df was also coupled to mAbs via a thioether linkage as previously described by Meijs et al. (7). In short, the amine group of Df was reacted with N-succinimidyl-S-acetylthioacetate (SATA; Pierce). Modification of mAb-lysine groups into maleimide groups was performed by reaction with sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (sulfo-SMCC; Pierce); the number of maleimide groups per mAb molecule was 0.8–0.9 as determined chemically with Ellman’s reagent (7). mAb-SMCC was purified by the use of a 10DG column (Bio-Rad). After incubation of the purified mAb-SMCC sample with SATA-Df (activated in 0.1 mol/L Na2CO3) for 90 min, unreacted maleimide groups were quenched with an excess of hydroxylamine for 15 min. Purification and labeling of the thus modified mAb with 89Zr and the subsequent purification of the conjugate were performed essentially the same as described for the novel procedure (Fig. 1, steps 5 and 6).
Analyses
A Ge(Li) detector coupled to a multichannel analyzer was used to quantify 89Zr, to monitor 89Zr purification, and (after decay of most of 89Zr) to identify and quantify radionuclidic impurities (Table 1). For quantification of 89Zr activities in a dose calibrator, the 54Mn mode was used, multiplying the displayed amount of activity by a factor of 0.67. Quantification in a γ-counter (LKB Wallac 1282 CompuGamma; Pharmacia) was performed on the 909-keV γ-energy of 89Zr (efficiency, 21.7%). Samples of the 89Zr oxalic acid fractions and of the purified mAb-89Zr solution were analyzed for the presence of any remaining 89Y by particle-induced x-ray emission (PIXE) according to the method described by Vis et al. (15).
High-performance liquid chromatography (HPLC) was performed to monitor the Df derivatives, the chemically modified mAbs, and the radiolabeled mAbs. N-sucDf, N-sucDf-Fe, and TFP-N-sucDf-Fe were analyzed with a Chromspher 5 C18 column (250 × 4.6 mm; Chrompack) with a gradient elution. Solvent A consisted of 10 mmol/L sodium phosphate, pH 6, and solvent B of 100% MeCN. The gradient was as follows (flow rate, 1 mL/min): 5 min of 100% A, linear increase of eluent B to 35% during 25 min, 10 min of 35% B. HPLC analysis of the chemically modified mAbs and the radiolabeled mAbs was performed as described before (10). mAb compounds were monitored by ultraviolet absorbance at 280 nm, Df compounds at 215 nm, and Df-Fe(III) complexes at 430 nm, whereas radiolabeled compounds were monitored either by continuous radioanalytic detection or by measurement of collected fractions. The equipment used has been described before (10).
HPLC analyses of PD-10 column (Pharmacia Biotech) elution profiles were performed as described before (16). Profiles of EDTA, N-sucDf, and TFP (Fig. 1, step 5) were determined with a gradient elution by ultraviolet absorbance at 210 nm. Solvent A consisted of 10 mmol/L sodium phosphate, pH 6, and solvent B of 100% MeCN. The gradient was as follows (flow rate, 1 mL/min): 2 min of 100% A, linear increase of eluent B to 40% during 13 min, 5 min of 40% B (retention times of 2.9, 12.3, and 15.6 min, respectively). Oxalic acid and N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) (Fig. 1, step 6) were analyzed by HPLC with 10 mmol/L sodium phosphate, pH 6, as eluent, a flow rate of 0.4 mL/min (retention times of 6.2 and 7.8 min, respectively), and a wavelength of 210 nm. For determination of the PD-10 column elution profile of [Fe(III)EDTA]-, 59Fe (370 MBq/mL in 0.5 mol/L HCl; Amersham Pharmacia Biotech) was used as tracer and the fractions were counted with a γ-counter.
For the measurement of the serum stability of radioimmunoconjugates, samples were incubated in freshly prepared human serum (1:1 v/v dilution) at 37°C in a humidified incubator maintained at 5% CO2 and 95% air. Radioimmunoconjugates were also incubated in heat-inactivated serum (treated for 40 min at 56°C); in 20% human serum albumin (HSA), pH 7.2 and 9.0; and in 20% HSA, pH 9.0, supplemented with an excess of l-cysteine (9 μmol/mL). At various intervals, samples were taken and analyzed by HPLC.
Radiochemical purity and integrity of the radiolabeled mAbs were monitored by instant thin-layer chromatography (ITLC) (eluent: citric acid, 20 mmol/L, pH 5), sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and analysis by a PhosphorImager (Molecular Dynamics), and immunoreactivity determination as previously described (9,10).
Animal Studies
For the biodistribution and PET studies, female nude mice (athymic nu/nu, 23–32 g, 8–10 wk old; Harlan CPB) bearing human HNX-OE xenografts were used. HNX-OE xenografts were established after subcutaneous injection of HNSCC cell-line VU-SCC-OE (17) and repeated transplantation as xenografts in nude mice. Two to 3 wk after implantation, the mice were used for experiments. All animal experiments were performed according to the U.S. National Institutes of Health principles of laboratory animal care (18) and the Dutch national law “Wet op de Dierproeven” (Stb 1985, 336).
For the biodistribution study, mice (tumor size, 30–200 mg) were anesthetized with ether, and 0.37 MBq cmAb U36-N-sucDf-89Zr (100 μL, 100 μg mAb) or 0.37 MBq of the reference conjugate cmAb U36-SMCC-SATA-Df-89Zr (100 μL, 100 μg mAb) was injected into the retroorbital plexus. The specific activities of the radioimmunoconjugates were 41 MBq/mg and 39 MBq/mg, respectively, and unlabeled mAb was added to bring the total mAb dose to 100 μg per mouse. At indicated times after injection, the mice were anesthetized, bled, killed, and dissected. After blood, tumor, normal tissues, and gastrointestinal contents were weighed, the amount of radioactivity in each was measured in a γ-counter. Radioactivity uptake was calculated as the percentage of the injected dose per gram of tissue (%ID/g). Differences in tissue uptake between groups were statistically analyzed with the Student t test for unpaired data. Differences were considered significant for P < 0.05.
PET studies were performed using a prototype single-crystal-layer HRRT 3-dimensional PET scanner (CTI). The axial field of view of the PET scanner is 25.5 cm, and the transaxial field of view is 31.2 cm; radial and transaxial resolutions are 2.6 mm in full width at half maximum at the center of the field. Transmission scans for attenuation correction were obtained in 2-dimensional mode (consisting of 52 scans, using a 740-MBq 137Cs point source), and emission scans were obtained in 3-dimensional mode. Images were reconstructed by filtered backprojection and consisted of 207 image planes of 256 × 256 pixels, and each voxel equaled 1.21 × 1.21 × 1.21 mm.
Mice bearing HNX-OE xenografts (19–154 mg) were anesthetized with ether and injected with 3.7 MBq cmAb U36-N-sucDf-89Zr (109 MBq/mg, 250 μL, 100 μg mAb). Before scanning, the mice were anesthetized with sodium pentobarbital (75 mg/kg, intraperitoneal) and positioned in the PET scanner. A transmission scan of 380 s was followed by a 60-min emission scan. Mice were scanned and then immediately dissected at 24 h (2 mice), 48 h (2 mice), and 72 h (6 mice) after injection. In addition, 2 mice were scanned 3 times: at 24, 48, and 72 h after injection; after 72 h these mice were dissected. Activity in tumors at the time of scanning was quantified by Ge(Li) and in a γ-counter together with a standard.
RESULTS
Production and Purification of 89Zr
After irradiation, the 89Y-layer was dissolved and the amount of 89Zr was determined, as well as that of contaminating radionuclides (Table 1). The crude yield of 89Zr was between 6.5 and 13.5 GBq (110–190 min of irradiation), with less than 1.3–2.7 MBq (0.02%, Table 1) of radionuclidic impurities. A small amount (0.00015%) of 88Zr was observed, being the result of a (p,2n) reaction on 89Y. The isotope 88Y is formed as a daughter product from 88Zr and possibly from a (p,pn) reaction on 89Y. The isotope 65Zn is formed by a (p,n) reaction on the copper support, whereas 48V, 56Co, and 156Tb are formed on titanium, iron, and gadolinium impurities, respectively, in the 89Y-target. Removal of the bulk nonradioactive 89Y and of the radionuclidic impurities 88Y, 48V, 56Co, 65Zn, and 156Tb (except 88Zr) was achieved with a hydroxamate column (Table 1). As a result, more than 99.99% pure 89Zr was obtained in 1 mol/L oxalic acid, with an overall yield of 97.0% ± 3.3%. PIXE analysis of the isolated 89Zr oxalate revealed the absence of nonradioactive 89Y.
Preparation of mAb-N-sucDf-89Zr
Step 1 in the preparation of mAb-N-sucDf via the TFP ester approach, as depicted in Figure 1, is the carboxylation of the primary amine of Df. After isolation of the product, N-sucDf was obtained in a yield of 68%–73%. After temporary blocking of the hydroxamate groups of N-sucDf with Fe(III) in step 2 (59Fe was used as a tracer to facilitate analytic monitoring) and nearly quantitative esterification (Figs. 2A and 2B), TFP-N-sucDf-Fe was isolated by a Sep-Pak (Waters) procedure in yields of about 80% in step 3. The TFP ester could be stored for at least 8 wk in MeCN at −70°C (data not shown).
HPLC profiles (absorbance at 430 nm and radioactivity after fraction collection) during synthesis of TFP-N-sucDf ester and preparation of mAb-N-sucDf after reaction of N-sucDf with 59Fe-FeCl3 (A), after esterification of N-sucDf-59Fe-Fe (B), after reaction of TFP-N-sucDf-59Fe-Fe with mAb (C), and after removal of iron from mAb-N-sucDf-59Fe-Fe with EDTA (D). Analyses were performed with Chromspher (Chrompack) 5 C18 column (A and B) or Superdex (Pharmacia Biotech) 200 HR column (C and D). Peak 1 = unreacted 59Fe-FeCl3 (retention time, 3.3 min; reaction performed with an excess of 1.1 to 1); peak 2 = N-sucDf-59Fe-Fe (retention time, 16.7 min for A and B and 39 min for C); peak 3 = TFP-N-sucDf-59Fe-Fe (retention time, 25.9 min); peak 4 = mAb-N-sucDf-59Fe-Fe (retention time, 23 min); peak 5 = [59Fe-Fe(III)EDTA]- (retention time, 38 min).
Conjugation conditions selected for the technical protocol comprised the addition of 63 nmol of ester to 33 nmol of mAb in step 4. For cmAb U36, mmAb E48, and mmAb 425, these conditions resulted in a reproducible conjugation efficiency of 54% ± 5% (Fig. 2C) and, consequently, in a reproducible chelate:mAb substitution ratio of about 1:1. After removal of Fe(III) by transchelation to EDTA in step 5, less than 5% iron remained complexed by mAb-N-sucDf after 30 min (Fig. 2D). Isolation of the mAb-N-sucDf on a PD-10 column efficiently removed more than 97% of EDTA, [Fe(III)EDTA]-, TFP, and N-sucDf (data not shown).
In step 6, mAb-N-sucDf was labeled with 89Zr in HEPES buffer (final concentration, 0.25 mol/L) at pH 7.2–7.4. After 30 min at this pH optimum, the amount of 89Zr transchelated from oxalate to mAb-N-sucDf is always more than 80%, even in the presence of 0.1 mol/L oxalic acid. A similar transchelation rate was found in reaction volumes of 1–9 mL, that is, enabling the use of 100–900 μL of 1 mol/L oxalic acid, provided the mAb concentration was more than 0.5 mg/mL. Below pH 6 and above pH 8, less than 15% labeling was obtained under the same conditions. The use of phosphate buffer instead of HEPES strongly affected the kinetics: At pH 7.3, the labeling efficiency was less than 30% after 30 min. Labeling efficiency was not influenced by the amount of 89Zr used, in the range of 0.004–1.5 GBq. Radioimmunoconjugates could be prepared with a specific activity from 10 to 550 MBq/mg mAb. Assessment of the PD-10 column elution profile of unreacted 89Zr, oxalic acid, and HEPES buffer by HPLC revealed that the compounds were quantitatively collected in the waste fractions eluted after the mAb-containing fraction.
Labeling of mmAb E48, mmAb 425, and cmAb U36 resulted in an overall yield of 80% (±6%), a radiochemical purity of more than 95% (determined with ITLC), and immunoreactive fractions of more than 90%. In general, PhosphorImager analysis of the SDS-PAGE gel revealed a major 150-kDa IgG band containing at least 93% of the radioactivity, a minor band with a higher molecular weight (≤3%), a minor band with a lower molecular weight (≤2%), and free 89Zr (≤2%). Gentisic acid was introduced during labeling and storage to prevent deterioration of the mAb integrity by radiation. The chemoprotective potency of gentisic acid (5 mg/mL, pH 5.0) was evaluated with 93 MBq 89Zr per milliliter as the challenging condition. Upon storage at 4°C for 2 h, 95.3% of the radioactivity was present in the 150-kDa IgG band. After 2 d, this band contained 88.7% of the radioactivity when gentisic acid was present and 79.7% when gentisic acid was absent.
Preparation of cmAb U36-SMCC-SATA-Df and Labeling with 89Zr
To arrive at about 1 SMCC group per mAb molecule, a SMCC/mAb molar ratio of 2 was used during the reaction. After reaction with SATA-Df and labeling with 89Zr, the conjugates showed a radiochemical purity of more than 95% and immunoreactive fractions of more than 90%. The overall yield was more than 80%.
Biodistribution in HNSCC-Bearing Nude Mice
Both radioimmunoconjugates, cmAb U36-N-sucDf-89Zr and the reference conjugate cmAb U36-SMCC-SATA-Df-89Zr, were injected into HNX-OE–bearing nude mice (n = 4 per conjugate per time point). The conjugates had only 1 modified lysine group per mAb molecule to avoid impairment of pharmacokinetics due to overmodification (19–21). At 24, 48, and 72 h after injection, the average %ID/g (mean ± SE) of tumor, blood, normal tissues, and gastrointestinal contents was determined (Fig. 3). Between 24 and 72 h after injection, the blood level of cmAb U36-N-sucDf-89Zr decreased from 14.6 ± 1.2 to 12.3 ± 0.7 %ID/g, whereas the tumor level increased from 14.1 ± 1.0 to 26.0 ± 1.9 %ID/g (P < 0.005). Blood clearance of cmAb U36-SMCC-SATA-Df-89Zr was significantly faster: a decrease from 10.7 ± 0.8 to 5.1 ± 0.6 %ID/g (P < 0.005). This enhanced blood clearance was reflected in lower tumor levels (not exceeding 10 %ID/g), lower levels in most organs, but significantly increased levels in colon content at 24 and 48 h after injection (4.8 ± 0.8 vs. 1.7 ± 0.2 %ID/g, P < 0.05, and 5.3 ± 1.8 vs. 1.2 ± 0.2 %ID/g, P < 0.05, respectively) and in ileum content at 24 h after injection (1.5 ± 0.3 vs. 0.5 ± 0.2 %ID/g, P < 0.05).
Biodistribution of cmAb U36-N-sucDf-89Zr (hatched bars, n = 4) and cmAb U36-SMCC-SATA-Df-89Zr (open bars, n = 4). Each conjugate (100 μg of mAb; 0.37 MBq) was injected into retroorbital plexus of HNX-OE–bearing nude mice. At 24 h (A), 48 h (B), and 72 h after injection (C), mice were bled, sacrificed, and dissected, and radioactivity levels (%ID/g ± SEM) of blood, tumor, organs, and gastrointestinal contents were assessed.
In Vitro Serum Stability of cmAb U36-N-sucDf-89Zr and cmAb U36-SMCC-SATA-Df-89Zr
The enhanced blood clearance of the reference conjugate cmAb U36-SMCC-SATA-Df-89Zr, as observed in vivo, was found to be related to the succinimide ring–thioether unit in the linker. Whereas HPLC analysis of cmAb U36-N-sucDf-89Zr showed no loss of radiolabel on incubation during 24 h at 37°C in human serum (Fig. 4A), the reference conjugate cmAb U36-SMCC-SATA-Df-89Zr gave an increased shoulder at 20 min; elution of radioactivity at 26 min, which corresponds to the HPLC retention time of albumin; and a broad radioactivity peak at 36–40 min (Fig. 4B). In a series of experiments, the nature of this phenomenon was evaluated. Incubation in heat-inactivated serum gave the same results as in serum, excluding enzyme-induced instability. To verify the possible transfer of a 89Zr-chelate fragment to serum proteins, HSA (which contains 1 –SH group per molecule) was chosen as a representative protein. Incubation in HSA gave a pattern (data not shown) closely resembling that in Figure 4B. Increasing the pH of the HSA incubation solution to 9 resulted in a more extensive transfer of radioactivity from the monomeric mAb peak to the HSA peak, an enhanced shoulder at 20 min, and an increased peak at 36–40 min (Fig. 4C). The presence of l-cysteine reduced the transfer to HSA, with a concomitant increase in radioactivity at 39 min and a decrease in the shoulder at 20 min (Fig. 4D). Under the same conditions, mAb-N-sucDf-89Zr, as produced by the novel coupling method, did not show these instability phenomena (data not shown).
In vitro stability of cmAb U36-N-sucDf-89Zr and cmAb U36-SMCC-SATA-Df-89Zr, monitored by HPLC. Conjugates were incubated for 24 h at 37°C: cmAb U36-N-sucDf-89Zr in human serum (A); cmAb U36-SMCC-SATA-Df-89Zr in human serum (B); cmAb U36-SMCC-SATA-Df-89Zr in HSA, at pH 9 (C); and cmAb U36-SMCC-SATA-Df-89Zr in HSA, at pH 9 with an excess of l-cysteine (D). Retention times of IgG and HSA are indicated.
PET Studies
To determine the feasibility of visualizing small tumors with the novel radiolabeled cmAb U36-N-sucDf-89Zr, 12 HNX-OE–bearing nude mice were subjected to immuno-PET. In the coronal as well as transaxial images obtained at 24 h (4 mice imaged), 48 h (4 mice imaged), and 72 h (8 mice imaged) after injection, all tumors could be clearly seen as hot spots (Fig. 5). Tumors as small as 19 mg, containing about 17 kBq at the time of the scanning (72 h), could be visualized with the HRRT PET scanner. From the nontarget tissues, only the blood pool in the heart and the liver area (and nose at 24 h) was visible.
HNX-OE–bearing nude mouse, injected with cmAb U36-N-sucDf-89Zr (100 μg of mAb; 3.7 MBq). (A–C) Coronal (upper) and transaxial (lower) PET images were obtained from same mouse at 24 h (A), 48 h (B), and 72 h (C). Image planes are those for which both tumors of same animal were visible. (D) Photographs of imaged mouse and excised tumors (left, 47 mg; right, 45 mg).
DISCUSSION
The assessment of tumor localizations may be improved by combining the selective tumor-targeting properties of a mAb with the excellent sensitivity and contrast resolution of PET. In the present report, we have described a novel method for labeling mAbs with the long-lived positron emitter 89Zr via amide bond (-NH-CO-) coupled Df, taking all the requirements for clinical application into account. Radioimmunoconjugates produced by this method were stable in serum in vitro and showed high accumulation in tumors in nude mice. The feasibility of visualizing small tumors was shown with 89Zr immuno-PET. The required 89Zr was produced in high amounts and was efficiently isolated in a consistent way.
The chelate Df has frequently been used for radiolabeling of mAbs. However, a diversity of problems was met in getting chemical control over the coupling process and the quality of the resulting final product (22–24). In the mAb-Df-89Zr labeling method developed by Meijs et al. (7), which was used as the reference method in the present study, lysine groups of the mAb were modified into maleimide groups, giving a succinimide ring–thioether unit in the linkage on reaction with SATA-Df. Evaluation of cmAb U36-SMCC-SATA-Df-89Zr prepared according to the latter method, however, revealed instability in serum in vitro, with the transfer of the 89Zr-chelate complex to serum proteins. In vivo experiments in tumor-bearing nude mice demonstrated such instability under physiologic conditions. Lewis and Shively (8) described comparable transfer phenomena for their 90Y-DOTA–labeled mAb, containing a linker with 2 succinimide ring–thioether units. They argued a succinimide ring cleavage, possibly Y catalyzed, to be responsible for release of chelate from the mAb. Our in vitro results with HSA strongly suggest that on opening of the succinimide ring, the chelate can break off at either side of the sulfur atom (-S-). When the S-containing Df-89Zr fragment is formed, coupling with HSA takes place, while cleavage at the other side of the S-bond leaves the reactive S-atom at the mAb side, making the mAb susceptible to aggregation (Fig. 4C). These reactions could be counteracted by quenching reactive S-atoms with l-cysteine (Fig. 4D).
To circumvent adverse in vivo instability of the mAb-Df linker, we have here described a novel coupling method based on the use of an active TFP ester. This method creates a chemically stable amide bond between the chelator Df and mAb and gives good control over the number of groups to be coupled to the mAb. To allow conjugation, the amine group present on Df was converted into a carboxylic acid functionality. For the synthesis of the corresponding TFP-N-sucDf ester, it was necessary to temporarily protect the 3 hydroxamate groups of Df against reactions with EDC. Without blocking of the hydroxamate groups, intractable results were obtained that were not further analyzed. Interestingly, blocking with 89Zr did not yield the required TFP ester. Therefore, direct conjugation of mAbs with TFP-N-sucDf-89Zr was not possible. Blocking with iron proved adequate, enabling the development of a postconjugation labeling method. After reaction of this TFP-N-sucDf-Fe ester with mAb, the iron was efficiently detached from Df with the aid of EDTA, under conditions that did not affect the integrity of the mAb.
For the labeling of the premodified mAb with 89Zr, reaction conditions were established that enabled efficient labeling in the presence of oxalic acid, the solution in which 89Zr is isolated. Therefore, the very time-consuming sublimation step, as proposed by Meijs et al. (12) for the removal of the potent 89Zr-chelating agent oxalic acid, is not required any longer. The transchelation of 89Zr from oxalate to Df appeared to take place efficiently within a narrow pH range (7.2–7.4). This observed sharp pH-optimum required the presence of a strong but indifferent buffer. HEPES was found to fulfill the latter requirements.
The need for protection of the mAb against radiation damage has been shown in previous mAb studies (9,25–27). Ascorbic acid as an antioxidant could not be used because this reagent causes detachment of 89Zr from Df by reducing Zr4+ to Zr2+. The presence of gentisic acid during the labeling, purification, and storage proved to be beneficial and, as such, has been made part of the protocol.
Because the novel labeling procedure for 89Zr described in this report uses lysine residues of the mAb for the stable coupling of the Df moiety, the method is applicable to each intact mAb (as confirmed by labeling of mAbs E48 and 425), as well as to mAb fragments or peptides that contain a lysine group. The labeling of the conjugate is based on the postlabeling of a premodified mAb, which makes it possible to centrally produce sterile batches of mAb-N-sucDf, with easy coupling of 89Zr at the user’s site. By our strongly modified 89Zr isolation, it is now possible to obtain large batches of 89Zr (on the order of 10 GBq) in a reproducible way, at low cost, and with a radionuclidic purity of more than 99.99%.
The feasibility of visualizing small tumors with radiolabeled cmAb U36-N-sucDf-89Zr with an HRRT PET scanner was demonstrated in xenograft-bearing nude mice. After injection of the conjugate, HNSCC xenografts in the range of 19–154 mg were detected after 24 h. Target-to-background ratios improved when imaging was performed after prolonged periods, and none of the normal organs showed an adverse high uptake. Similar results were obtained with an ECAT EXACT HR+ positron scanner (CTI) (data not shown).
Sensitive detection of small tumors in nude mice was recently also observed for the positron emitter 64Cu (half-life, 12.7 h) using a 64Cu-DOTA-minibody of T84.66 (molecular weight, 80 kDa), albeit with a high accumulation of 64Cu in the liver (32.4 %ID/g after 4 h) (28). With regard to 124I, a more suitable candidate for use in combination with whole IgG (29,30), recently the capacity of 124I-complementarity-determining region–grafted humanized A33 for detection of colon carcinoma xenografts in nude mice was evaluated (31,32). In this case, high-resolution images of tumors ranging from 200 to 700 mg were obtained 24 h after injection. In ongoing studies at our laboratory, we are evaluating how 124I compares with 89Zr with respect to radiopharmacokinetic behavior, tumor retention, and suitability for PET imaging of small tumors.
A direct practical clinical application of immuno-PET would be its combination with radioimmunotherapy. In this way, imaging can be used for the selection of, for example, 186Re-, 131I-, or 90Y-radioimmunotherapy candidates by confirmation of tumor targeting with additional estimation of radiation delivery to tumor and normal organs. The possibility of using PET with a 89Zr-labeled mAb to predict and quantify targeting of a 90Y-labeled mAb during treatment for more extended periods would be of special interest because of the lack of γ-emission of 90Y.
Because of the encouraging results presented here, the 89Zr-labeled cmAb U36 IgG is currently being evaluated for its capacity to detect primary tumors and metastases in operable HNSCC patients. To this end, results from 89Zr-immuno-PET are being compared with results from CT, MRI, 18FDG PET, and histopathologic evaluation.
CONCLUSION
This study provided practical protocols for reproducible isolation of the long-lived positron emitter 89Zr and its coupling to mAbs via the chelate Df using new linker chemistry. Resulting mAb-89Zr conjugates appeared optimal with respect to radiochemical purity, integrity, immunoreactivity, and stability. Moreover, specific targeting and sensitive detection (with a PET camera) of head and neck cancer xenografts were demonstrated. These achievements justify clinical evaluation of mAb-89Zr conjugates.
Acknowledgments
The authors thank the staff of the BV Cyclotron (VU University) for performing the irradiations, Jan H. Rector (Solid State Physics, VU University) for sputtering 89Y on copper supports, and Fred L. Buijs (Radionuclide Center, VU University) for the PET analyses.
Footnotes
Received Jul. 22, 2002; revision accepted Dec. 11, 2002.
For correspondence or reprints contact: Guus A.M.S. van Dongen, PhD, Department of Otolaryngology/Head and Neck Surgery, VU University Medical Center, De Boelelaan 1117, 1081 HV, Amsterdam, The Netherlands.
E-mail: gams.vandongen{at}VUmc.nl
REFERENCES
In this issue
Jump to section
Related Articles
Cited By...
- 68Ga-DOTATATE Prepared from Cyclotron-Produced 68Ga: An Integrated Solution from Cyclotron Vault to Safety Assessment and Diagnostic Efficacy in Neuroendocrine Cancer Patients
- 89Zr-DFO-Durvalumab PET/CT Before Durvalumab Treatment in Patients with Recurrent or Metastatic Head and Neck Cancer
- Tracking Innate Immune Activation in a Mouse Model of Parkinsons Disease Using TREM1 and TSPO PET Tracers
- DOTA chelation through click chemistry enables favorable biodistribution of 89Zr-radiolabeled antibodies: A comparison with DFO chelation
- Mesothelin/CD3 Half-Life-Extended Bispecific T-Cell Engager Molecule Shows Specific Tumor Uptake and Distributes to Mesothelin and CD3-Expressing Tissues
- Molecular Imaging of Prostate Cancer Targeting CD46 Using ImmunoPET
- The Biodistribution of a CD3 and EpCAM Bispecific T-Cell Engager Is Driven by the CD3 Arm
- 89Zr-pembrolizumab biodistribution is influenced by PD-1-mediated uptake in lymphoid organs
- Detection of incipient pancreatic cancer with novel tumor-specific antibodies in mouse models
- Probody Therapeutic Design of 89Zr-CX-072 Promotes Accumulation in PD-L1-Expressing Tumors Compared to Normal Murine Lymphoid Tissue
- Labeling Antibodies
- Light-Induced Radiosynthesis of 89Zr-DFO-Azepin-Onartuzumab for Imaging the Hepatocyte Growth Factor Receptor
- Total-Body PET and Highly Stable Chelators Together Enable Meaningful 89Zr-Antibody PET Studies up to 30 Days After Injection
- 89Zr-labeled Bispecific T-cell Engager AMG 211 PET Shows AMG 211 Accumulation in CD3-rich Tissues and Clear, Heterogeneous Tumor Uptake
- Fully Automated 89Zr Labeling and Purification of Antibodies
- Molecular Imaging of Radiolabeled Bispecific T-Cell Engager 89Zr-AMG211 Targeting CEA-Positive Tumors
- PET Imaging of Receptor Tyrosine Kinases in Cancer
- In Vivo Characterization of Platinum(II)-Based Linker Technology for the Development of Antibody-Drug Conjugates: Taking Advantage of Dual Labeling with 195mPt and 89Zr
- 89Zr-Lumretuzumab PET Imaging before and during HER3 Antibody Lumretuzumab Treatment in Patients with Solid Tumors
- Theranostics Using Antibodies and Antibody-Related Therapeutics
- Human Epidermal Growth Factor Receptor 3-Specific Tumor Uptake and Biodistribution of 89Zr-MSB0010853 Visualized by Real-Time and Noninvasive PET Imaging
- Molecular Drug Imaging: 89Zr-Bevacizumab PET in Children with Diffuse Intrinsic Pontine Glioma
- A Novel Platinum(II)-Based Bifunctional ADC Linker Benchmarked Using 89Zr-Desferal and Auristatin F-Conjugated Trastuzumab
- Bevacizumab Targeting Diffuse Intrinsic Pontine Glioma: Results of 89Zr-Bevacizumab PET Imaging in Brain Tumor Models
- ImmunoPET with Anti-Mesothelin Antibody in Patients with Pancreatic and Ovarian Cancer before Anti-Mesothelin Antibody-Drug Conjugate Treatment
- Immuno-PET Imaging of CD30-Positive Lymphoma Using 89Zr-Desferrioxamine-Labeled CD30-Specific AC-10 Antibody
- Antibody Positron Emission Tomography Imaging in Anticancer Drug Development
- Immuno-PET and Immuno-SPECT of Rheumatoid Arthritis with Radiolabeled Anti-Fibroblast Activation Protein Antibody Correlates with Severity of Arthritis
- Development of Novel ADCs: Conjugation of Tubulysin Analogues to Trastuzumab Monitored by Dual Radiolabeling
- Everolimus Reduces 89Zr-Bevacizumab Tumor Uptake in Patients with Neuroendocrine Tumors
- Multicenter Harmonization of 89Zr PET/CT Performance
- Effective dose to staff members in a positron emission tomography/CT facility using zirconium-89
- Designing the Magic Bullet? The Advancement of Immuno-PET into Clinical Use
- Placental Growth Factor (PlGF)-Specific Uptake in Tumor Microenvironment of 89Zr-Labeled PlGF Antibody RO5323441
- Review on Production of 89Zr in a Medical Cyclotron for PET Radiopharmaceuticals
- Advances in Immuno-Positron Emission Tomography: Antibodies for Molecular Imaging in Oncology
- Immuno-PET of the Hepatocyte Growth Factor Receptor Met Using the 1-Armed Antibody Onartuzumab
- Molecular Imaging of Tumors with Radioactive Labeled Antibodies from Laboratory to the Clinic
- PET with the 89Zr-Labeled Transforming Growth Factor-{beta} Antibody Fresolimumab in Tumor Models
- Imaging of Prostate Cancer with Immuno-PET and Immuno-SPECT Using a Radiolabeled Anti-EGP-1 Monoclonal Antibody
- 89Zr-Nanocolloidal Albumin-Based PET/CT Lymphoscintigraphy for Sentinel Node Detection in Head and Neck Cancer: Preclinical Results
- Targeting the Internal Epitope of Prostate-Specific Membrane Antigen with 89Zr-7E11 Immuno-PET
- Magnitude of Enhanced Permeability and Retention Effect in Tumors with Different Phenotypes: 89Zr-Albumin as a Model System
- ImmunoSPECT and ImmunoPET of IGF-1R Expression with the Radiolabeled Antibody R1507 in a Triple-Negative Breast Cancer Model
- 89Zr-DFO-J591 for ImmunoPET of Prostate-Specific Membrane Antigen Expression In Vivo
- Cerenkov Luminescence Imaging of Medical Isotopes
- PET of Hypoxia with 89Zr-Labeled cG250-F(ab')2 in Head and Neck Tumors
- Radiation Dosimetry of 89Zr-Labeled Chimeric Monoclonal Antibody U36 as Used for Immuno-PET in Head and Neck Cancer Patients
- Development and Characterization of Clinical-Grade 89Zr-Trastuzumab for HER2/neu ImmunoPET Imaging
- Disparity Between In Vivo EGFR Expression and 89Zr-Labeled Cetuximab Uptake Assessed with PET
- Immuno-PET: A Navigator in Monoclonal Antibody Development and Applications
- Potential of immuno-positron emission tomography for tumor imaging and immunotherapy planning.
- Performance of immuno-positron emission tomography with zirconium-89-labeled chimeric monoclonal antibody u36 in the detection of lymph node metastases in head and neck cancer patients.
- 89Zr as a PET Surrogate Radioisotope for Scouting Biodistribution of the Therapeutic Radiometals 90Y and 177Lu in Tumor-Bearing Nude Mice After Coupling to the Internalizing Antibody Cetuximab
- The Promise of Immuno-PET in Radioimmunotherapy
- Quantitative 89Zr Immuno-PET for In Vivo Scouting of 90Y-Labeled Monoclonal Antibodies in Xenograft-Bearing Nude Mice
- Immuno-PET for Tumor Targeting