Abstract
Introduction
In pretargeted radioimmunotherapy (PRIT), a bifunctional antibody is administered and allowed to pre-localize to tumor cells. Subsequently, a chelated radionuclide is administered and captured by cell-bound antibody while unbound hapten clears rapidly from the body. We aim to engineer high-affinity binders to DOTA chelates for use in PRIT applications.
Methods
We mathematically modeled antibody and hapten pharmacokinetics to analyze hapten tumor retention as a function of hapten binding affinity. Motivated by model predictions, we used directed evolution and yeast surface display to affinity mature the 2D12.5 antibody to 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA), reformatted as a single chain variable fragment (scFv).
Results
Modeling predicts that for high antigen density and saturating bsAb dose, a hapten binding affinity of 100 picomolar (pM) is needed for near-maximal hapten retention. We affinity matured 2D12.5 with an initial binding constant of about 10 nanomolar (nM) to DOTA-yttrium chelates. Affinity maturation resulted in a 1000-fold affinity improvement to biotinylated DOTA-yttrium, yielding an 8.2 ± 1.9 picomolar binder. The high-affinity scFv binds DOTA complexes of lutetium and gadolinium with similar picomolar affinity and indium chelates with low nanomolar affinity. When engineered into a bispecific antibody construct targeting carcinoembryonic antigen (CEA), pretargeted high-affinity scFv results in significantly higher tumor retention of a 111In-DOTA hapten compared to pretargeted wild-type scFv in a xenograft mouse model.
Conclusions
We have engineered a versatile, high-affinity DOTA-chelate-binding scFv. We anticipate it will prove useful in developing pretargeted imaging and therapy protocols to exploit the potential of a variety of radiometals.
Keywords: pretargeting, radioimmunotherapy, DOTA, 2D12.5, antibody
INTRODUCTION
Radioimmunotherapy (RIT) uses radionuclides directly conjugated to antibodies to target tumor-specific antigens, aiming to deliver high doses of radiation to neoplasms while mostly sparing healthy tissues. RIT has shown clinical efficacy in non-Hodgkins lymphoma and other blood cancers. However, RIT has not been successful in the treatment of solid cancer, where the antibody must traverse the tumor vasculature before encountering its target cells. Due to the low permeability coefficient of antibodies across the capillary wall, large concentrations are needed in order to achieve sufficient tumor penetration. At the same time, systemic exposure of healthy tissues to radiation resulting from slow plasma clearance limits the dose that can be safely administered.
PRIT decouples the pharmacokinetics of antibody targeting and radionuclide delivery, and has been shown to increase efficacy and decrease toxicity in both preclinical and clinical models [1-5]. In PRIT, a bifunctional antibody is administered first and allowed to bind to cancer antigen. Because it is not directly attached to a radioactive metal, high doses can be administered. After sufficient tumor uptake of the antibody, a chelated radionuclide is administered and is captured by the pretargeted antibody while the unbound hapten is cleared rapidly from the body.
The first PRIT reagents used the high-affinity binding of streptavidin to biotin for radionuclide capture. However, this approach has significant disadvantages. Streptavidin is a bacterial protein and is consequently immunogenic in humans. In addition, streptavidin localizes to the kidneys, where it appears to remain accessible to bind biotinylated hapten resulting in renal toxicity [6]. Endogenous biotin and the biotinylated hapten compete for streptavidin binding sites [7]. Finally, endogenous biotinidase can cleave biotin from the hapten molecule, requiring the use of biotinidase-resistant linkers [8].
Second generation PRIT approaches employ bispecific antibodies (bsAb) with specificity for both cancer antigen and chelated radionuclide [2]. An approach with a bispecific antibody recognizing an indium EDTA derivative has been studied previously [9]. Because antibodies to metal chelates generally exhibit relatively weak binding, researchers have taken advantage of avidity and developed bivalent haptens to improve tumor retention of the radiometal chelate [2, 10-13]. Another approach to improve hapten tumor retention uses an engineered redox-reactive group in the radiometal chelate to attach covalently to a free thiol in the antibody [14]. However, it remains a challenge to maintain the free thiol during antibody production, purification and delivery.
We present here an alternative approach using DOTA as the radionuclide-carrying hapten. DOTA-metal-complexes are essentially irreversible under physiological conditions and demonstrate higher thermodynamic stability than linear DTPA and EDTA complexes for many metals including gadolinium, yttrium, and lutetium [15-17]. DOTA-gadolinium (DOTA-Gd) has extensive clinical history as a magnetic resonance imaging (MRI) contrast agent and has an excellent safety profile in humans [18]. DOTA-Gd diffuses rapidly, and exhibits rapid renal clearance. A monoclonal DOTA-binding antibody, 2D12.5, was previously isolated from an immunized mouse [3, 19]. 2D12.5 binds to DOTA chelates of all lanthanides with similar nanomolar affinity [3, 20] and to DOTA chelated to indium and copper with weaker affinity [14]. This promiscuity in binding is an unusual property, as other anti-metal-chelate antibodies generally bind only one or two chelated metals with significant affinity [21, 22].
The goal of the present study was to develop a high-affinity antibody to DOTA chelates, starting from the 2D12.5 antibody. A high-affinity DOTA binder will enable the use of simple DOTA as the pretargeting hapten.
MATERIALS AND METHODS
Modeling
The PRIT models developed here are straightforward extensions of two model systems developed and described by Thurber and colleagues [23]. The micrometastasis model uses spherical geometry and assumes diffusion-only transport. The vascularized tumor model uses cylindrical geometry around capillaries. Numerical simulations were performed in MATLAB (The MathWorks, Framingham, MA). Details of the mathematical models are provided in the supplemental materials.
PRIT simulations were performed for a 1 g vascularized tumor and a 400 μm diameter micrometastasis assuming a 70 kg human with 3.5 L of blood volume. An IgG-like bispecific antibody is given as a bolus dose of 7 μmol at time zero. The hapten is given as a bolus dose of 350 nmol with 5 GBq initial activity at 72 h. The model implements a clearing/blocking step 24 h before hapten dosing, in which 99.9% of bsAb hapten binding sites are blocked in the blood compartment. After hapten dosing, unbound hapten concentration in the blood is calculated as the initial hapten concentration minus hapten binding sites in the blood from unblocked residual antibody. The model assumes a 90Y radionuclide that has a residualization half life of 120 h after cellular internalization.
DOTA Complexes
Stock solutions of DOTA and S-2-(4-Aminobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid (DOTA-Bn) where purchased from Macrocyclics (Dallas, TX) and dissolved in 0.4 M sodium acetate, pH 5.2.
DOTA-Bn-biotin (Figure 2) was synthesized by dissolving Amine-PEG3-Biotin purchased from Thermo Scientific (Rockford, IL) and S-2-(4-Isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane-tetraacetic acid (p-SCN-Bn-DOTA) purchased from Macrocyclics in dimethyl sulfoxide (DMSO) with a 10 fold molar excess of triethylamine. The reaction mixture was vortexed at room temperature for 3 h, and then purified by high performance liquid chromatography (HPLC). HPLC purification was performed on a C-18 reverse-phase column (Agilent Model 1100 HPLC, 1 × 25 cm, buffer A = 0.05% trifluoroacetic acid (TFA), buffer B = 0.0425% TFA in 80% acetonitrile, 2 – 100% B gradient for 98 min). Flow through was monitored by absorbance detection at 280 nanometers. Fractions containing DOTA-Bn-biotin were confirmed using matrix assisted laser desorption instrument time of flight (MALDI-TOF) mass spectrometry (Applied Biosystems Model Voyager DE-STR). Chemical purity was assessed by analytical HPLC (Agilent Model 1100 HPLC, 2.1 × 150 mm, buffer A = 0.05% TFA, buffer B = 0.0425% TFA in 80% acetonitrile, 2-100% B gradient for 45 min). DOTA-Bn-biotin concentration was determined using a biotin quantitation kit (Thermo Scientific) following the manufacturer’s instructions.
Fig. 2. Chemical structures.
Chemical structures of the DOTA variants used in this study with trivalent metal cations.
(+)-(2S)-2-(4-Aminobutyl)-1,4,7,10-tetra-azacyclododecane-1,4,7,10-tetrayltetra-acetic acid (DOTA-alkyl) has been synthesized previously [24]. Here, it was synthesized following the procedure of Takenouchi et al. [25] starting with the compound H-Lys(Boc)-OMe (Bachem, E-1620). H-Lys(Boc)-OMe was treated stepwise with methyl bromoacetate and diethylenetriamine to obtain tert-butyl 4-(3,12-dioxo-1,4,7,10-tetraazacyclododecan-2-yl)butylcarbamate. Borane·THF complex [26] was used to reduce the carboxylic amides followed by trifluoroacetic acid Boc deprotection to obtain 4-(1,4,7,10-tetraazacyclododecan-2-yl)butan-1-amine. This compound was subsequently reacted with 6-((6-((biotinoyl)amino)hexanoyl)amino)hexanoic acid, sulfosuccinimidyl ester, sodium salt (biotin-xx, SSE) purchased from Invitrogen (Carlsbad, CA) in DMSO with a 10 fold molar excess of triethylamine for 3 h vortexing at room temperature to form DOTA-alkyl-biotin (Figure 2). In all synthesis steps compounds were purified by HPLC and their identity and purity confirmed by mass spectrometry with a Waters (Milford, MA) LCT electrospray time-of-flight (ES-TOF) liquid chromatography mass spectrometry (LC/MS) or by MALDI-TOF as described above.
Metal complexes of each DOTA derivative (see Figure 2 for chemical structures) were prepared as follows. Yttrium nitrate hexahydrate, lutetium (III) chloride hexahydrate, indium (III) chloride, gallium (III) nitrate hydrate, and gadolinium (III) chloride hexahydrate were purchased from Sigma (St. Louis, MO) and prepared as stock solutions in 0.4 M sodium acetate pH 5.2. To a 2 mM (for DOTA and DOTA-Bn) or 400 μM (for DOTA-Bn-biotin) solution of the chelating agent, a 5-fold molar excess of the metal stock solution was added and chelated by overnight rotation at room temperature. The pH was adjusted to 7 with 10 M NaOH and the complex was diluted with phosphate buffered saline with 0.1% bovine serum albumin (PBSA) to a final concentration of 1 mM (for DOTA and DOTA-Bn) or 200 μM (for DOTA-Bn-biotin). For gadolinium chelates, an identical metal loading procedure was used except that the complexation reaction took place at 80°C for 12 h in a thermocycler. Complete complexation of the chelator was confirmed by LC/MS using a 75 μm × 150 mm C18 column (Magic C18 from Michrom Bioresources).
Kinetic Characterization
KD Measurements for DOTA-Bn-biotin-metal
Equilibrium dissociation constants (KD) for binding of yeast surface-displayed scFv to biotinylated DOTA complexes at 37°C were determined in triplicate by titration as described by Chao et al. [27]. Briefly, yeast expressing an scFv clone on their surface were grown, washed with PBSA and incubated with various concentrations of DOTA-Bn-biotin-metal long enough to allow for at least a 95% approach to equilibrium. Bound DOTA-Bn-biotin-metal was detected by fluorescence activated cell sorting (FACS) after incubation with streptavidin-phycoerythrin (Invitrogen, Carlsbad, CA). Generally, 5 × 105 induced cells were used for each concentration point. When antigen concentrations less than 10 pM were assayed, the titration was performed with 2.5 × 104 induced and 7.5 × 105 non-induced cells to ensure antigen excess over the scFv without requiring impractically large volumes. The addition of non-induced cells aids pelleting during centrifugation [28]. When antigen concentrations greater than 100 nM were used, non-specific antigen binding to the yeast surface was taken into account. Yeast expressing an irrelevant scFv on their surface were treated in the same manner as the yeast displaying the scFv of interest, and mean total phycoerythrein fluorescence (MFUtot) due to non-specific binding was measured by flow cytometry and averaged over three replicates. This value was subtracted from the MFUtot for the yeast of interest, and the data was fit by least-squares regression.
KD Measurements for DOTA-metal and DOTA-Bn-metal
To determine the KD for scFv binding to nonbiotinylated haptens, the above protocol was modified to a competition-based assay as follows. After determining the KD for scFv binding to DOTA-Bn-biotin-Y, a titration was set up with 100 pM DOTA-Bn-biotin-Y, 2.5 × 105 cells per tube, and varying concentrations of the nonbiotinylated complex. Incubation, staining, and flow cytometry analysis was the same as that for biotinylated antigen. MFUtot as a function of the concentration of the nonbiotinylated antigen ([Ag]), normalization constant (MFUrange), minimal total mean fluorescence (MFUmin), KD for DOTA-Bn-biotin-Y (KD,biot), DOTA-Bn-biotin-Y concentration ([Agbiot]), and KD for the antigen of interest (KD) follows this modified equation:
The data was fit by least squares regression as before, varying MFUmin, MFUrange and KD.
Dissociation Kinetics
To determine the dissociation rate, koff, for DOTA-Bn-biotin complexes, cells were induced and washed as above, and 1 × 107 cells were incubated in 1 mL PBSA with 1 nM DOTABn-biotin-metal for 1 h to reach saturation. Subsequently, the yeast were washed with 1 mL PBSA, resuspended in 1 mL PBSA with 1 μM (excess) non-biotinylated antigen as competitor and split into 100 uL aliquots. These aliquots were incubated at 37°C for different lengths of time, then washed with cold PBSA and left on ice. All samples were simultaneously stained with streptavidin-phycoerythrein for 10-20 min and analyzed by flow cytometry. The data was fit to the following equation by least squares regression, varying MFUmin, MFUrange and koff:
For nonbiotinylated antigens, the procedure was identical except that initial saturation was with the nonbiotinylated antigen and DOTA-Bn-biotin-metal was used as competitor. The data followed the expression
Affinity maturation
The 2D12.5 scFv served as our starting point and was subjected to nine rounds of directed evolution by random mutagenesis and subsequent selection for improved binding using yeast surface display as described by Chao and colleagues [27] and adapted as follows.
Mutagenesis
To counteract the mutational bias of error-prone PCR, mutagenesis at each round was also performed with the Mutazyme mutagenesis kit (Agilent, Santa Clara, CA) according to the manufacturer’s instructions, and the resulting mutagenized DNA was pooled with that obtained by error-prone PCR. All other steps were carried out as described [27].
Selection
Each round of mutagenesis resulted in a library size of 0.5–4 ×107 and was sorted 2-3 times by flow cytometry for improved binders. At least five times the estimated library diversity was labeled for cell sorting. Staining was performed by equilibrium incubation at a biotinylated DOTA-Y concentration of approximately 1/3 of the average KD of the previous library (in early rounds) or by saturation with antigen followed by dissociation for 2-3 dissociation half-times (in later rounds), and subsequent labeling with streptavidin-phycoerythrein (Invitrogen). To label for full-length scFv expression, the yeast were also stained with a mouse anti-HA (clone 12CA5, Roche Applied Science, Basel, Switzerland) or a mouse anti-cmyc (clone 9e10, Covance, Princeton, NJ) primary antibody and a goat anti-mouse Alexa-647 (Invitrogen) secondary antibody. Yeast expressing the best 0.01–0.1% of binders were collected. Periodically, the antigen was alternated between DOTA-Bn-biotin-Y and DOTA-alkyl-biotin-Y.
Disulfide stabilization and glycosylation knockout
The N-linked glycosylation site in the heavy chain of the scFv was removed and a disulfide bond between the heavy and light chain was introduced during the seventh mutagenesis of the affinity maturation. This was accomplished by introducing through PCR site-directed mutagenesis the mutations N88E or N88D, Q111C, and L179C (numbering corresponds to the scFv sequence; Figure 3C).
Fig. 3. Mutations resulting from affinity maturation.
Mutations accrued through eight rounds of affinity maturation are highlighted in blue in the x-ray crystal structure of the 2D12.5 Fab (A) and a magnified view of the binding pocket (B). Panels A and B were generated with PyMol based on the research collaboration for structural bioinformatics (RCSB) protein data bank entry 1NC2 [38]. Panel C shows the sequence alignment of the 2D12.5ds and C8.2.5 scFvs. Residues within 5 Angstroms of the hapten from the 2D12.5 crystal structure are highlighted in yellow; the (Gly4Ser)3 linker is highlighted in grey. Note that the residue numbering is different from that for the crystal structure of the Fab.
Selection of clones
Individual clones were isolated by transforming XL-1 blue chemically competent E. coli (Agilent) with plasmid DNA isolated from the yeast library (Zymoprep II Kit, Zymo Research, Orange, CA) and plating on agar plates containing ampicillin. Individual colonies were picked and grown in liquid medium overnight and plasmid DNA was isolated using a Miniprep kit (Qiagen, Valencia, CA). The plasmid DNA was sequenced and transformed back into yeast with the EZ yeast transformation kit (Zymo Research). Clonal yeast cultures were grown and their kinetic parameters determined.
Bispecific Antibody Construction
An IgG-like bispecific antibody that binds to CEA and DOTA was engineered from a high-affinity Sm3e antibody [29] and the C8.2.5 scFv as described [30]. An analogous bispecific antibody with the wild-type 2D12.5ds scFv was also constructed by ligating the 2D12.5ds scFv into the light chain plasmid between the Nhe1 and Sal1 restriction sites. The bispecific antibodies were produced in transient HEK293 culture and purified by protein A chromatography as described [30].
Radiolabeling
The HPLC/mass spectrometry platform used for purification of radioactive small molecules has been described in detail [31, 32]. DOTA-Bn was dissolved at 5 mM in ammonium acetate pH 5.5. 1-2 mCi 111InCl3 (Cardinal Health, Dublin, OH) were added to the metal chelate and incubated for 1 h at 90°C. The radiolabled compound was purified by RP-HPLC with gamma detection on a 4.6 × 75 mm Symmetry C18 column using a linear gradient from 0% to 100% B over 15 minutes, at a flow rate of 1 mL/min, where A = water and B = acetonitrile with 0.1% formic acid.
Animal Model
All animal handling was performed in accordance with Beth Israel Deaconess Medical Center Animal Research Committee guidelines. LS174T human colorectal carcinoma cells (CL 188) were obtained from American Type Culture Collection and maintained under standard conditions and confirmed to be negative for mycoplasma and mouse pathogens by the Yale Virology Lab. Xenografts were established in 5-6 week-old male NCRU-nu/nu mice (Taconic Farms) by subcutaneous injection of 1-2 × 106 LS174T cells into the flank of the mouse. After 8-10 days, tumors were 0.1 – 0.5 g in size. 30 ug of bispecific antibody was injected intravenously followed 24 h later by intravenous injection of 1.3 pmol 111In-labeled DOTA-Bn. Mice were euthanized 24 h later by intraperitoneal injection of pentobarbital, a method consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association. Tumors were resected, washed in PBS, weighed, and counted with a model 1470 Wallac Wizard (Perken Elmer, Wellesley, MA) 10-detector gamma counter. A Students t-test was used to examine the differences between the experimental groups.
RESULTS
Mathematical Modeling
We mathematically modeled the effect of DOTA-binding affinity on the delivery of ionizing radiation in pretargeted radioimmunotherapy. Two mathematical models were implemented that simulate PRIT based on previously validated models, one that simulates antibody distribution in vascularized tumors and the other in micrometastases [23]. These two types of tumors were considered separately due to different modes of transport. For micrometastases, antibody and hapten diffuses into the tumor mass from the surrounding interstitial fluid. While there may be some transport from surrounding interstitial fluid into the edges of large vascularized tumors, the majority of antibody and hapten transport occurs across the tumor vasculature.
We extended both models to account for hapten kinetics, assuming an IgG-like bsAb with an affinity of 100 pM to CEA [30], and a 15 h internalization half-life [33]. We used a bsAb blood concentration of 2 μM as an input variable. We expect this initial concentration to essentially saturate the antigen binding sites for vascularized tumors from both modeling predictions and from Fenwick and colleagues [34] demonstrating that antibody doses of several hundreds of micrograms or more are required to obtain saturation in a mouse xenograft model. PRIT model timing and dosing parameters are similar to that of a recent Phase II human trial [35] with bsAb dosing at time 0, followed by a clearing/blocking step at 48 h and hapten dosing at 72 h, with an initial hapten blood concentration of 100 nM. Our model predicts that this hapten dose will saturate the pretargeted bsAb binding sites in the vascularized tumor. Note that bsAb and hapten doses are orders of magnitude above those predicted to saturate micrometastases. The model assumes a 70 kg man and 2-compartment pharmacokinetic parameters for antibody and hapten. A detailed description of all model parameters is provided in the supplemental materials.
PRIT model simulations were run, varying the hapten dissociation rate while keeping the association rate constant. We simulated hapten concentration in the tumor as a function of time and total cumulative activity assuming a 90Y radionuclide over a time interval of 15 days. Hapten retention in vascular tumors (Figure 1A) and micrometastases (Figure 1B) were predicted over a hapten KD range of six orders of magnitude. The half-time of residualization of DOTA chelates after internalization is assumed to be 120 h (estimated from [36]). We looked at the effect of varying the association rate while maintaining a constant KD and found no significant difference in hapten retention for typical hapten association rates (5×105 – 5×107 M−1s−1), demonstrating that the relevant parameter is KD.
Fig. 1. Hapten retention in tumors as a function of hapten binding affinity.
PRIT simulations were performed assuming a vascularized tumor (A) and a small micrometastasis (B) with human pharmacokinetic parameters. The hapten concentration in the tumor as a function of time was plotted for various dissociation constants (indicated by arrows). The cumulative activity (calculated over a time interval of 15 days) for a 90Y radionuclide is tabulated for various KD values and also for a theoretical koff equal to zero. Cumulative activity units are gigabecquerel seconds (GBq s) for the vascularized tumor and megabequerel seconds (MBq s) for micrometasteses. The cumulative activity values for the vascularized tumor are larger than those for the micrometastases due to the size difference between the two tumor models (1 g versus ~0.03 mg, respectively).
For the aforementioned PRIT conditions, we predict that a hapten KD greater than 100 pM will allow significant hapten retention for both vascularized tumors and micrometastases. The affinity of the bsAb to the tumor antigen has a negligible impact on hapten retention for bsAb/antigen affinites of ~10 nM or better (data not shown). This is due to lower capillary permeability of large proteins such that they are able to re-bind repeatedly and remain in the tumor even at modest affinities [37].
Affinity maturation
We affinity matured the 2D12.5 antibody fragment against biotinylated DOTA-Y by directed evolution. We used biotinylated DOTA-Y in order to probe binding using a streptavidin-fluorophore secondary label and flow cytometry, as DOTA-Y itself possesses no intrinsic fluorescent properties. The gene encoding the variable domains of the 2D12.5 DOTA-binding antibody in an scFv format (Figure 3C) was synthesized from its published sequence [38]. The scFv was subsequently subjected to nine rounds of affinity maturation. Yeast expressing 2D12.5 scFv variants were labeled for expression with either DOTA-Bn-biotin-Y or DOTA-alkyl-biotin-Y (Figure 2) followed by streptavidin-phycoerythrein and sorted by flow cytometry to select the highest affinity clones. The antigen was periodically switched to minimize selection of variants with mutations that conferred binding improvement to the linker region. During the seventh mutagenesis, we introduced an intramolecular disulfide bond between the heavy and light variable regions of the scFv [39] and removed the N-linked glycosylation site in the heavy chain. These additional mutations may improve stability and result in simpler downstream processing of the scFv.
Sequences and kinetic constants were determined for several clones from libraries 8.2 (8 rounds of mutagenesis followed by 2 sorts) and 9.3 (9 rounds of mutagenesis and 3 sorts). All clones from library 9.3 had lost the disulfide bond between the heavy and light chain and were consequently discarded. Of the clones from library 8.2, C8.2.5 retained the disulfide bond and bound most tightly to DOTA-Bn-biotin-Y.
Improved mutant C8.2.5
The sequence (Figure 3C) of the high-affinity C8.2.5 scFv differs from 2D12.5ds (the original 2D12.5 scFv with the addition of the intramolecular disulfide bond and removed glycosylation site) at 15 amino acid positions. The spatial distribution in the crystal structure of wild-type 2D12.5 is depicted in Figure 3A and 3B. Only one mutation, N53(L)H (numbering corresponds to the 2D12.5 antigen-binding fragment (Fab) for which the crystal structure was determined [38]), occurred within 5 Angstroms of the bound hapten, indicating that most mutations enhanced affinity via subtle structural perturbations remote from the binding interface.
Kinetic characterization
The kinetic properties of both 2D12.5ds and C8.2.5 were characterized and are summarized in Tables 1 and 2 and in Figures 4 and 5. The affinity of the scFv to DOTA-Bnbiotin-Y was improved by 3 orders of magnitude, from nanomolar to single-digit picomolar. The dissociation half-time for DOTA-Bn-biotin-Y increased from 5.5 min for 2D12.5ds to just over 5 hours for C8.2.5 (Table 2 and Figure 5).
TABLE 1.
Equilibrium dissociation constants for yeast surface-displayed scFvs bound to DOTA complexes
scFv | Hapten | Metal | KD* | n | |
---|---|---|---|---|---|
2D12.5ds | DOTA-Bn-Biotin | Y | 3.7 ± 3.6 | nM | 3 |
C8.2.5 | DOTA-Bn-Biotin | Y | 8.2 ± 1.9 | pM | 3 |
DOTA-Bn | Y | 15.4 ± 2.0 | pM | 3 | |
Lu | 10.8 ± 2.5 | pM | 3 | ||
Gd | 34.0 ± 5.3 | pM | 2 | ||
In | 1.01 ± 0.04 | nM | 2 | ||
Ga | 52 ± 12 | nM | 2 | ||
DOTA | Y | 103 ± 35 | pM | 3 | |
Lu | 390 ± 14 | pM | 2 | ||
Gd | 149 ± 6 | pM | 2 | ||
In | 23.7 ± 3.7 | nM | 2 | ||
Ga | 216 ± 26 | nM | 2 | ||
|
KD given as mean ± SD.
TABLE 2.
Dissociation half-lives for yeast surface-displayed scFvs bound to DOTA complexes
scFv | Hapten | Metal | Dissociation half-life* |
---|---|---|---|
2D12.5ds | DOTA-Bn-Biotin | Y | 5.5 ± 1.3 |
C8.2.5 | DOTA-Bn-Biotin | Y | 302 ± 13 |
DOTA-Bn | Y | 53.1 ± 2.3 | |
DOTA | Y | 3.5 ± 0.7 | |
Lu | 3.8 ± 0.4 | ||
|
Calculated as ln(2)/koff and given in minutes as mean ± SD for n=3 experiments.
Fig. 4. Metal specificity of improved mutant.
Competition equilibrium isotherms for C8.2.5 scFv displayed on the surface of yeast for binding to DOTA-metal (A) and DOTA-Bn-metal (B) complexes measured with a competition-based assay. It should be noted that 100 pM DOTA-Bnbiotin-Y is the detected label. The binding constant for the competitor will follow a Cheng-Prusoff relationship as described in Materials & Methods.
Fig. 5. Improved mutant dissociation curves to DOTA variants.
Dissociation curves for C8.2.5 scFv and yttrium complexes of different DOTA variants. For the nonbiotinylated haptens, the fractional saturation plotted is 1-(fractional saturation with competitor (DOTA-Bn-biotin-Y)).
The high-affinity clone C8.2.5 binds DOTA-Bn-biotin-Y, DOTA-Bn-Y, and DOTA-Y with equilibrium dissociation constants of 8.2 ± 1.9 pM, 15.4 ± 2.0 pM, and 103 ± 35 pM, respectively. The addition of the benzene ring and biotin moiety may change the charge distribution of the DOTA epitope, altering the affinity. It is also possible that there are some binding interactions between the scFv and the benzene ring and biotin linker region. The affinity differences between these various yttrium chelates are reflected in their dissociation half-lives (Table 2 and Figure 5).
DOTA complexes of lutetium and gadolinium were bound by C8.2.5 similarly to those of yttrium (Table 1 and Figure 4). The high-affinity scFv also binds indium and gallium chelates with nanomolar affinity. All DOTA-metal chelates were bound by C8.2.5 with about an order of magnitude weaker affinity than the respective DOTA-Bn metal chelate.
Complete metal complexation of DOTA was confirmed using LC/MS as described in the materials and methods. However, LC/MS may not be able to distinguish the thermodynamically stable complex from intermediates that may form [40]. To confirm that the kinetic characterization described above was not influenced by the presence of intermediate complexes, the DOTA-Bn metal complexes were incubated for an additional 12 h at 80°C; the measured binding affinities of these DOTA-Bn complexes to C8.2.5 were essentially the same as those described above (data not shown).
Analysis of tumor targeting in vivo
The C8.2.5 scFv was engineered into an IgG-like bispecific antibody as a C-terminal fusion to the light chain of a CEA-targeting Sm3e IgG [30]. The bispecific antibody retains parental affinities of both the C8.2.5 scFv and Sm3e IgG, with an affinity of ~100 pM to CEA-expressing LS174T cells, and also exhibits IgG-like blood clearance and tumor targeting in vivo [30]. Mice injected with 30 ug C8.2.5 bispecific antibody 24 hours prior to 111In-DOTA-Bn administration exhibit significantly greater tumor uptake of the hapten 24 hours later, compared to an analogous bispecific antibody containing the wild-type 2D12.5ds scFv and 111In-DOTA-Bn only (Figure 6), demonstrating improved retention of the 111In-DOTA-Bn at the site of the tumor for the affinity matured scFv.
Fig. 6. Comparison of high-affinity and wild-type scFv pretargeting in vivo.
24 h post injection hapten tumor retention in xenograft mice pretargeted with C8.2.5 bispecific antibody compared to 2D12.5ds bispecific antibody and 111In-DOTA-Bn hapten alone. n = 3, error bars are s.d.; * P < 0.05.
DISCUSSION
In this study, we have used a mathematical model of PRIT to predict the maximum achievable cumulative activity in the tumor as a function of antibody binding affinity to the radiometal chelate. While other mathematical models of PRIT have been developed [41, 42], for the present work we exclusively used published measured parameters without any curve fitting. Hapten retention at the site of the tumor depends on a large number of factors, including the hapten dissociation rate, rebinding of dissociated hapten, diffusion rate, capillary permeability, hapten pharmacokinetics in the blood, and antigen internalization. For therapeutic applications, it is desired that the hapten is retained at the site of the tumor until the majority of radioactive decays have occurred.
From predicted hapten tumor concentrations, we have calculated tumor cumulative activities for a 90Y radionuclide. It would be straightforward to further calculate doses to the tumor assuming spherical geometry and published S-values for 90Y [43]. However, we have not presented dose estimates, as we expect the model to provide qualitative trends but due to significant heterogeneity in many of the model parameters, we do not expect accurate predictions of true clinical doses.
Motivated by mathematical modeling, we used directed evolution and yeast surface display to affinity mature the 2D12.5 scFv to biotinylated DOTA-Y with a goal of low picomolar affinity. The resulting high-affinity clone not only binds DOTA-Y chelates but also lutetium and gadolinium chelates with low picomolar affinity and indium and gallium chelates with nanomolar affinity. While the 2D12.5 Fab binds the DOTA chelates of all lanthanides with similar nanomolar affinity and indium DOTA chelates with micromolar affinity [44], it was not a priori obvious that this versatility would persist throughout the affinity maturation. The engineered high-affinity clone possesses mutations that significantly improve binding to several metal chelates, despite selective pressure only toward the yttrium chelate. This promiscuous binding is highly advantageous in further preclinical and clinical development of a bispecific agent containing the C8.2.5 scFv for pretargeting, as it allows metals (with their different radioactive properties) to be varied and used with the same bsAb.
Interestingly, the metals whose complexes are bound by C8.2.5 most tightly – yttrium, lutetium, and gadolinium – are chelated by DOTA with identical coordination chemistry, having a coordination number of 9 with one crystalline water molecule in the complex [45-47]. Indium and gallium, in contrast, are chelated by DOTA with coordination numbers of 8 and 6, respectively [48, 49]. This leads us to hypothesize that C8.2.5 may bind all lanthanides with low-picomolar affinities, as they are known to form nonacoordinate chelates with DOTA that bind with similar nanomolar affinities to wild-type 2D12.5 [20] . This suggests further potential biotechnological applications for C8.2.5, exploiting for example the luminescence of Tb and Eu.
The nature of the accumulated mutations in C8.2.5 is similar to that previously observed for engineering extremely high-affinity binders from a moderate binder to the same antigen [50]. Most mutations occurred away from the binding interface, many of them at second-shell residues, and are conservative with respect to physicochemical properties. This indicates that enhanced binding stems from slight structural adjustments, rather than novel direct binding contacts to the hapten. Determining a crystal structure for C8.2.5 bound to DOTA-Y would enable a more detailed analysis of the binding interactions.
Based on our model, we predict that our high-affinity scFv will effect approximately 4-fold higher cumulative activities in vascularized tumors and 8-fold higher cumulative activities in micrometastases when compared to the wild type 2D12.5 antibody (for 90Y radionuclides) for the model conditions. We also predict that any further improvement in affinity to yttrium chelates would result in no more than a two-fold increase in cumulative activity in micrometastases and would have no significant effect in vascularized tumors. Four-fold and eight-fold improvements in cumulative activity may seem modest for an affinity improvement of three orders of magnitude. However, these predictions are based on the selection of a highly expressing cancer antigen and saturating bsAb and hapten doses (similar to optimized doses tested previously in pretargeted radioimmunotherapy studies in the clinic [35]). While sub-saturating bsAb doses, lower antigen densities, and smaller hapten doses would result in more striking cumulative activity increases for hapten affinity improvements, they would also result in significantly lower tumor cumulative activities overall.
We investigated the effect of the in vitro affinity maturation of 2D12.5 in vivo in a xenograft mouse model, where we compare bispecific antibodies constructed with the high-affinity C8.2.5 scFv and the wild-type 2D12.5 scFv (Figure 6). The comparative tumor uptake data show that the affinity maturation resulted in a significant improvement in hapten retention in vivo in this model. These results encourage the design of more elaborate in vivo studies in mice including optimization of the injected dose of bsAb and blockage of the circulating bsAb with a clearing agent. Detailed studies with optimized dosing will be presented elsewhere. While we have utilized a bispecific antibody that is a C-terminal scFv fusion to the light chain of an IgG, other bispecific constructs could be constructed with the C8.2.5 scFv such as diabodies [51], scFv-IgGs, di-diabodies, and scDb-Fc fusions (reviewed in [52]).
While IgG-like bispecific antibodies are expected to result in significantly higher tumor accumulation than smaller antibody fragments due to slower blood clearance [53, 54], a considerable amount of antibody will likely remain in the blood at the time of hapten dosing. Thus a clearing/blocking step may be necessary to minimize hapten binding to residual bsAb in the blood. This may be achieved with a DOTA conjugated dextran or albumin blocking agent, a cold dose of DOTA chelate, a galactosylated DOTA-dextran or DOTA-albumin clearing agent, or a combination thereof. While three-step pretargeted radioimmunotherapy is more complex than proposed two-step approaches [12, 55], it may result in higher tumor doses for a given amount of bispecific antibody (due to a higher concentration of hapten binding sites accumulating in the tumor) and possible antibody dependent cell-mediated cytotoxicity (ADCC) and complement dependent cycotoxicity (CDC) due to the retained Fc domain.
Based on the model results, hapten retention is expected to be similar for DOTA-Bn-Y with a 10 pM affinity and DOTA-Y with a 100 pM affinity under antigen saturation for highly expressed tumor targets. This is a hypothesis that will be tested in vivo. Either DOTA or DOTABn could be used with this system; we have recently shown that 111In-DOTA-Bn clears rapidly from the blood, and is not retained significantly in any organ (K.D. Orcutt et al., manuscript in press).
We have engineered a versatile, DOTA-chelate-binding scFv with picomolar binding to yttrium, lutetium, and gadolinium chelates and nanomolar binding to indium and gallium chelates. Our approach comprised mathematical modeling of the pharmacokinetics of the bsAb and the metal chelate for the treatment of both micrometastatic disease and vascularized tumors to derive design specifications, and protein engineering via directed evolution using yeast surface display to achieve the desired outcome experimentally. We anticipate that the high-affinity DOTA-binding C8.2.5 scFv will prove useful for pretargeted imaging with positron emission tomography using 86Y and single photon emission computed tomography using 111In and pretargeted therapy with beta-emitters 177Lu and 90Y. C8.2.5 may also be useful for targeted MRI with multivalent macromolecular contrast agents containing DOTA-Gd.
Supplementary Material
ACKNOWLEDGMENTS
We thank Ioannis Papayannopoulos for LC-MS analysis, the MIT Biopolymers Laboratory for HPLC purification and MALDI-TOF analysis, and the MIT Flow Cytometry Core Facility for technical assistance. We thank Greg Thurber, Steven Sazinsky, Mike Schmidt and Margaret Ackerman for helpful discussions and Stefan Zajic for designing the wild type scFv with 2D12.5 variable domains. This work was supported by the National Institutes of Health CA101830 and a National Science Foundation Graduate Research Fellowship to K.D.O.
Footnotes
CONFLICT OF INTEREST The authors declare that they have no conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- [1].Axworthy DB, Reno JM, Hylarides MD, Mallett RW, Theodore LJ, Gustavson LM, et al. Cure of human carcinoma xenografts by a single dose of pretargeted yttrium-90 with negligible toxicity. Proc Natl Acad Sci U S A. 2000;97:1802–7. doi: 10.1073/pnas.97.4.1802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Goldenberg DM, Chatal JF, Barbet J, Boerman O, Sharkey RM. Cancer Imaging and Therapy with Bispecific Antibody Pretargeting. Update Cancer Ther. 2007;2:19–31. doi: 10.1016/j.uct.2007.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Goodwin DA, Meares CF, Watanabe N, Mctigue M, Chaovapong W, Ransone CM, et al. Pharmacokinetics of Pretargeted Monoclonal-Antibody 2d12.5 and Y-88 Janus-2-(PNitrobenzyl)-1,4,7,10-Tetraazacyclododecanetetraacetic Acid (Dota) in Balb/C Mice with Khjj Mouse Adenocarcinoma - a Model for Y-90 Radioimmunotherapy. Cancer Research. 1994;54:5937–46. [PubMed] [Google Scholar]
- [4].Liu G, Dou S, Mardirossian G, He J, Zhang S, Liu X, et al. Successful radiotherapy of tumor in pretargeted mice by 188Re-radiolabeled phosphorodiamidate morpholino oligomer, a synthetic DNA analogue. Clin Cancer Res. 2006;12:4958–64. doi: 10.1158/1078-0432.CCR-06-0844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Pantelias A, Pagel JM, Hedin N, Saganic L, Wilbur S, Hamlin DK, et al. Comparative biodistributions of pretargeted radioimmunoconjugates targeting CD20, CD22, and DR molecules on human B-cell lymphomas. Blood. 2007;109:4980–7. doi: 10.1182/blood-2006-11-056895. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Forster GJ, Santos EB, Smith-Jones PM, Zanzonico P, Larson SM. Pretargeted radioimmunotherapy with a single-chain antibody/streptavidin construct and radiolabeled DOTA-biotin: strategies for reduction of the renal dose. J Nucl Med. 2006;47:140–9. [PubMed] [Google Scholar]
- [7].Hamblett KJ, Kegley BB, Hamlin DK, Chyan MK, Hyre DE, Press OW, et al. A streptavidin-biotin binding system that minimizes blocking by endogenous biotin. Bioconjug Chem. 2002;13:588–98. doi: 10.1021/bc010087t. [DOI] [PubMed] [Google Scholar]
- [8].Foulon CF, Alston KL, Zalutsky MR. Astatine-211-labeled biotin conjugates resistant to biotinidase for use in pretargeted radioimmunotherapy. Nucl Med Biol. 1998;25:81–8. doi: 10.1016/s0969-8051(97)00166-2. [DOI] [PubMed] [Google Scholar]
- [9].Stickney DR, Anderson LD, Slater JB, Ahlem CN, Kirk GA, Schweighardt SA, et al. Bifunctional antibody: a binary radiopharmaceutical delivery system for imaging colorectal carcinoma. Cancer Res. 1991;51:6650–5. [PubMed] [Google Scholar]
- [10].Le Doussal JM, Martin M, Gautherot E, Delaage M, Barbet J. In vitro and in vivo targeting of radiolabeled monovalent and divalent haptens with dual specificity monoclonal antibody conjugates: enhanced divalent hapten affinity for cell-bound antibody conjugate. J Nucl Med. 1989;30:1358–66. [PubMed] [Google Scholar]
- [11].Janevik-Ivanovska E, Gautherot E, Hillairet de Boisferon M, Cohen M, Milhaud G, Tartar A, et al. Bivalent hapten-bearing peptides designed for iodine-131 pretargeted radioimmunotherapy. Bioconjug Chem. 1997;8:526–33. doi: 10.1021/bc970083h. [DOI] [PubMed] [Google Scholar]
- [12].Sharkey RM, McBride WJ, Karacay H, Chang K, Griffiths GL, Hansen HJ, et al. A universal pretargeting system for cancer detection and therapy using bispecific antibody. Cancer Res. 2003;63:354–63. [PubMed] [Google Scholar]
- [13].Kraeber-Bodere F, Rousseau C, Bodet-Milin C, Ferrer L, Faivre-Chauvet A, Campion L, et al. Targeting, toxicity, and efficacy of 2-step, pretargeted radioimmunotherapy using a chimeric bispecific antibody and 131I-labeled bivalent hapten in a phase I optimization clinical trial. J Nucl Med. 2006;47:247–55. [PubMed] [Google Scholar]
- [14].Corneillie TM, Whetstone PA, Lee KC, Wong JP, Meares CF. Converting weak binders into infinite binders. Bioconjug Chem. 2004;15:1389–91. doi: 10.1021/bc049825e. [DOI] [PubMed] [Google Scholar]
- [15].Penfield JG, Reilly RF. Nephrogenic systemic fibrosis risk: is there a difference between gadolinium-based contrast agents? Semin Dial. 2008;21:129–34. doi: 10.1111/j.1525-139X.2007.00408.x. [DOI] [PubMed] [Google Scholar]
- [16].Camera L, Kinuya S, Garmestani K, Wu C, Brechbiel MW, Pai LH, et al. Evaluation of the serum stability and in vivo biodistribution of CHX-DTPA and other ligands for yttrium labeling of monoclonal antibodies. J Nucl Med. 1994;35:882–9. [PubMed] [Google Scholar]
- [17].Byegard J, Skarnemark G, Skalberg M. The stability of some metal EDTA, DTPA and DOTA complexes: Application as tracers in groundwater studies. Journal of Radioanalytical and Nuclear Chemistry. 1999;241:281–90. [Google Scholar]
- [18].Le Mignon MM, Chambon C, Warrington S, Davies R, Bonnemain B. Gd-DOTA. Pharmacokinetics and tolerability after intravenous injection into healthy volunteers. Invest Radiol. 1990;25:933–7. [PubMed] [Google Scholar]
- [19].Lubic SP, Goodwin DA, Meares CF, Song C, Osen M, Hays M. Biodistribution and dosimetry of pretargeted monoclonal antibody 2D12.5 and Y-Janus-DOTA in BALB/c mice with KHJJ mouse adenocarcinoma. J Nucl Med. 2001;42:670–8. [PubMed] [Google Scholar]
- [20].Corneillie TM, Whetstone PA, Fisher AJ, Meares CF. A rare earth-DOTA-binding antibody: probe properties and binding affinity across the lanthanide series. J Am Chem Soc. 2003;125:3436–7. doi: 10.1021/ja029363k. [DOI] [PubMed] [Google Scholar]
- [21].Love RA, Villafranca JE, Aust RM, Nakamura KK, Jue RA, Major JG, Jr., et al. How the anti-(metal chelate) antibody CHA255 is specific for the metal ion of its antigen: X-ray structures for two Fab’/hapten complexes with different metals in the chelate. Biochemistry. 1993;32:10950–9. doi: 10.1021/bi00092a004. [DOI] [PubMed] [Google Scholar]
- [22].Blake DA, Jones RM, Blake RC, 2nd, Pavlov AR, Darwish IA, Yu H. Antibody-based sensors for heavy metal ions. Biosens Bioelectron. 2001;16:799–809. doi: 10.1016/s0956-5663(01)00223-8. [DOI] [PubMed] [Google Scholar]
- [23].Thurber GM, Zajic SC, Wittrup KD. Theoretic criteria for antibody penetration into solid tumors and micrometastases. J Nucl Med. 2007;48:995–9. doi: 10.2967/jnumed.106.037069. [DOI] [PubMed] [Google Scholar]
- [24].Cox JPL, Craig AS, Helps IM, Jankowski KJ, Parker D, Eaton MAW, et al. SYNTHESIS OF C-FUNCTIONALIZED AND N-FUNCTIONALIZED DERIVATIVES OF 1,4,7-TRIAZACYCLONONANE-1,4,7-TRIYLTRIACETIC ACID (NOTA), 1,4,7,10-TETRAAZACYCLODODECANE-1,4,7,10-TETRAYLTETRA-ACETIC ACID (DOTA), AND DIETHYLENENETRIAMINEPENTA-ACETIC ACID (DTPA) - BIFUNCTIONAL COMPLEXING AGENTS FOR THE DERIVATIZATION OF ANTIBODIES. Journal of the Chemical Society-Perkin Transactions. 1990;1:2567–76. [Google Scholar]
- [25].Takenouchi K, Watanabe K, Kato Y, Koike T, Kimura E. Novel Bifunctional Macrocyclic Chelating-Agents Appended With A Pendant-Type Carboxymethylamino Ligand And Nitrobenzyl Group And Sability Of The Y-88(III) Complexes. Journal of Organic Chemistry. 1993;58:1955–8. [Google Scholar]
- [26].Ashweek NJ, Coldham I, Haxell TFN, Howard S. Preparation of diamines by lithiation-substitution of imidazolidines and pyrimidines. Organic & Biomolecular Chemistry. 2003;1:1532–44. doi: 10.1039/b301502e. [DOI] [PubMed] [Google Scholar]
- [27].Chao G, Lau WL, Hackel BJ, Sazinsky SL, Lippow SM, Wittrup KD. Isolating and engineering human antibodies using yeast surface display. Nat Protoc. 2006;1:755–68. doi: 10.1038/nprot.2006.94. [DOI] [PubMed] [Google Scholar]
- [28].Hackel BJ, Kapila A, Wittrup KD. Picomolar affinity fibronectin domains engineered utilizing loop length diversity, recursive mutagenesis, and loop shuffling. J Mol Biol. 2008;381:1238–52. doi: 10.1016/j.jmb.2008.06.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Graff CP, Chester K, Begent R, Wittrup KD. Directed evolution of an anti-carcinoembryonic antigen scFv with a 4-day monovalent dissociation half-time at 37 degrees C. Protein Eng Des Sel. 2004;17:293–304. doi: 10.1093/protein/gzh038. [DOI] [PubMed] [Google Scholar]
- [30].Orcutt KD, Ackerman ME, Cieslewicz M, Quiroz E, Slusarczyk AL, Frangioni JV, et al. A modular IgG-scFv bispecific antibody topology. Protein Eng Des Sel. doi: 10.1093/protein/gzp077. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Humblet V, Misra P, Frangioni JV. An HPLC/mass spectrometry platform for the development of multimodality contrast agents and targeted therapeutics: prostate-specific membrane antigen small molecule derivatives. Contrast Media Mol Imaging. 2006;1:196–211. doi: 10.1002/cmmi.106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Misra P, Humblet V, Pannier N, Maison W, Frangioni JV. Production of multimeric prostate-specific membrane antigen small-molecule radiotracers using a solid-phase 99mTc preloading strategy. J Nucl Med. 2007;48:1379–89. doi: 10.2967/jnumed.107.040303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Schmidt MM, Thurber GM, Wittrup KD. Kinetics of anti-carcinoembryonic antigen antibody internalization: effects of affinity, bivalency, and stability. Cancer Immunol Immunother. 2008 doi: 10.1007/s00262-008-0518-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [34].Fenwick JR, Philpott GW, Connett JM. Biodistribution and histological localization of anti-human colon cancer monoclonal antibody (MAb) 1A3: the influence of administered MAb dose on tumor uptake. Int J Cancer. 1989;44:1017–27. doi: 10.1002/ijc.2910440614. [DOI] [PubMed] [Google Scholar]
- [35].Knox SJ, Goris ML, Tempero M, Weiden PL, Gentner L, Breitz H, et al. Phase II trial of yttrium-90-DOTA-biotin pretargeted by NR-LU-10 antibody/streptavidin in patients with metastatic colon cancer. Clin Cancer Res. 2000;6:406–14. [PubMed] [Google Scholar]
- [36].Shih LB, Thorpe SR, Griffiths GL, Diril H, Ong GL, Hansen HJ, et al. The processing and fate of antibodies and their radiolabels bound to the surface of tumor cells in vitro: a comparison of nine radiolabels. J Nucl Med. 1994;35:899–908. [PubMed] [Google Scholar]
- [37].Schmidt MM, Wittrup KD. A modeling analysis of the effects of molecular size and binding affinity on tumor targeting. Mol Cancer Ther. 2009;8:2861–71. doi: 10.1158/1535-7163.MCT-09-0195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Corneillie TM, Fisher AJ, Meares CF. Crystal structures of two complexes of the rare-earth-DOTA-binding antibody 2D12.5: ligand generality from a chiral system. J Am Chem Soc. 2003;125:15039–48. doi: 10.1021/ja037236y. [DOI] [PubMed] [Google Scholar]
- [39].Reiter Y, Brinkmann U, Lee B, Pastan I. Engineering antibody Fv fragments for cancer detection and therapy: disulfide-stabilized Fv fragments. Nat Biotechnol. 1996;14:1239–45. doi: 10.1038/nbt1096-1239. [DOI] [PubMed] [Google Scholar]
- [40].Moreau J, Guillon E, Pierrard JC, Rimbault J, Port M, Aplincourt M. Complexing mechanism of the lanthanide cations Eu3+, Gd3+, and Tb3+ with 1,4,7,10-tetrakis(carboxymethyl)-1,4,7,10-tetraazacyclododecane (dota)-characterization of three successive complexing phases: study of the thermodynamic and structural properties of the complexes by potentiometry, luminescence spectroscopy, and EXAFS. Chemistry. 2004;10:5218–32. doi: 10.1002/chem.200400006. [DOI] [PubMed] [Google Scholar]
- [41].Liu G, Dou S, He J, Liu X, Rusckowski M, Hnatowich DJ. Predicting the biodistribution of radiolabeled cMORF effector in MORF-pretargeted mice. Eur J Nucl Med Mol Imaging. 2007;34:237–46. doi: 10.1007/s00259-006-0222-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Zhu H, Jain RK, Baxter LT. Tumor pretargeting for radioimmunodetection and radioimmunotherapy. J Nucl Med. 1998;39:65–76. [PubMed] [Google Scholar]
- [43].Siegel JA, Stabin MG. Absorbed fractions for electrons and beta particles in spheres of various sizes. J Nucl Med. 1994;35:152–6. [PubMed] [Google Scholar]
- [44].Butlin NG, Meares CF. Antibodies with infinite affinity: origins and applications. Acc Chem Res. 2006;39:780–7. doi: 10.1021/ar020275e. [DOI] [PubMed] [Google Scholar]
- [45].Aime S, Barge A, Botta M, Fasano M, Ayala JD, Bombieri G. Crystal structure and solution dynamics of the lutetium(III) chelate of DOTA. Inorganica Chimica Acta. 1996;246:423–9. [Google Scholar]
- [46].Cosentino U, Villa A, Pitea D, Moro G, Barone V, Maiocchi A. Conformational characterization of lanthanide(III)-DOTA complexes by ab initio investigation in vacuo and in aqueous solution. J Am Chem Soc. 2001;124:4901–9. doi: 10.1021/ja017666t. [DOI] [PubMed] [Google Scholar]
- [47].Aime S, Botta M, Fasano M, Marques MPM, Geraldes CFGC, Pubanz D, et al. Conformational and coordiation equiliibria on DOTA complexes of lanthanide metal ions in aqueous solution studied by 1H-NMR spectroscopy. Inorganic Chemistry. 1997;36:2059–68. doi: 10.1021/ic961364o. [DOI] [PubMed] [Google Scholar]
- [48].Liu S, He Z, Hsieh W, Fanwick PE. Synthesis, characterization, and x-ray crystal structure of In(DOTA-AA) (AA = p-aminoanilide): A model for 111In-labeled DOTA-biomolecule conjugates. Inorganic Chemistry. 2003;42:8831–7. doi: 10.1021/ic0349914. [DOI] [PubMed] [Google Scholar]
- [49].Heppeler A, Froidevaux S, Macke HR, Jermann E, Behe M, Powell P, et al. Radiometal-labelled macrocyclic chelator-derivatised somatostatin analogue with superb tumour-targeting properties and potential for receptor-mediated internal radiotherapy. Chem Eur J. 1999;5:1974–81. [Google Scholar]
- [50].Midelfort KS, Hernandez HH, Lippow SM, Tidor B, Drennan CL, Wittrup KD. Substantial energetic improvement with minimal structural perturbation in a high affinity mutant antibody. J Mol Biol. 2004;343:685–701. doi: 10.1016/j.jmb.2004.08.019. [DOI] [PubMed] [Google Scholar]
- [51].DeNardo DG, Xiong CY, Shi XB, DeNardo GL, DeNardo SJ. Anti-HLA-DR/anti-DOTA diabody construction in a modular gene design platform: bispecific antibodies for pretargeted radioimmunotherapy. Cancer Biother Radiopharm. 2001;16:525–35. doi: 10.1089/10849780152752128. [DOI] [PubMed] [Google Scholar]
- [52].Marvin JS, Zhu Z. Recombinant approaches to IgG-like bispecific antibodies. Acta Pharmacol Sin. 2005;26:649–58. doi: 10.1111/j.1745-7254.2005.00119.x. [DOI] [PubMed] [Google Scholar]
- [53].Schneider DW, Heitner T, Alicke B, Light DR, McLean K, Satozawa N, et al. In vivo biodistribution, PET imaging, and tumor accumulation of 86Y- and 111In-antimindin/RG-1, engineered antibody fragments in LNCaP tumor-bearing nude mice. J Nucl Med. 2009;50:435–43. doi: 10.2967/jnumed.108.055608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [54].Olafsen T, Kenanova VE, Wu AM. Tunable pharmacokinetics: modifying the in vivo half-life of antibodies by directed mutagenesis of the Fc fragment. Nat Protoc. 2006;1:2048–60. doi: 10.1038/nprot.2006.322. [DOI] [PubMed] [Google Scholar]
- [55].Barbet J, Peltier P, Bardet S, Vuillez JP, Bachelot I, Denet S, et al. Radioimmunodetection of medullary thyroid carcinoma using indium-111 bivalent hapten and anti-CEA × anti-DTPA-indium bispecific antibody. J Nucl Med. 1998;39:1172–8. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.