Improved Tumor Selectivity of Radiolabeled Peptides by Receptor and

and Antigen Dual Targeting in the Neurotensin Receptor Model ... and a tumor antigen show increased selectivity to target tumor cells as compared to c...
0 downloads 0 Views 105KB Size
654

Bioconjugate Chem. 2002, 13, 654−662

Improved Tumor Selectivity of Radiolabeled Peptides by Receptor and Antigen Dual Targeting in the Neurotensin Receptor Model Marc Hillairet de Boisferon,§,†,| Olivier Raguin,§,†,| Cynthia Thiercelin,†,| Monique Dussaillant,†,| William Roste`ne,† Jacques Barbet,‡,| Andre´ Pe´legrin,#,| and Anne Gruaz-Guyon*,†,| INSERM U339, Faculte´ de Me´decine et Hoˆpital Saint-Antoine, Paris, France; INSERM U463, Nantes, France; EA2989 CRLC Val d'Aurelle-Paul Lamarque, Montpellier, France; CNRS GDR 2352, Immunociblage des tumeurs, France. Received November 28, 2001; Revised Manuscript Received March 1, 2002

Radiolabeled peptides are emerging tools for diagnosis and therapy of tumors overexpressing receptors. However, binding to receptors expressed by nontumor tissues may cause toxicity. The objective of this study was to specifically enhance the binding affinity of labeled peptides to tumor cells, as opposed to receptor-positive nontumor cells, to ensure targeting selectivity. This was achieved by the simultaneous binding of hapten-bearing peptides to their receptor and to a tumor-associated antigen, mediated by a bispecific antibody directed to the tumor antigen and to the hapten. Binding of labeled neurotensin analogues bearing the DTPA(indium) hapten (NT-DTPA(111In)) to human colorectal carcinoma cells (HT29), which express the neurotensin receptor (NTR1) and carcinoembryonic antigen (CEA), was studied in the presence of a bispecific antibody (BsmAb) directed to CEA and to DTPA(indium). In vitro dual binding of NT-DTPA in the presence of BsmAb was about 6.5-fold higher than monovalent binding to NTR1 and 3.5-fold higher than the sum of the monovalent bindings to NTR1 or to CEA, suggesting cooperativity. Increased binding under bivalent conditions translated into increased internalization. In vivo pretargeting with BsmAb enhanced tumor uptake and tumor retention. Hapten bearing peptides binding simultaneously an overexpressed cell-surface receptor and a tumor antigen show increased selectivity to target tumor cells as compared to cells only expressing the cell surface receptor. Better resistance to enzymatic degradation and optimized administration protocols should further enhance in vivo targeting selectivity and may allow the development of radiopharmaceuticals labeled with isotopes suitable for radiotherapy such as 131I or 90 Y.

INTRODUCTION

Many tumors overexpress receptors for small biologically active peptides, such as hormones or neuromediators. Their radiolabeled analogues are potential tools for tumor targeting. Their fast diffusion in tumor tissue offers a major pharmacokinetic advantage as compared to antibodies. In addition, internalization of receptorbound ligands offers an opportunity for isotope accumulation in target cells. The potential of peptide radiopharmaceuticals is illustrated by the successful development of somatostatin analogues to visualize tumors and more recently by therapeutic applications (1, 2). Labeled analogues of peptide ligands for several other receptors, overexpressed by tumors (3), such as CCK (4, 5), gastrin (4), vasoactive intestinal peptide (6), melanocyte stimulating hormone (7), and neurotensin (8) are under investigation. However, the background due to receptor-positive nontumor tissues contributes to de* Correspondence to Anne Gruaz-Guyon INSERM U.339, Biophysique, Faculte´ de Me´decine Saint-Antoine, 27 Rue Chaligny, 75012 Paris, France. (Tel: (33) 1 40 01 14 66, Fax: (33) 1 43 43 89 46). e-mail: [email protected]. † INSERM U339, Faculte ´ de Me´decine et Hoˆpital SaintAntoine. ‡ INSERM U463, Nantes. # EA2989 CRLC Val d'Aurelle-Paul Lamarque. | CNRS GDR 2352, Immunociblage des tumeurs. § The contributions of Marc Hillairet de Boisferon and Olivier Raguin are to be considered as equal.

crease tumor to normal tissue uptake ratios and is expected to cause toxicity (9). Enhancement of the affinity of a peptide for tumor cell surface, as opposed to receptor-positive nontumor tissue, would increase targeting selectivity and might be a key step to the development of more efficient therapeutic agents. It is well-known that affinity is enhanced by the cooperativity of multivalent binding. The aim of the present study was to evaluate the feasibility of increasing the binding affinity of a labeled peptide for receptorpositive tumor cells by the simultaneous targeting of an endogenous receptor and an antigen at the tumor cell surface. With an adequate choice of the antigen and of the receptor, no affinity enhancement will occur on nontumor tissue expressing only one of these targets. Hence, double targeting will enhance selectivity for tumor cells. The binding of an anti-hapten x anti-antigen bispecific antibody (BsmAb) to a selected tumor-associated antigen provides hapten-binding sites at the tumor cell surface. Then a labeled hapten-bearing ligand can bind simultaneously to a tumor-overexpressed receptor and to the bispecific antibody (Figure 1). To evaluate this approach we have chosen carcinoembryonic antigen (CEA) and type 1 neurotensin receptor (NTR1) dual targeting as a model. CEA and NTR1 expression is encountered on some colorectal carcinomas, breast, small cell lung carcinoma, and pancreatic carcinoma cells. The HT29 colorectal carcinoma cell line was used to investigate this approach (10, 11). The diethyl-

10.1021/bc015585g CCC: $22.00 © 2002 American Chemical Society Published on Web 04/25/2002

Receptor and Antigen Dual Targeting

Figure 1. Schematic representation of enhanced tumor targeting specificity of a hapten-bearing peptide by dual targeting of a receptor and an antigen. At the tumor cell surface the dual binding of DTPA(indium)-NT analogue to NTR1 and to the BsmAb bound to CEA will provide affinity enhancement. This will not occur at the surface of receptor-positive antigen-negative nontumor cells. (DTPA: diethylenetriaminepentaacetic acid, NT: neurotensin, BsmAb: anti-CEA x anti-DTPA(indium) bispecific antibody).

enetriaminepentaacetic acid (DTPA)-indium complex (DTPA(indium)) was used as a hapten. Two neurotensin analogues bearing the DTPA(indium) hapten (DTPANT) were synthesized. Their dual binding to NTR1 and to CEA at HT29 cell surface, in the presence of anti-CEA x anti-DTPA(indium) BsmAb, was studied in vitro and in vivo. MATERIAL AND METHODS

Cells. HT29 cells (ATCC, Rockville, MD) are human colorectal carcinoma cells that express the high affinity neurotensin receptor (NTR1) (10) and CEA (11) at their surface. These cells were grown in DMEM medium (Gibco, France) supplemented with 10% fetal calf serum (FCS), 1% glutamine, and 0.1% gentamycin at 37 °C in 5% CO2. For binding experiments, cells were treated as described by Velcich et al. (11) in order to increase CEA expression. Briefly, cells were distributed into 12-well Corning dishes (6 × 105 cells per well in 1 mL) and incubated at 37 °C for 24 h using the culture conditions described above. Culture medium was then replaced by 1 mL of low serum medium (0.05% FCS). Twenty-four hours later, cells were treated with low serum medium supplemented with 10 µg/mL forskolin (Sigma, France). Synthesis of the DTPA-NT Analogues. DTPAdianhydride (Aldrich, France) was reacted with the lysine -NH2 of neurotensin (pGlu-Leu-Tyr-Glu-Asn-Lys-ProArg-Arg-Pro-Tyr-Ile-Leu) (Neosystem, France) or with the R-NH2 of a neurotensin analogue (Gly-Glu-Leu-TyrGlu-Asn-Lys(Ac)-Pro-Arg-Arg-Pro-Tyr-Ile-Leu) (Neosystem, France) to obtain, respectively, pGlu-Leu-Tyr-GluAsn-Lys(DTPA)-Pro-Arg-Arg-Pro-Tyr-Ile-Leu (hereafter referred to as [Lys(DTPA)]-NT) and DTPA-Gly-GluLeu-Tyr-Glu-Asn-Lys(Ac)-Pro-Arg-Arg-Pro-Tyr-Ile-Leu (DTPA-Gly-NT). Briefly, 1.8 µmol of a neurotensin solution (220 µL) in HEPES buffer (0.6 M, pH 8.2, filtered through Chelex 100 (Biorad, France)) was added with stirring to 14.5 µmol of DTPA dianhydride (Aldrich, France) in solution in DMSO (220 µL). The pH was maintained at 8.2 until the reaction was completed. After evaporation of the solvents, the crude product was dissolved in water and trifluoroacetic acid to reach pH 2. The solution was filtered through Chelex 100. DTPA-NT analogues were purified by C18 reverse phase chromatography (Nucleosil, Shandon, France) using a linear 15-min gradient (A:

Bioconjugate Chem., Vol. 13, No. 3, 2002 655

H2O/TFA(0.05%), B: acetonitrile) from 0% to 26% B and a linear 20-min gradient from 26% to 30%. Purity was checked by C18 reverse phase chromatography under isocratic conditions (acetonitrile 30%, H2O/TFA 0.05%). Both peptides were purified to at least 95% purity. The DTPA-NT analogues were characterized by mass spectrometry 2048.16 found, 2048.32 calculated for [Lys(DTPA)]-NT, 2165.50 found and 2165.37 calculated for DTPA-Gly-NT using a Nermag R10-10 mass spectrometer with an electrospray ion source (Analytica of Branford) equipped with an iron guide. Antibodies. All antibodies used in this work were kindly provided by Immunotech S.A., France. The antiCEA (clone F6) is a mouse IgG1,κ monoclonal antibody specific for human CEA (12). The antibody directed to the DTPA(indium) complex is an IgG1,λ (13). The anti-DTPA(indium) Fab′ fragment was biotinylated and immobilized on avidin-coated tubes (Immunotech S.A., France) to determine DTPA(indium)-NT immunoreactivity (1 µg per tube) or equilibrium affinity constants (0.1 µg per tube). The anti-CEA x anti-DTPA(indium) BsmAb was prepared as already described (14) by chemical coupling of the two reduced Fab′ fragments using o-phenylenedimaleimide. Equilibrium affinity constant (Ka) for the binding of anti-CEA x anti-DTPA(indium) BsmAb to CEA expressed at HT29 cell surface was determined as 2.6 ( 1.4 × 108 M-1 by competition binding experiments between 125I-labeled Bsmab (prepared by the iodogen method (15)) and increasing concentrations of unlabeled Bsmab (1 × 10-10 M to 1 × 10-8 M). Nonspecific binding was evaluated in the presence of excess unlabeled anti-CEA IgG (2 × 10-7 M). The number of binding sites per HT29 cell was 2.1 ( 0.1 × 104 after a 24-h treatment by forskolin, about 6-fold higher than that determined without forskolin treatment (3.5 ( 0.7 × 103). All subsequent experiments were performed on forskolin-treated HT29 cells. The anti-CEA immunoreactivity of the BsmAb was 87% (16) and the anti-DTPA(indium) immunoreactivity of the BsmAb determined by incubating trace amounts of 125I-labeled BsmAb in DTPA(indium)-coated tubes was 85%. Radiolabeling. The DTPA-NT analogues (0.2 nmol) were labeled with indium-111 (111InCl3, 11.1 MBq, CIS bio international, France) in 100 mM acetate-10 mM citrate buffer pH 5.0 during 22 h at room temperature. Then, since the anti-DTPA(indium) antibody is specific to the DTPA(indium) complex, free DTPA groups were saturated with nonradioactive InCl3 (2 nmol) in the same buffer. Excess free indium was removed on a Sep-Pak cartridge (Waters, Milford, MA) in acetate-citrate buffer, followed by elution of the labeled DTPA-NT analogues in acetate-citrate buffer/ethanol (50/50, v/v). 111In-labeling of both DTPA-NT analogues led to a chelation of indium-111 of 88 ( 5% for 111In-labeled [Lys(DTPA)]NT and 85 ( 4% for DTPA(111In)-Gly-NT. Their specific activities were very similar: 179 ( 16 and 177 ( 12 MBq/ nmol, respectively. Computer Simulations. Delaage and co-workers (17, 18) have introduced several years ago a methodology to calculate binding equilibrium under complex situations. This methodology has been implemented as an add-in to Microsoft Excel 97 run on an ordinary PC (19). The multiple equilibrium system is described by a partition function, and the program calculates the free concentrations of all interacting species by minimizing the partition function for a given set of parameters. Then parameter estimation is obtained by variance-weighted nonlinear least-squares regression. Results are given as the bestfit estimation ( SD. This program allowed us to deter-

656 Bioconjugate Chem., Vol. 13, No. 3, 2002

mine unknown parameters such as binding affinity, receptor density, or accessibility factors to interpret binding experiments. In addition it allowed us to test for the cooperativity of peptide-binding to target cells (19). This program is available on request to J. B. ([email protected]). Briefly, steps to model the multiple binding system are described as follows. F1 is the free NT-DTPA(In) concentration, F2 the free BsmAb concentration, F3 the NTR concentration, F4 the CEA concentration, P1 the affinity binding constant (Ka) of NT-DTPA to the antibody, P2 the Ka for binding of the antibody to CEA, P3 the Ka for binding of NT-DTPA to NTR, P4 an accessibility factor (taking into account a loss in affinity for a simultaneous binding of NT-DTPA to the receptor and the BsmAb), and P5 a cooperativity factor (reflecting the enhanced binding in dual binding condition). The concentrations of the various complexes formed are [NT-DTPA(In)+BsmAb]:F1.F2.P1; [BsMAb+CEA]: F2.F4.P2; [NTR+NT-DTPA(In)]: F1.F3.P3; [NTR+NT-DTPA(In)+ BsmAb]:F1.F3.P3. F2.P1.P4; [NTR+NT-DTPA(In)+ BsmAb+CEA]:F1.F3.P3. F2.P1.P4.F3.P3.P5. Binding affinity constants were set to the values determined in independent experiments, and the number of CEA and NTR binding sites and the accessibility and cooperativity factors were estimated, from experimental data, by the program. NT-DTPA Analogues Binding Parameters Determination. Equilibrium affinity binding constant (Ka) and immunoreactivity of ligands were evaluated from binding experiments on cells or on antibody-coated tubes. Binding parameters were then determined from experimental results by curve-fitting to experimental boundto-total ratios for all available data using the program described above. Every result is expressed as mean ( standard deviation. Determination of DTPA-NT Analogues Binding Affinities to the Anti-DTPA(Indium) Antibody. Affinity constants of each 111In-labeled DTPA-NT analogue for anti-DTPA(indium) antibody were determined by incubating trace amounts of the 111In-labeled [Lys(DTPA)]-NT and increasing concentrations of unlabeled DTPA-NT analogue saturated with nonradioactive InCl3 (5.1 × 10-11 M to 10-6 M) in anti-DTPA(indium)-coated tubes (0.1 µg) in a final volume of 1 mL of phosphatebuffered saline (PBS) supplemented with 0.2% bovine serum albumin (BSA) for 2.5 h at 37 °C under shaking. Then 100 µL of supernatant were collected and counted. Tubes were washed twice with 0.9% NaCl and 0.002% Tween 20 and counted. Nonspecific binding was evaluated in the presence of excess unlabeled DTPA-NT analogues (saturated with nonradioactive indium, 10-6 M) (two experiments in duplicate). Determination of DTPA-NT Analogues Binding Affinities to NTR1. Neurotensin (NT) was labeled with iodine-125 by the lactoperoxidase method (20) and purified by ion-exchange chromatography. Cells were incubated in the presence of trace amounts of 125I-neurotensin and increasing concentrations of unlabeled DTPA-NT analogues saturated with nonradioactive indium (4 × 10-10 M to 1 × 10-6 M) in 500 µL of binding medium (DMEM/0.2% BSA/0.8 mM 1,10-phenanthroline). For each experiment, nonspecific binding was evaluated in the presence of an excess of unlabeled neurotensin (1 × 10-6 M). After 1 h at 37 °C, 50 µL of the supernatant were taken off. Cells were washed twice with ice-cold DMEM/BSA 0.2% and removed from the wells by addi-

Hillairet de Boisferon et al.

tion of 900 µL 0.1 M NaOH and counted. An aliquot of the supernatant was counted (three experiments in triplicate). Binding of DTPA-NT Analogues to HT29 Cells in the Presence of Anti-CEA x Anti-DTPA(Indium) Bispecific Antibody. As HT29 cells simultaneously express CEA and NTR1, monovalent binding of DTPANT analogues to CEA, mediated by the BsmAb, was evaluated in the presence of excess unlabeled neurotensin in order to saturate NTR1 at cell surface. This condition mimics cells expressing CEA alone. Monovalent binding of DTPA-NT analogues to NTR1 in the presence of BsmAb was studied using excess anti-CEA IgG to saturate CEA at cell surface, mimicking cells expressing NTR1 alone. Dual binding was studied in parallel in absence of neurotensin or anti-CEA IgG. The following protocol (protocol 1) was used in most experiments unless otherwise stated. A solution of BsmAb (300 µL, binding medium) was preincubated on the cells. After 2 h of incubation at 37 °C, 111In-labeled-DTPANT analogues (200 µL binding medium) were added without washing. Then, the cells were incubated for 1 h at 37 °C. Final concentrations: DTPA(111In)-NT 0.15 × 10-9 M, BsmAb as indicated. The cells and the supernatant were counted as above. For NTR1 monovalent binding experiments, the BsmAb solution contained excess anti-CEA IgG (2 × 10-7 M, final concentration), and for CEA monovalent binding experiments, the 111Inlabeled-DTPA-NT analogue solution contained excess neurotensin (1 × 10-6 M, final concentration). In protocol 2, BsmAb was preincubated as above, and after 2 h at 37 °C, cells were washed once with ice-cold DMEM/0.2% BSA and incubated with 500 µL of 111Inlabeled-DTPA-NT analogues for 1 h at 37 °C. In protocol 3, the 111In-labeled-DTPA-NT analogues were incubated with BsmAb in 500 µL of binding medium for 30 min at 37 °C before a 1-h incubation with the cells. Binding kinetics and internalization experiments were performed with a 2-h pretargeting of the BsmAb without washing (protocol 1), with a final BsmAb concentration of 6 × 10-9 M and of 0.15 × 10-9 M 111In-labeled-DTPANT analogues. At selected time intervals (15, 30, 60, 120, and 240 min), supernatant and cells were removed and counted as above for binding determination. The internalized fraction was evaluated by treating the cells for 15 min at 4 °C with DMEM/0.2% BSA adjusted to pH 2.0 with 1 M HCl. This last procedure dissociated surfacebound ligands. Cells were then washed twice with icecold DMEM/0.2% BSA. The radioactivity of the cells was counted. Internalization was expressed as the percentage of the total incubated activity that remained associated with the cells after acid wash. As above, for each condition the monovalent binding to CEA or NTR1 was determined in the presence of excess neurotensin or antiCEA IgG, respectively. In Vivo Experiments. Animal experiments were performed in compliance with the regulations of the institution and with generally accepted guidelines governing such work. Animals. Female SWISS nu/nu mice, 6-8 weeks old (Iffa Credo, France), were grafted by subcutaneous injection in the flank with 2 × 106 HT29 human colorectal carcinoma cells. Biodistribution studies were performed two weeks later. Biodistribution. Mice were given anti-CEA x antiDTPA(indium) or an irrelevant anti-CD20 x anti-DTPA(indium) BsmAb (40 pmol in 150 µL of PBS) or vehicle only by intravenous injection. Twenty-four hours later, 111In-labeled [Lys(DTPA)]-NT (15-30 pmol, 25 MBq/

Receptor and Antigen Dual Targeting

nmol) or an irrelevant 111In-DTPA-bearing peptide AG5.1 (Ac-Lys(DTPA)-GABA-D-Tyr-GABA-Lys(Gly succinylhistamine)-NH2) (19) (15 pmol, 25 MBq/nmol) was injected iv. Mice were weighed and sacrificed at selected time intervals (1, 3, and 24 h). Tumor, organs, and blood were collected, and the radioactivity in these samples was determined. The injected doses were corrected for losses during injection and subcutaneously injected material remaining in the animal tail. Tumor Imaging. For imaging studies, mice were administered with relevant BsmAb (40 pmol) or vehicle 24 h before 111In-labeled [Lys(DTPA)]-NT (30 pmol) as described above. Images of the two groups were recorded simultaneously at 3, 6, and 24 h after injection of the labeled analogue. Mice were anesthetized and positioned on the head of the gamma-camera (Picker Prism2000XP). Acquisition was performed for 15 min with 163-183 keV and 237-257 keV energy windows. Statistical Analysis. Student t test was used to evaluate the significance of in vitro results. For in vivo experiments, statistical analysis of differences in the tissue uptake values and tumor-to-organ ratios was performed using ANOVA variance analysis followed by the Newman-Keuls test. Differences at p < 0.05 were considered significant.

Bioconjugate Chem., Vol. 13, No. 3, 2002 657

Figure 2. Binding of [Lys(DTPA)]-NT to CEA and NTR1 in the presence of increasing concentrations of pretargeted antiDTPA(indium) x anti-CEA BsmAb. A: monovalent binding to NTR1 in the presence of excess anti-CEA IgG (2 × 10-7 M) (4: experimental, curve: computer simulation), B: monovalent binding to CEA determined in the presence of excess NT (1 × 10-6 M) (]: experimental, curve: computer simulation), C: dual binding (0: experimental, curve: computer simulation). Final concentration of [Lys(DTPA)]-NT: 0.15 × 10-9 M. Fraction of radioactivity associated to the cells (B/T%) was counted. The mean of triplicate determinations are plotted ( standard deviation unless smaller than the point as plotted.

RESULTS

Synthesis of the DTPA-NT Analogues. Two neurotensin analogues, pGlu-Leu-Tyr-Glu-Asn-Lys(DTPA)Pro-Arg-Arg-Pro-Tyr-Ile-Leu ([Lys(DTPA)]-NT) and DTPA-Gly-Glu-Leu-Tyr-Glu-Asn-Lys(Ac)-Pro-Arg-ArgPro-Tyr-Ile-Leu (DTPA-Gly-NT), bearing the hapten DTPA, which is a chelating agent suitable for 111Inlabeling, have been synthesized by coupling DTPA to an NH2 group as described in Materials and Methods. Both peptides were purified to at least 95% purity and identified by mass spectrometry. Equilibrium Affinity Constants of DTPA-NT Analogues Binding to Anti-DTPA(Indium) Antibody. Equilibrium affinity constants for the binding of DTPANT analogues to anti-DTPA(indium) antibody were determined by competition experiments. The affinity constant of [Lys(DTPA)]-NT for binding to anti-DTPA(indium) antibody was 8.4 ( 0.7 × 108 M-1, which is in agreement with Ka values determined for peptides bearing DTPA coupled to the -NH2 of a lysine (19). The Ka binding value decreased when DTPA was coupled to the R-NH2 of the glycine residue (1.6 ( 0.1 × 108 M-1 for DTPA-Gly-NT). Affinity Constants of DTPA-NT Analogues Binding to NTR1. Affinity parameters for the binding of DTPA-NT analogues to NTR1 were determined by competition experiments. The affinity constants were similar for both DTPA-NT analogues (1.0 ( 0.4 × 108 M-1 for [Lys(DTPA)]-NT and 8.7 ( 3.4 × 107 M-1 for DTPA-Gly-NT) and about 10-fold lower than that of neurotensin (9.7 ( 3.0 × 108 M-1) determined in the same experiment. The number of binding sites was 4.9 ( 1.1 × 104 per cell. Monovalent Binding of DTPA-NT Analogues to NTR1 at HT29 Cell Surface in the Presence of AntiCEA x Anti-DTPA(Indium) Bispecific Antibody. To study the monovalent binding of DTPA-NT analogues to NTR1 in the presence of BsmAb, CEA at cell surface was saturated with 2 × 10-7 M anti-CEA IgG (no difference was observed between 2 × 10-7 M and 1 × 10-6 M). Increasing the BsmAb concentration (7.2 × 10-11 M to 5.4 × 10-8 M) decreased the binding of DTPA-NT

Figure 3. Binding of DTPA-Gly-NT to CEA and NTR1 in the presence of increasing concentrations of pretargeted antiDTPA(indium) x anti-CEA BsmAb. A: monovalent binding to NTR1 in the presence of excess anti-CEA IgG (2 × 10-7 M) (4: experimental, curve: computer simulation), B: monovalent binding to CEA determined in the presence of excess NT (1 × 10-6 M) (]: experimental, curve: computer simulation), C: dual binding (0: experimental, curve: computer simulation). Final concentration of DTPA-Gly-NT: 0.15 × 10-9 M. Fraction of radioactivity associated to the cells (B/T%) was counted. The mean of triplicate determinations are plotted ( standard deviation unless smaller than the point as plotted.

analogues to NTR1 by about 75% (from 1189 ( 234 to 256 ( 5 molecules per cell, P < 0.001) for [Lys(DTPA)]NT (Figure 2A) and about 62% (from 1231 ( 159 to 428 ( 41 molecules per cell, P ) 0.001) for DTPA-Gly-NT (Figure 3A) at the highest BsmAb concentration tested. This reduced binding was interpreted in terms of accessibility factors (0.13 ( 0.02 for [Lys(DTPA)]-NT and 0.11 ( 0.04 for DTPA-Gly-NT expressing steric hindrance in the simultaneous binding of the relatively small hapten-bearing peptide to the NT receptor and the antihapten antibody. Monovalent Binding of DTPA-NT Analogues to HT29 Cell Surface CEA. The monovalent binding of the labeled DTPA-NT analogues to CEA was monitored in the presence of increasing concentrations of BsmAb (7.2 × 10-11 M to 5.4 × 10-8 M). As HT29 cells simultaneously express NTR1 and CEA, the binding to CEA was evaluated in the presence of excess neurotensin (1 × 10-6 M) to saturate NTR1. Binding increased with increasing concentrations of BsmAb to reach a maximum of 535 (

658 Bioconjugate Chem., Vol. 13, No. 3, 2002

173 molecules bound per cell at a BsmAb concentration of 2 × 10-9 M for [Lys(DTPA)]-NT (Figure 2B) and of 493 ( 42 at a BsmAb concentration of 6 × 10-9 M for DTPA-Gly-NT (Figure 3B). Then, at higher BsmAb concentrations, labeled DTPA-NT analogues were trapped by excess antibody in the supernatant and binding to cells decreased. Bivalent Binding of DTPA-NT Analogues to CEA and NTR1. The dual binding of the 111In-labeled DTPANT analogues to NTR1 and CEA was studied in the presence of increasing bispecific antibody concentrations (7.2 × 10-11 M to 5.4 × 10-8 M). It was compared to monovalent binding to NTR1 and to CEA in experiments run in parallel. Dual binding exhibited a bell-shaped curve, as expected from the increasing concentration of free antibody in the supernatant (Figure 2C, 3C). Maximal bivalent binding of [Lys(DTPA)]-NT was observed at a bispecific antibody concentration of 2 × 10-9 M, and of DTPA-Gly-NT at a BsmAb concentration of 6 × 10-9 M. Bivalent binding was greater than each monovalent binding to NTR1 or to CEA (p < 0.001) except for DTPAGly-NT binding to NTR at a 7.2 × 10-11 M BsmAb concentration (Figures 2 and 3). In addition, it was higher than the sum of monovalent bindings in a large range of BsmAb concentration (2.2 × 10-10 M to 5.4 × 10-8 M, p < 0.001). The highest ratio between bivalent binding and the sum of the monovalent bindings was observed at a BsmAb concentration of 6 × 10-9 M. At this BsmAb concentration, 3123 ( 406 molecules of [Lys(DTPA)]NT were bound per cell versus 461 ( 44 for monovalent binding to NTR1 and 507 ( 108 for monovalent binding to CEA, and 4218 ( 135 molecules of DTPA-Gly-NT versus 640 ( 41 for monovalent binding to NTR1 and 493 ( 42 for monovalent binding to CEA. Because incubation time is not sufficient for receptor and antigen reexpression, the bound/internalized fractions should reflect the formation of bivalent complexes at the cell surface. Thus, binding on HT 29 cells was evaluated by computer simulation of the binding curves and nonlinear least-squares regression to estimate cooperativity factors. Affinity constants calculated in independent experiments were used to analyze experimental results. Typical curve fits are presented in Figures 2 and 3. Cooperativity factors were estimated to 1090 ( 15 for [Lys(DTPA)]-NT and 2440 ( 668 for DTPA-GlyNT. Cooperativity far exceeded steric hindrance represented in equilibrium calculations by the accessibility factors determined above. Without the assumption of cooperative binding, satisfactory curve fitting could not be obtained in the bivalent situation. Kinetics of Bivalent Binding and Internalization of DTPA-NT Analogues to CEA and NTR1. Kinetic experiments were performed with a 2-h pretargeting of the BsmAb (6 × 10-9 M). Binding of [Lys(DTPA)]-NT (Figure 4A) and DTPA-Gly-NT (Figure 4B) increased over the 4-h incubation under the monovalent NTR1 binding condition and bivalent binding condition. However, the ratios between bivalent binding and the sum of monovalent bindings were similar for both analogues and were not significantly increased between 2 and 4 h of incubation (3.2 ( 0.01 at 2 h and 3.7 ( 0.4 at 4 h for [Lys(DTPA)]-NT and 3.5 ( 0.01 and 3.8 ( 0.7, respectively, for DTPA-Gly-NT). A high fraction of the peptide bound to NTR1 was internalized (86 ( 4% and 79 ( 2% for [Lys(DTPA)]NT and DTPA-Gly-NT, respectively). A similar fraction of the bound peptide was internalized when bivalent binding was allowed (94 ( 6% and 89 ( 10%, respec-

Hillairet de Boisferon et al.

Figure 4. Binding kinetics and kinetics of internalization of DTPA-NT analogues at 37 °C in the presence of pretargeted BsmAb. Binding kinetics [Lys(DTPA)]-NT (A) and DTPA-GlyNT (B) and kinetics of internalization of [Lys(DTPA)]-NT (C) and DTPA-Gly-NT (D) (final concentrations: 0.15 × 10-9 M DTPA-NT analogues and 6 × 10-9 M BsmAb). Bivalent binding to CEA and NTR1 (solid line) was performed after a 2 h pretargeting of the BsmAb. Monovalent binding to CEA was evaluated in the presence of excess NT (1 × 10-6 M) (dotted line) and monovalent binding to NTR1 in the presence of excess anti-CEA IgG (2 × 10-7 M) (broken line). Fraction of radioactivity associated to the cells (B/T%) or inside the cells (I/T%) was counted. The mean of triplicate determinations are plotted ( standard deviation unless smaller than the point as plotted.

tively). The amount of internalized radioactivity under bivalent conditions was higher than that observed after binding to NTR1 (p < 0.001) for both analogues (7.7 ( 1.1-fold for [Lys(DTPA)]-NT and 6.3 ( 0.9 for DTPAGly-NT). Influence of the Protocol. Pretargeting of the BsmAb without a wash prior to the incubation of the peptide (protocol 1), which was performed in the above-described experiments, was compared to two other experimental conditions using a fixed BsmAb concentration (6 × 10-9 M). Protocol 2 was performed similarly except for a wash of the cells prior to the incubation of the peptide. In protocol 3, the 111In-DTPA-NT analogues were mixed with the BsmAb and incubated 30 min at 37 °C before incubation (1 h) with the cells. Monovalent binding to NTR1 or to CEA were performed in parallel in the presence of excess anti-CEA IgG or excess neurotensin as described above. Whatever the protocol used, when simultaneous binding to NTR1 and to CEA was allowed, the fraction of peptide bound to cells was increased compared to each monovalent binding conditions (Table 1). In addition, for the three protocols, the bivalent binding was higher than the sum of the monovalent bindings to NTR1 and to CEA suggesting that cooperativity is not dependent on the experimental conditions. Biodistribution Studies in Human Colorectal Cancer Xenografted Nude Mice. In vivo dual targeting of NTR1 and CEA was studied in nude mice grafted with HT29 cells and preinjected with anti-DTPA(indium) x anti-CEA BsmAb 24 h before 111In-labeled [Lys(DTPA)]NT injection (Table 2). Radioactivity persistence in blood was increased in mice pretargeted with BsmAb (1.6 ( 0.4 at 1 h, versus 0.32 ( 0.31% ID/g, p < 0.01). Pretargeting with BsmAb increased tumor uptake of 111In-labeled [Lys(DTPA)]-NT at 1, 3, and 24 h post peptide injection (p < 0.01). In addition, BsmAb pretargeting protracted tumor retention (with BsmAb 20% decrease of tumor activity between 1 h and 3 h postinjection, versus 57% in the absence of BsmAb). Dual

Receptor and Antigen Dual Targeting

Bioconjugate Chem., Vol. 13, No. 3, 2002 659

Table 1. Bivalent and Monovalent Bindings of DTPA-Gly-NT and [Lys(DTPA)-NT] to HT29 Cells: Comparison of Three Different Protocolsa

B/T %

bivalent protocol binding

DTPA-Gly-NT [Lys(DTPA)-NT]

1 2 3 1 2 3

monovalent monovalent binding binding to NTR1 to CEA

6.1 ( 0.2 1.3 ( 0.1 0.61 ( 0.04 13 ( 1 3.5 ( 0.2 1.9 ( 0.1 2.9 ( 0.2 1.1 ( 0.1 0.23 ( 0.02 5.3 ( 0.2 0.62 ( 0.03 1.7 ( 0.1 16 ( 1 2.6 ( 0.1 5.4 ( 1.1 2.3 ( 0.2 0.56 ( 0.04 0.84 ( 0.04

aProtocol 1: after 2 h preincubation of BsmAb, 111In-labeledDTPA-NT analogues were incubated for 1 h; protocol 2 was performed as protocol 1 except for a wash of the cells prior to peptide incubation; protocol 3: the 111In-labeled-DTPA-NT analogues were incubated with BsmAb for 30 min before a 1-h incubation with the cells. Final concentrations: BsmAb 6 × 10-9 M, DTPA(111In)-NT 0.15 × 10-9 M, 37 °C. (Fraction of radioactivity associated to the cells: B/T%, mean ( standard deviation).

targeting (111In-labeled [Lys(DTPA)]-NT and BsmAb) enhanced tumor to organ uptake ratios, particularly for the gastrointestinal tract at 3 and 24 h (p < 0.01) (stomach p < 0.01 at 3 and 24 h, colon and small intestine p < 0.05 at 3 h and p < 0.01 at 24 h) in comparison with NTR1 targeting (111In-labeled [Lys(DTPA)]-NT, without BsmAb). Tumor-to-kidney ratios were also improved (p < 0.01 at 3 h and p < 0.05 at 24 h). Pretargeting with an irrelevant BsmAb (anti-DTPA x anti-CD20), able to bind to DTPA but not to CEA was compared to pretargeting with the relevant BsmAb. Tumor-to-organ ratios at 3 h post labeled NT-DTPA injection was significantly improved by relevant BsmAb pretargeting as compared to irrelevant for organs known to express NT receptors (stomach: p < 0.01, colon: p < 0.01 and small intestine: p < 0.05), while no significant difference was observed for other organs. In mice pretargeted with the relevant BsmAb 3 h after injection of 111In-labeled [Lys(DTPA)]-NT tumor uptake was higher (p < 0.01) than tumor uptake (0.65 ( 0.09% ID/g) of a 111In-DTPA bearing peptide with no affinity Table 2. Biodistribution Studies of Cellsa,b

tumor blood kidneys T/blood T/heart T/lung T/liver T/spleen T/kidneys T/muscle T/GIT T/stomach T/colon T/small intestine

DISCUSSION

Within the past few years, the use of low molecular weight radiolabeled peptides, which exhibit a faster clearance than antibodies and provide a higher accessibility to tumor cells, has been described. Taking advantage of the overexpression of numerous peptide receptors by various tumor cells, labeled peptides (somatostatin, MSH, VIP, neurotensin, CCK, and gastrin analogues, for example) have been developed for diagnosis and therapeutic purposes (1-8), but binding also in Nude Mice Grafted with Human Colon Carcinoma

[Lys(DTPA)]-NT + relevant BsmAb, 40 pmol

3 h* (n ) 4) 0.65 ( 0.11 0.15 ( 0.03 1.6 ( 0.1

26.0 ( 4.2 11.8 ( 3.9 9.80 ( 3.76 5.14 ( 1.11 4.06 ( 2.06 0.10 ( 0.06 20.2 ( 3.9 3.07 ( 1.34 9.85 ( 6.45 3.23 ( 1.92 2.33 ( 0.74

Tumor (T)/Organ Ratio 1.92 ( 0.22 16.8 ( 4.5 39.2 ( 7.9 6.84 ( 0.95 27.4 ( 3.4 22.9 ( 4.7 4.02 ( 0.17 14.8 ( 2.1 20.8 ( 6.9 6.75 ( 0.44 8.51 ( 3.95 8.25 ( 1.44 8.28 ( 1.41 16.0 ( 2.2 6.79 ( 1.42 0.25 ( 0.06 0.19 ( 0.04 0.24 ( 0.07 9.99 ( 2.55 33.4 ( 11.8 35.1 ( 7.4 6.10 ( 1.98 7.54 ( 3.68 9.79 ( 1.10 nd 42.6 ( 5.6 32.6 ( 12.8 nd 6.91 ( 1.26 10.3 ( 2.8 nd 8.14 ( 2.04 7.93 ( 0.50

3 h* (n ) 7)

irrelevant (DTPA)-peptide + relevant BsmAb, 40 pmol

24 h** (n ) 4)

17.6 ( 6.6 19.6 ( 5.6 9.80 ( 3.46 9.01 ( 1.82 9.38 ( 1.44 0.073 ( 0.014 14.6 ( 8.2 2.79 ( 1.43 19.1 ( 4.6 3.37 ( 0.40 3.50( 0.36

24 h** (n ) 4) 1 h* (n ) 3) 3 h* (n ) 7) 24 h** (n ) 4)

[Lys(DTPA)]-NT + irrelevant BsmAb, 40 pmol

Uptake (% ID/g) 1.7 ( 0.5 0.72 ( 0.22 0.54 ( 0.13 2.9 ( 0.6 2.4 ( 0.3 1.3 ( 0.2 1.1 ( 0.4 0.74 ( 0.15 0.32 ( 0.31 0.043 ( 0.007 0.021 ( 0.003 1.6 ( 0.4 0.15 ( 0.03 0.034 ( 0.004 0.062 ( 0.011 0.044 ( 0.008 8.4 ( 3.7 9.8 ( 1.8 6.4 ( 2.9 11.7 ( 0.7 12.5 ( 2.0 5.8 ( 1.1 8.2 ( 0.6 7.2 ( 4.2 12.0 ( 11.8 16.2 ( 11.1 4.45 ( 4.20 13.2 ( 7.1 10.4 ( 9.0 0.24 ( 0.15 6.17 ( 9.1 5.44 ( 1.60 nd nd nd

3 h* (n ) 7)

for NTR1 (irrelevant-DTPA peptide: Ac-Lys(DTPA)GABA-D-Tyr-GABA-Lys(Gly succinyl-histamine)-NH2 (19)). At 3, 6, and 24 h post [Lys(DTPA)]-NT injection, tumors were visualized in relevant BsmAb-pretargeted mice whereas tumors were hardly or not detected in vehicle preinjected mice. Kidneys and bladder were the major radioactivity accumulation sites in normal organs, indicating that the major route of excretion was renal.

111In-Labeled-[Lys(DTPA)]-NT

[Lys(DTPA)]-NT 1 h* (n ) 3)

Figure 5. Tumor imaging in mice with pretargeted111In-labeled [Lys(DTPA)]-NT. Images (ventral views) were recorded at 3, 6, and 24 h postinjection (Picker Prism 2000XP gamma-camera). Subcutaneously HT29-xenografted nude mice were injected with A: anti-CEA x anti-DTPA(indium) BsmAb (40 pmol) (tumor weights: left 355.6 mg, right 497.9 mg) or B: vehicle (tumor weights: left 481.7 mg, right 358.8 mg), 24 h before 111In-labeled [Lys(DTPA)]-NT (30 pmol). t: tumor, b: bladder, k: kidney.

17.2 ( 4.4 21.5 ( 4.9 13.3 ( 4.6 8.72 ( 1.55 12.8 ( 2.6 0.13 ( 0.05 24.0 ( 8.3 3.52 ( 0.89 14.8 ( 8.3 2.50 ( 0.89 4.06 ( 0.51

17.1 ( 2.0 13.9 ( 1.5 10.4 ( 1.6 4.40 ( 0.66 6.46 ( 4.47 0.12 ( 0.04 20.0 ( 8.7 4.29 ( 0.88 15.3 ( 6.2 3.46 ( 1.00 3.84 ( 0.77

4.3 ( 0.6 9.7 ( 0.6 4.2 ( 0.8 4.8 ( 0.9 8.8 ( 0.8 0.41 ( 0.05 24.8 ( 2.1 5.53 ( 2.090 24.4 ( 9.9 2.94 ( 0.89 8.69 ( 4.70

a Nude mice, grafted with HT29 cells in the flank, were given anti-CEA x anti-DTPA(indium) or vehicle or irrelevant BsmAb by iv injection. Twenty-four hours later 15 (*) or 30 (**) pmol of 111In-labeled [Lys(DTPA)]-NT or 15 pmol of an irrelevant 111In-DTPA-bearing peptide (Ac-Lys(DTPA)-GABA-D-Tyr-GABA-Lys(Gly succinyl-histamine)-NH2) was injected iv. Mice were sacrificed at selected time intervals. Tumor, organs, and blood were collected, and the radioactivity in these samples was determined. Tissue radioactivity is expressed as percentage injected dose/g tissue (% ID/g mean in bold ( standard deviation). b Number of animals (n) indicated, except for stomach, colon, and small intestine, n ) 4.

660 Bioconjugate Chem., Vol. 13, No. 3, 2002

occurs in nontumor tissues expressing the receptors, which is expected to cause toxicity (9). We have shown that bispecific antibody-mediated dual binding enhances radiolabeled bivalent hapten binding in vitro (13, 21) and in vivo targeting in animal models (13) and in the clinic (14, 22, 23) and that it can also provide an improved selectivity of antigen targeting (19, 21). Our objective is to specifically enhance the binding affinity of peptides to receptor-positive tumor cells, as opposed to receptor-positive nontumor cells, to ensure a higher targeting selectivity. The dual binding of a radiolabeled peptide to a receptor and to an antigen on tumor cell surface should enhance the binding to these cells in comparison with a monovalent binding to receptorpositive antigen-negative cells. Numerous tumor cells simultaneously bear receptors and antigens at their surface. In the present paper, we chose neurotensin receptor (NTR1) and CEA dual targeting as a model to test the feasibility of this approach. NTR1 (24, 25) and CEA (26) are present on breast, small cell lung carcinoma, pancreatic carcinoma cells, and some colorectal carcinomas. HT29 cells were selected as a model of NTR1 and CEA coexpression. A BsmAb formed by the covalent linking of two Fab′, one directed to CEA and the other one to the DTPA(indium) hapten, was used. Binding of the BsmAb to CEA provided DTPA(indium) binding sites to the HT29 cell surface, allowing the simultaneous binding of DTPA(indium) bearing neurotensin analogues to two sites (NTR1 and CEA) at the target cell surface. All in vitro experiments with HT29 cells were performed after a 24-h forskolin treatment as this increases CEA levels at the cell surface (11) with no modification of the binding affinity to the CEA or to the NTR1 (data not shown). Under these conditions, expression of CEA and NTR1 was of the same order of magnitude, which is the optimal condition to evaluate the feasibility of dual binding. Two neurotensin analogues bearing the hapten DTPA were synthesized: [Lys(DTPA)]-NT and DTPA-GlyNT. The affinity of DTPA-Gly-NT for the anti-DTPA(indium) antibody was about 5-fold lower than that of [Lys(DTPA)]-NT in agreement with our previous observation that the highest Ka values are obtained when the DTPA is coupled to the -NH2 of a lysine as it is in the immunogen (19). Affinity constant for the binding of [Lys(DTPA)]-NT to the neurotensin receptor NTR1 was 10-fold lower than that of neurotensin itself, determined in the same experimental conditions, in agreement with the observation of a decreased affinity of neurotensin derivatives substituted on the -amino group of the lysine (27). Unexpectedly the affinity constant of DTPA-Gly-NT was similar to that of [Lys(DTPA)]-NT. As both DTPA-NT analogues were able to bind to NTR1 and to anti-DTPA(indium) antibody, we evaluated both the ability of the anti-CEA x anti-DTPA(indium) BsmAb to target DTPA-NT analogues to CEA, and the influence of BsmAb on the binding of these peptides to NTR1. The 111In-labeled DTPA-NT analogues were efficiently targeted to CEA by the BsmAb. Binding in the presence of increasing concentrations of antibody followed a bell-shaped curve, due to trapping of 111In-labeled DTPA-NT by excess antibody in the supernatant at high BsmAb concentrations. The BsmAb concentration necessary for maximal binding of the peptide was higher for DTPA-Gly-NT in agreement with its lower affinity for the anti-DTPA(indium) as compared to [Lys(DTPA)]-NT. The presence of BsmAb dramatically decreased but did not abrogate the monovalent binding of DTPA-NT

Hillairet de Boisferon et al.

analogues to NTR1 (more than 60% inhibition at the highest antibody concentration tested). This result suggested that the affinity of the BsmAb-DTPA(111In)-NT complex for NTR1 was lower than that of the free DTPA(111In)-NT, probably due to steric hindrance. The simultaneous binding of DTPA(111In)-NT analogues to NTR1 and to CEA (bivalent binding) in the presence of BsmAb was tested using different incubation protocols with controls of monovalent binding to CEA or NTR1 performed in parallel. Whatever the operating procedure, the incubation time, and BsmAb concentration dual, binding was higher than binding to NTR1 (with a 6 × 10-9 M BsmAb concentration 6.6-fold for DTPAGly-NT and about 6.8-fold for [Lys(DTPA)]-NT, after 2 h incubation of the peptides). These results demonstrated the feasibility of enhanced selectivity of radioisotopes targeting to tumor cells bearing simultaneously one receptor and one antigen, compared to nontarget cells bearing only the targeted receptor. The bivalent binding was higher than the monovalent binding to CEA (8.6-fold for DTPA-Gly-NT and 6.2-fold for [Lys(DTPA)]-NT in the same conditions). The increased selectivity toward receptor-positive/antigen-positive target cells, compared to receptor-negative/antigenpositive cells, may allow nontumor specific antigens, such as clusters of differentiation, to be considered in this context. In addition, the binding enhancement provided by dual binding was higher than the sum of the monovalent bindings for both DTPA-NT analogues in a wide range of bispecific antibody concentrations (2.2 × 10-10 M to 1.8 × 10-8 M). Fitting these binding curves required the assumption of cooperativity for the simultaneous binding of the DTPA-NT analogues to NTR1 and to anti-CEA x anti-DTPA(indium) BsmAb bound to cell surface CEA. These results suggest that, as previously observed for double antigen targeting (19, 21, 28), dual targeting to a receptor and an antigen provides an affinity enhancement due to the cooperativity of dual binding. Cooperativity factors far exceeded steric hindrance, represented in calculations by an accessibility factor. Thus, the decreased binding of DTPA-NT analogues to NTR1 observed in the presence of BsmAb should translate into reduced uptake in receptor-positive antigen-negative tissues and cooperativity should restore or even provide an enhanced binding on receptor positive antigen-positive tumor cells. In in vivo pretargeting approaches, unlabeled bispecific antibodies are injected first, and the labeled derivative is injected a few days later to allow tumor uptake and clearance of the bispecific antibody. However, at this time a significant fraction of the circulating antibody remains in the circulation. This was better reproduced in vitro by incubating the peptide without washing the BsmAb. As expected, the specific binding of the 111In-labeled DTPA-NT analogue was lower under these conditions, due to trapping of the peptide by excess antibody in the supernatant, but the ratios of dual binding to the sum of the monovalent bindings were higher. Binding of neurotensin analogues to NTR1 results in the internalization of the ligand. Bivalent binding did not prevent analogues from internalization. In fact, increased binding under bivalent conditions translated into increased internalization. This property is of major interest, as internalization is an important factor for internal radiotherapy since it may result in the accumulation of the radioisotope in the tumor and allow the use of shortrange emitting isotopes such as R emitters. In this

Receptor and Antigen Dual Targeting

respect, dual binding to a receptor and an antigen could afford a major advantage compared to single receptor targeting. The method was evaluated in nude mice grafted with HT29 human colorectal cells in the flank. [Lys(DTPA)]NT was labeled with indium-111 in order to perform biodistributions and imaging studies. The clearance of the hapten-bearing peptide was reduced in the presence of the BsmAb due to trapping in the plasma by excess antibody, as we demonstrated earlier for pretargeted bivalent DTPA haptens (13). Dual targeting (in the presence of BsmAb) enhanced tumor uptake in comparison with NTR1 targeting (without BsmAb). Tumor uptake was also increased by dual targeting in comparison with CEA targeting (irrelevant 111In labeled-DTPApeptide and BsmAb). In addition tumor to normal organ uptake ratios were enhanced particularly for the gastrointestinal tract (stomach, colon, and small intestine) where NT receptors are located (29), demonstrating that uptake increase provided by the BsmAb is higher in the tumor than in NTR1-positive tissues. Thus the targeting selectivity for tumor cells is enhanced by simultaneous targeting of NT receptor and CEA compared to monovalent targeting (NTR or CEA). Pretargeting with a relevant BsmAb (anti-CEA x anti-DTPA(indium)) provided a significant enhancement of tumor to NTR-positive organ ratios as compared to irrelevant BsmAb pretargeting (pretargeting with an anti-CD20 x anti-DTPA(indium) unable to bind to CEA). Thus tumor to NTR1positive organ ratios may be attributed to dual binding rather than to reduced clearance due to binding of the NT-DTPA(indium) analogue to a BsmAb. In vitro, the analysis of NT-DTPA binding to target cells in the presence and absence of BsmAb and inhibitors allowed us to conclude that NT-DTPA binds simultaneously cooperatively to NTR1 and CEA via the BsmAb. In vivo, pretargeting with BsmAb of NT-DTPA increased tumor uptake, protracted tumor retention, and enhanced tumor to organ uptake ratios, particularly for the gastrointestinal tract, but cooperativity could not be demonstrated. One reason is the reduced clearance of haptenbearing-NT in the presence of BsmAb, which improves tumor uptake even if the BsmAb does not bind the target cells. Other reasons are the enzymatic degradation of the peptide (in in vitro experiments 1,10-phenanthroline was added to prevent enzymatic degradation) and the possible use of nonoptimal doses of bispecific antibody and peptide. 111In-labeling of NT-DTPA analogues were used in this study to evaluate the feasibility of simultaneous targeting of a receptor and an antigen. To optimize the in vivo targeting selectivity to tumor cells and to develop radiopharmaceuticals suitable for therapy, new hapten bearing receptor-ligands resistant to enzymatic degradation must be developed. Tumor-to-kidneys ratios were also improved, but peptides exhibiting a low renal accretion, such as those described by Chavatte K. et al. (8), labeled with isotopes suitable for internal radiotherapy, such as 131I or 90Y, will be necessary. In conclusion, the double targeting of a receptor and an antigen at the tumor cell surface may be beneficial to peptide targeting of radioisotopes by increasing targeting specificity, tumor uptake, and isotope residence time in the tumor. DTPA-NT analogues, resistant to enzymatic degradation, in combination with an anti-CEA x antiDTPA(indium) bispecific antibody, might be used in several clinical situations, including pancreatic tumors, small cell lung carcinoma, or colon carcinoma expressing both CEA and a neurotensin receptor (24-26). Successful targeting of radioactive isotopes to MTC has been per-

Bioconjugate Chem., Vol. 13, No. 3, 2002 661

formed using either anti-CEA antibodies (22, 30) or gastrin or CCK analogues (4, 5). Tumor uptake and contrast ratios in this disease could be improved by means of dual targeting. This approach might be of interest for radiopeptide therapy in many other cases of tumors overexpressing peptide receptors. For example VIP/PACAP receptors are expressed by most of human tumors such as breast, colon, or small cell lung carcinomas (9). These receptors are potentiel targets for VIP/ PACAP ligands in vivo. The limitations for a radiotherapy using such agents include the destruction by irradiation of receptor-positive normal target tissues (9). Enhancing the selectivity by dual targeting of the VIP/ PACAP receptor and a selected antigen may provide tumor to normal tissue uptake ratios suited for therapy. Antigens such as MDR P-glycoproteins, expressed on multidrug resistant tumors, CEA, or MUC1 expressed by various carcinomas, might be used for dual targeting of VIP/PACAP receptor positive tumors. Numerous other receptor and antigens might be selected for dual targeting of tumors, but the antigen should not be internalizing for successful BsmAb pretargeting. With internalizing antigens, simultaneous injection of the labeled analogue and the BsmAb may be considered (21). Very often antigen and receptor expression differs among tumors of the same type. Nevertheless, it is expected that a set of radiopharmaceuticals recognizing different receptors and bearing the same hapten, combined with a set of BsmAb directed to various antigens and to the hapten, will allow combinations of one radiopharmaceutical and one BsmAb able to perform dual targeting on a wide range of tumors. ACKNOWLEDGMENT

The authors gratefully acknowledge Pr. G. Milhaud and Dr. D. Pelaprat for helpful discussions, Pr. S. Askienazy for research facilities, and M. Brissac, B. Vinson, and S. Hermant for in vivo experiments. This work was supported in part by the Association pour la Recherche contre le Cancer through Grant 4042 allocated to A.G.G. The authors thank the Ligue Nationale Contre le Cancer for awarding a fellowship to O.R. LITERATURE CITED (1) Breeman, W. A. P., de Jong, M., Kwekkeboom, D. J., Valkema, R., Bakker, W. H., Kooij, P. P. M., Visser, T. J., and Krenning, E. P. (2001) Somatostatin Receptor-Mediated Imaging and Therapy: Basic Science, Current Knowledge, Limitations and Future Perspectives. Eur. J. Nucl. Med. 28, 1421-1429. (2) Kwekkeboom, D. J., Krenning, E. P., and de Jong, M. (2000) Peptide Receptor Imaging and Therapy. J. Nucl. Med. 41 (10), 1704-1713. (3) Boerman, O. C., Oyen, W. J. G., and Corstens, F. H. M. (2000) Radio-Labeled Receptor-Binding Peptides: A New Class of Radiopharmaceuticals. Semin. Nucl. Med. 30 (3), 195-208. (4) Behr, T. M., Jenner, N., Behe, M., Angerstein, C., Gratz, S., Raue, F., and Becker, W. (1999) Radiolabeled peptides for targeting cholecystokinin-B/Gastrin receptor-expressing tumors. J. Nucl. Med. 40, 1029-1044. (5) De Jong, M., Bakker, W. H., Bernard, B. F., Valkema, R., Kwekkeboom, D. J., Reubi, J. C., Srinivasan, A., Schmidt, M., and Krenning, E. P. (1999) Preclinical and initial clinical evaluation of 111In-labeled nonsulfated CCK8 analog: a peptide for CCK-B receptor-targeted scintigraphy and radionucleide therapy. J. Nucl. Med. 40, 2081-2087. (6) Virgolini, I., Raderer, M., Kurtaran, A., Angelberger, P., Banyal, S., Yang, Q., Shuren, L., Banyai, M., Pidlich, J.,

662 Bioconjugate Chem., Vol. 13, No. 3, 2002 Niederle, B., Scheithauer, W., and Valent, P. (1994) Vasoactive intestinal peptide-receptor imaging for the localization of intestinal adenocarcinomas and endocrine tumors. N. Engl. J. Med. 331, 1116-1121. (7) Bard, D. R. (1995) An improved imaging agent for malignant melanoma, based on [Nle4, D-Phe7]R-melanocyte stimulating hormone. Nucl. Med. Commun. 16, 860-866. (8) Chavatte, K., Terriere, D., Jeannin, L., Iterbeke, K., Briejer, M., Schuurkes, J., Mertens, J. J. R., Bruyneel, E., Tourwe´, D., Leysen, J. E., and Bossuyt, A. (1999) Labelling and evaluation of new stabilised neurotensin (8-13) analogues for single photon emission tomography (SPET). J. Labelled Compd. Radiopharm. 42, 423-435. (9) Reubi, J. C. (2000) In vitro evaluation of VIP/PACAP receptors en healthy and diseased human tissues. Clinical implications. Ann. N. Y. Acad. Sci. 921, 1-25. (10) Evers, B. M., Ishizuka, J., Chung, D. H., Townsend, C. M., and Thompson, J. C. (1992) Neurotensin expression and release in human colon cancers. Ann. Surg. 216 (4), 423431. (11) Velcich, A., Palumbo, L., Jaury, A., Laboisse, C., Racevskis, J., and Augenlicht, L. (1995) Patterns of expression of lineagespecific markers during the in vitro-induced differenciation of HT29 colon carcinoma cells. Cell Growth Differ. 6, 749757. (12) Chetanneau, A., Baum, R. P., Lehur, P. A., Liehn, J. C., Perkins, A. C., Bares, R., Bourguet, P., Herry, J. Y., Saccavini, J. C., and Chatal, J. F. (1990) Multi-centre immunoscintigraphic study using indium-111-labeled CEA-specific and/or 19-9 monoclonal antibody F(ab′)2 fragments. Eur. J. Nucl. Med. 17, 223-229. (13) Le Doussal, J. M., Gruaz-Guyon, A., Martin, M., Gautherot, E., Delaage, M., and Barbet, J. (1990) Targeting of indium111-labeled bivalent hapten to human melanoma mediated by bispecific monoclonal antibody conjugates: imaging of tumors hosted in nude mice. Cancer Res. 50, 3445-3452. (14) Le Doussal, J. M., Chetanneau, A., Gruaz-Guyon, A., Martin, M., Gautherot, E., Lehur, P. A., Chatal, J. F., Delaage, M., and Barbet, J. (1993) Bispecific monoclonal antibody-mediated targeting of an indium-111-labeled DTPA dimer to primary colorectal tumors: pharmacokinetics, biodistribution, scintigraphy and immune response. J. Nucl. Med. 34 (10), 1662-1671. (15) Salacinsky, P. R., McLean, C., Sykes, J. E., Clement-Jones, V. V., and Lowry, P. J. (1981) Iodination of proteins and peptides using a solide phase oxidizing agent 1,3,4,6-tetrachloro-3,6-diphenylglycoluryl (iodogen). Anal. Biochem. 117, 136-141. (16) Gautherot, E., Le Doussal, J. M., Bouhou, J., Manetti, C., Martin, M., Rouvier, E., and Barbet, J. (1998) Delivery of therapeutic doses of radioiodine using bispecific antibodytargeted bivalent haptens. J. Nucl. Med. 39 (11), 1937-1943. (17) Bellon, B., Jelsch, J., and Delaage, M. (1973) The´orie des e´quilibres multiples pour les polyme`res. Biochimie 55, 11591162. (18) Barbet J., Le Doussal, J. M., Gruaz-Guyon, A., Martin, M., Gautherot, E., and Delaage, M. (1993) Computer calculation of multiple binding equilibrium isotherms: application to the binding of bivalent ligands to antibodies interacting with cell surface Fc-receptor. J. Theor. Biol. 165, 321-340.

Hillairet de Boisferon et al. (19) Hillairet de Boisferon, M., Raguin, O., Dussaillant, M., Roste`ne, W., Barbet, J., and Gruaz-Guyon, A. (2000) Enhanced targeting specificity to tumor cells by simultaneous recognition of two antigens. Bioconjug. Chem. 11, 452-460. (20) Mazella, J., Poustis, C., Labbe, C., Checler, F., Kitabgi, P., Granier, C., van Rietschoten, J., and Vincent, J. P. (1983) Monoiodo-(trp11) neurotensin, a highly radioactive ligand of neurotensin receptors. Preparation, biological activity, and binding properties to rat brain synaptic membranes. J. Biol. Chem. 258 (6), 3476-3481. (21) Hillairet de Boisferon, M., Manetti, C., Raguin, O., Gautherot, E., Roste`ne, W., Barbet, J., and Gruaz-Guyon, A. (1997) Pretargeted radioimmunotherapy using iodine-131-labeled bivalent hapten-bearing peptides. Lett. Pept. Sci. 4, 331-339. (22) De Labriolle-Vaylet, C., Cattan, P., Sarfati, E., Wioland, M., Billotey, C., Broche´riou, C., Rouvier, E., de Roquancourt, A., Rostene, W., Askienazy, S., Barbet, J., Milhaud, G., and Gruaz-Guyon, A. (2000) Successful surgical removal of occult metastases of medullary thyroid carcinoma recurrences with the help of immunoscintigraphy and radioimmunoguided surgery. Clin. Cancer Res. 6 (2), 363-371. (23) Kraeber-Bodere, F., Bardet, S., Hoefnagel, C. A., Vieira, M. R., Vuillez, J. P., Murat, A., Ferreira, T. C., Bardies, M., Ferrer, L., Resche, I., Gautherot, E., Rouvier, E., Barbet, J., and Chatal, J. F. (1999) Radioimmunotherapy in medullary thyroid cancer using bispecific antibody and iodine 131labeled bivalent hapten: preliminary results of a phase I/II clinical trial. Clin. Cancer. Res. 5 (10 Suppl), 3190s-3198s. (24) Reubi, J. C., Waser, B., Schaer, J. C., and Laissue, J. A. (1999) Neurotensin receptors in human neoplasms: high incidence in Ewing’s sarcomas. Int. J. Cancer 82, 213-218. (25) Reubi, J. C., Waser, B., Friess, H., Bu¨chler, M., and Laissue, J. (1998) Neurotensin receptor: a new marker for human ductal pancreatic adenocarcinoma. Gut 42, 546-550. (26) Behr, T. M., Sharkey, R. M., Juweid, M. E., Dunn, R. M., Vagg, R. C., Ying, Z., Zhang, C. H., Swayne, L. C., Vardi, Y., Siegel, J. A., and Goldenberg, D. M. (1997) Phase I/II clinical radioimmunotherapy with an iodine-131-labeled anti-carcinoembryonic antigen murine monoclonal antibody IgG. J. Nucl. Med. 38 (6), 858-870. (27) Gaudriault, G., and Vincent, J. P. (1992) Selective labeling of R- or -amino groups in peptides by the Bolton-Hunter reagent. Peptides 13, 1187-1192. (28) Le Doussal, J. M., Gautherot, E., Martin, M., Barbet, J., and Delaage, M. (1991) Enhanced in vivo targeting of an asymmetric bivalent hapten to double-Ag-positive mouse B cells, using mAb conjugate cocktails. J. Immunol. 146 (1), 169-175. (29) Kitabgi, P., Checler, F., Mazella, J., and Vincent, J. P. (1985) Pharmacology and biochemistry of neurotensin receptors. Rev. Clin. Basic. Pharm. 5, 397-486. (30) Juweid, M., Sharkey, R. M., Behr, T., Swayne, L. C., Herskovic, T., Pereira, M., Rubin, A. D., Hanley, D., Dunn, R., Siegel, J., and Goldenberg, D. M. (1996) Radioimmunotherapy of medullary thyroid cancer with iodine-131-labeled anti-CEA antibodies. J. Nucl. Med. 37 (6), 905-911.

BC015585G