In vitro evaluation of avidin antibody pretargeting using 211At-labeled and biotinylated poly-L-lysine as effector molecule


  • The articles in this supplement were presented at the “12th Conference on Cancer Therapy with Antibodies and Immunoconjugates,” in Parsippany, New Jersey, October 16-18, 2008.



Pretargeting is an approach for enhancing the therapeutic index of radioimmunotherapy by separating the administrations of tumor-targeting substance and radiolabel. In this study, a pretargeting model system of avidin-conjugated monoclonal antibody trastuzumab and biotinylated, 211At-labeled poly-L-lysine was constructed and analyzed in vitro.


Avidin activated by 4-(N-maleimidomethyl)cyclohexane-1-carboxylic acid 3-sulfo-N-hydroxysuccinimide ester sodium salt (sulfo-SMCC) and thiolated trastuzumab were incubated overnight at 4°C. The monomeric fraction was extracted using size exclusion fast protein liquid chromatography (FPLC) and further purified on an iminobiotin affinity column. Poly-L-lysine was biotinylated with succinimidyl-6-(biotinamido)hexanoate (NHS-LC-biotin), followed by direct 211At-labeling with N-succinimidyl-3-(trimethylstannyl)benzoate (m-MeATE), and succinylation with succinic anhydride. The avidin-trastuzumab conjugate was characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and FPLC, together with cell-binding and biotin-binding analyses. The labeled poly-L-lysine conjugate was assessed in terms of radiochemical purity and avidin binding. Furthermore, the full pretargeting system was evaluated in a tumor cell binding assay.


The estimated size of the pretargeting molecule was 220 kDa, which corresponds to that of the expected avidin-trastuzumab monomer. Neither cell-binding ability (64%) nor biotin-binding ability (85%-95%) indicated any severe adverse effects from the chemical modifications. The radiochemical purity of the effector molecule was 92%-97%, and the avidin binding capacity was 91%-93%. The complete pretargeting assay resulted in a binding of 75.3 ± 6.2% of added effector molecules to cells.


The high binding of effector molecules to cells demonstrates a proof of concept for the synthesized molecules and pretargeting system, which will be further evaluated in vivo in future studies. Cancer 2010;116(4 suppl):1101–10. © 2010 American Cancer Society.

Radioimmunotherapy (RIT) uses antibodies as carriers of radionuclides that emit cell-killing radiation. Ideally, tumors are targeted by the labeled antibodies and subsequently eradicated as the radionuclide decays. Unfortunately, the large size (approximately 150 kDa) of the antibodies reduces the efficacy because of slow diffusion, which leads to slow uptake in tumors. The large size also impedes clearance, resulting in unwanted irradiation of normal tissue and a risk of ensuing toxicity.

Several researchers have reported encouraging results for enhancing the therapeutic index of RIT by using a technique that is generally referred to as pretargeted radioimmunotherapy, or simply pretargeting.1-4 Different schemes have been developed, all sharing 1 feature: separate administration of the tumor-targeting substance (pretargeting molecule) and the radiolabeled substance (effector molecule), thus allowing time for efficient targeting and enabling faster uptake of radioactivity in tumors. There are primarily 2 strategies used to ensure a high affinity between the effector and pretargeting molecules: bispecific antibodies5 and the avidin/streptavidin-biotin system.6 In addition, the use of complementary oligonucleotides has also been studied for pretargeted radioimmunotherapy.7, 8

Avidin is a 66 kDa egg white glycoprotein with 4 binding sites for biotin, a small, naturally occurring vitamin. The high affinity (Kd = 10−15 M) between avidin and biotin is one of the strongest known noncovalent interactions, which can be used in pretargeting by coupling avidin to the antibody and using an effector molecule comprising biotin and a radionuclide. Previous studies have concluded that this approach is preferred over the opposite, ie, biotinylating the antibody and using an avidin-based effector molecule.3 Avidin is, in in vitro research, generally interchangeable with streptavidin, its bacterially produced nonglycosylated analog. For in vivo applications, the difference in glycosylation makes the proteins suitable for different purposes.

There are several ways of conjugating avidin/streptavidin to antibodies. Robust methods using bifunctional chemical reagents have been developed that sustain the multivalency of avidin for biotin, ensuring high avidity between pretargeting and effector molecules.9-11 Fusion proteins, which are produced by DNA techniques, are advantageous because of superior homogeneity and straightforward scale-up, but they are time consuming and require great effort in their development and optimization.12, 13

The avidity of the effector molecule for avidin can also be adjusted by careful design of the biotin-conjugate. A few studies have been performed in which poly-L-lysine has been used as a joint between biotin and the radionuclide.14-17 The advantage of the polymer is the number of ϵ-amino groups available for N-acylation reagents, making bifunctionate conjugates with multiple biotins attached possible, as well as intermediate reagents or chelates for radiolabeling. Consequently, both avidity and specific radioactivity can be increased compared with labeling biotin. Another benefit of using polymers with variable molecular weights is the possibility of increased control over the in vivo distribution.

An array of nuclides, including both alpha and beta emitters, is available for radioimmunotherapy for different kinds of tumors. Among the beta emitters, 90Y is frequently used for pretargeting,1, 18, 19 but other radionuclides such as 131I20, 21 and 177Lu22, 23 have also been studied. For the treatment of disseminated cancer in the form of micrometastases, the relatively long range (0.05-12 mm) of beta emitters is unfavorable, yielding low tumor-to-normal tissue ratios. Alpha emitters, conversely, have a short particle range (40-100 μm) and high linear energy transfer (approximately 80-100 keV/μm), which make them advantageous for treating micrometastatic disease.24 Several alpha emitters have been evaluated for therapy, including 211At (t½ = 7.2 hours), 212Bi (t½ = 60.6 minutes), 213Bi (t½ = 45.6 minutes), 223Ra (t½ = 11.4 days), and 225Ac (t½ = 10.0 days).25 Among these nuclides, 211At is perhaps the most applicable, although only a few research facilities have the cyclotron needed for its production. The daughters of 211At decay are of no major concern, and astatine labeling of different conjugates is feasible. The manageable half-life also allows time for chemical synthesis and for the carrier molecule to distribute in vivo.

In this study, avidin-coupled monoclonal antibodies and biotinylated and radiolabeled poly-L-lysine conjugates were synthesized for pretargeted 211At-radioimmunotherapy. To simplify conjugation and purification, a large poly-L-lysine conjugate was constructed as a model for future studies of much smaller molecules. The molecules were evaluated in vitro, analytically using gel electrophoresis and chromatography, as well as functionally in tumor cell binding assays, to assess their potential usefulness for prospective treatment of disseminated cancer.



Avidin from egg white, succinimidyl-6-(biotinamido)hexanoate (EZ-link NHS-LC-biotin), 2-iminothiolane hydrochloride (2IT), 4-(N-maleimidomethyl)cyclohexane-1-carboxylic acid 3-sulfo-N-hydroxysuccinimide ester sodium salt (sulfo-SMCC), and poly-L-lysine were purchased from Sigma-Aldrich Sweden AB (Stockholm, Sweden). The human cell line SKOV-3, an ovarian carcinoma cell line, was obtained from American Type Culture Collection (Rockville, Md). The tumor cell lines were cultured at the Department of Oncology at Sahlgrenska University Hospital (Gothenburg, Sweden). No-carrier added (125I)NaI was purchased from PerkinElmer Sverige AB (Upplands Väsby, Sweden). The monoclonal antibody (MoAb) trastuzumab (Herceptin) was obtained from Apoteket AB, Sahlgrenska University Hospital (Gothenburg, Sweden). This MoAb is specific for human epidermal growth factor ErbB2 (Her2). The N-succinimidyl-3-(trimethylstannyl)benzoate reagent was purchased from Toronto Research Chemicals Inc. (Canada).

The 211At was produced at the Cyclotron and PET Unit, Rigshospitalet (Copenhagen, Denmark) via the 209Bi(α,2 n)211At reaction. The irradiated target was transported to the Department of Nuclear Medicine at Sahlgrenska University Hospital (Gothenburg, Sweden), where the astatine was transformed into a chemically useful form by dry distillation as previously described.26

Radioactivity Measurements

Radioactivity measurements of high-activity samples (>100 kBq) were performed in an ionization chamber (Capintec, CRC-15 dose calibrator, Iowa). Radioactivity measurements of low-activity samples (<10 kBq) were conducted using a NaI(Tl) γ-counter (Wizard 1480, Wallac, Finland).

Pretargeting Molecule

The avidin-antibody conjugation was performed essentially as previously described by Foulon et al.9

Activation of Avidin with Sulfo-SMCC

Avidin in 0.2 M carbonate buffer (pH 8.5) was mixed with 10 times molar excess of sulfo-SMCC in dimethyl sulfoxide (DMSO) and incubated at room temperature for 30-60 minutes. The conjugate was purified by passage through a Sephadex NAP-5 column (GE Healthcare, Uppsala, Sweden).

Activation of Trastuzumab with 2IT

Ten to 20 times molar excess of 2IT in 0.9% NaCl/2 mM EDTA was added to trastuzumab in phosphate buffered saline (PBS) and incubated at room temperature for 30-60 minutes. The thiolated antibody was purified by passage through a Sephadex NAP-5 column.

Preparation and Purification of the Avidin-Trastuzumab Conjugate

The thiolated MoAb and sulfo-SMCC-activated avidin were mixed at an equimolar ratio and incubated overnight at 4°C. The monomeric fraction was isolated from unreacted antibody and undesired larger complexes by size exclusion fast protein liquid chromatography (FPLC) on a Superdex-200 column using an ÄKTA-FPLC system (GE Healthcare, Uppsala, Sweden) with UV detection at 280 nm and collection of desired fractions. An iminobiotin affinity column was used to remove nonavidin-bound conjugates. The column was equilibrated with 50 mM sodium-carbonate buffer with 500 mM NaCl (pH 9.5) and the avidin-bound fraction was eluted using 50 mM sodium acetate-acetic acid with 500 mM NaCl (pH 4).

To assess the cell-binding and biotin-binding ability of the pretargeting molecule, the antibody conjugate was labeled with 211At using a direct procedure with the reagent N-succinimidyl-3-(trimethylstannyl)benzoate (m-MeATE).27

Conjugation with m-MeATE

The m-MeATE reagent was dissolved in chloroform at a concentration of 50 mg/mL, and 2 μL of that stock solution was extracted to a glass vial. The chloroform was evaporated and the reagent was redissolved in DMSO to a final concentration of 17 mM. The avidin-trastuzumab conjugate was prepared in 0.2 M sodium carbonate buffer (pH 8.5) at a concentration of 2 mg/mL. The m-MeATE reagent was used at a molar excess of between 5 and 10 over the antibody conjugate in the following conjugation. During vigorous agitation, 0.5-1.0 μL of m-MeATE was added to 100 μL of antibody conjugate, and the reaction proceeded for 30 minutes with gentle agitation at room temperature. The m-MeATE-avidin-trastuzumab conjugate was isolated on a Sephadex NAP-5 column and eluted with 0.9 mL of 0.2 M sodium acetate buffer (pH 5.5).


After dry distillation of astatine, the radionuclide was dispensed in chloroform, which was subsequently evaporated, leaving a dry residue in the vial. Immediately before labeling, the astatine was reacted with 0.1-1 nmole N-iodosuccinimide (130 μM) in methanol/1% acetic acid to oxidize the radionuclide into a reactive form. Without delay, 75-100 μg of the m-MeATE-avidin-trastuzumab conjugate, at a concentration of 0.25-0.44 mg/mL, was added to the vial and the reaction mixture incubated for 1 minute with agitation at room temperature. Remaining stannyl groups on the conjugate were iodo-substituted by the addition of a 20-fold molar excess of N-iodosuccinimide in methanol/1% acetic acid (18 mM) over antibody, and the solution was yet again incubated for 1 minute with agitation. Before purification of the labeled antibody conjugate on a Sephadex NAP-5 column, any unreacted 211At was reduced by the addition of 5 μL sodium ascorbate in water at a concentration of 50 mg/mL. The labeled conjugate was eluted with 0.9 mL PBS.

Effector Molecule

Biotinylation of Poly-L-lysine

Poly-L-lysine (approximately 130 kDa) was dissolved in 0.2 M carbonate buffer (pH 8.5) to 2 mg/mL. EZ-link NHS-LC-biotin was dissolved in dimethylformamide to 2 mg/mL and immediately added to the poly-L-lysine solution at a 10-fold molar excess. The reaction mixture was incubated at room temperature for 30 minutes with agitation.

Labeling of the unrefined biotinylated poly-L-lysine was performed by a direct procedure modified from the method of Lindegren et al.27

Conjugation with m-MeATE

Five to 10 equivalents of m-MeATE were added to 1-2 mg of biotinylated poly-L-lysine in 0.2 M carbonate buffer (pH 8.5) during vigorous agitation. After reacting for 30 minutes with gentle agitation, the biotinylated and m-MeATE-conjugated poly-L-lysine was isolated in a Microsep 30K centrifugal device (MW range 90-180 kDa, Pall Life Sciences, Port Washington, NY) by repeated centrifugation and washing with 0.2 M sodium acetate buffer (pH 5.5).

Labeling with 211At

The labeling of the polymer was performed mainly by the same route as the avidin-coupled antibody but with a few modifications. To the dry astatine residue, 0.1-0.2 nmole of N-iodosuccinimide in methanol/1% acetic acid (130 μM) was added, followed by 75-500 μg of the m-MeATE-poly-L-lysine conjugate (3.5-6.8 mg/mL). After incubation for 1 minute with agitation, N-iodosuccinimide (18 mM) was added in a 20-fold molar excess, and the solution was yet again incubated for 1 minute. Before eliminating unbound astatine, the labeled poly-L-lysine conjugate was succinylated by the addition of solid succinic anhydride to modify the charge of the molecule by the conversion of remaining amino groups to carboxylic residues. During the reaction, the pH was continuously adjusted by the addition of 1 M Na2CO3 to keep remaining amino groups unprotonated. The succinylation was performed for 15 minutes, after which the solution was added to a Sephadex NAP-5-column to isolate the labeled polymer fraction. The column was eluted with 0.9 mL PBS.

Labeling with 125I

The iodination of biotinylated poly-L-lysine was performed according to the same direct method as the 211At-labeling above, with the difference that N-bromosuccinimide was used instead of N-iodosuccinimide.

The synthesis of the effector molecule is described in Figure 1.

Figure 1.

Synthesis of 211At-labeled and biotinylated poly-L-lysine is depicted as follows: (1) Biotinylation with succinimidyl-6-(biotinamido)hexanoate (NHS-LC-biotin), pH 8.5, 30 minutes, (2) Conjugation with N-succinimidyl-3-(trimethylstannyl)benzoate (m-MeATE), pH 8.5, 30 minutes, (3) Labeling, 211At/N-iodosuccinimide alt. 125I/N-bromosuccinimide, pH 5.5, 2 minutes, (4) Succinylation with succinic anhydride, pH 8.5, 15 minutes. X = 211At or 125I.

In Vitro Quality Control

The radiochemical purity (RCP) of the labeled products was evaluated in triplicate at room temperature by methanol precipitation and trichloroacetic acid precipitation for 211At- and 125I-labeling, respectively, as well as by size exclusion chromatography.

Methanol Precipitation

Placed in tubes were 200 μL of 1% bovine serum albumin in PBS, 1-5 kBq aliquots of the labeled conjugate, and 500 μL of methanol. The total activity in each tube was measured in a gamma counter, and the samples were centrifuged for 3 minutes at 3000 rpm. From all test tubes, 200 μL of the supernatant was withdrawn and measured for radioactivity in the gamma counter. The radiochemical purity was then calculated according to the formula:

equation image

where Asup equals the activity in the supernatant, Abg is the background activity, Atot refers to the total activity added to the tube, and vtot and vsup represent the initial volume and the volume of the extracted supernatant, respectively.

Trichloroacetic Acid Precipitation

The trichloroacetic acid precipitation was performed according to the above description of methanol precipitation, but with 10% trichloroacetic acid instead of methanol.

Size Exclusion Chromatography

Size exclusion fast protein liquid chromatography was performed on a Superdex 200-column and an ÄKTA-FPLC system with UV detection at 225 nm. Aliquots of the radiolabeled effector molecule with an activity of approximately 1 MBq were added to the column and eluted with PBS at a flow rate of 0.5 mL/min. Fractions of 0.65 mL were collected and measured in a gamma counter. The distribution of the radioactivity was compared with the UV-chromatograms to determine the quality of the labeled product. FPLC analysis was also performed for the purified pretargeting molecule to determine the composition of the final product.

Gel Electrophoresis

The molecular weight of the purified pretargeting molecule was estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The samples were run on a 4-12% Bis-Tris gel (Invitrogen, Carlsbad, California) under nonreducing conditions, and the proteins were detected by Coomassie staining.

Biotin Binding Capacity

A volume of 100 μL iminobiotin-linked agarose beads was added to a polypropylene microcentrifuge filter tube (Corning Costar Spin-X; Sigma-Aldrich Sweden AB, Stockholm, Sweden) with 2.5 μg of the 211At-labeled pretargeting molecule in 25 μL PBS. The sample was incubated at room temperature for 1 hour with gentle agitation. The tubes were then centrifuged for 1 minute at 3000 rpm and washed twice with 100 μL PBS. The bead-containing filter was extracted from the tube, and the biotin binding capacity was determined as the bead-associated activity divided by the total applied activity.

Avidin Binding Capacity

The ability of the biotinylated effector molecule to bind avidin was evaluated by adding 5 ng of 211At-labeled and biotinylated poly-L-lysine in 25 μL PBS to 100 μL avidin-linked agarose beads in a Spin-X centrifuge tube filter. The analysis was then performed in the same manner as biotin binding described above.

Cell Binding

The cell binding was determined in duplicate for the 211At-labeled avidin-trastuzumab conjugate by binding to SKOV-3 cells. The cells were prepared in single-cell suspension at a concentration of 5 × 106 cells/mL. A constant amount (2.5 μg) of 211At-avidin-trastuzumab (0.2 MBq/μg) was added, and the cells were incubated at room temperature with gentle agitation for 2 hours. After washing the cells, the bound activity was measured and the cell binding was calculated as the ratio between bound activity and total applied activity.

Pretargeting Assay

The potency of the pretargeting system was also evaluated by binding to SKOV-3 cells. The cell assay was prepared as described above, and 2.5 μg of avidin-trastuzumab was added to each dilution and incubated for 2 hours. The amount of added pretargeting molecule was based on previous estimations of the number of binding sites for trastuzumab on SKOV-3 cells (roughly 2 × 106 sites/cell) and determined so as to ensure that the antibody-conjugate was in excess over available binding sites. After incubation, the cells were centrifuged for 3 minutes at 3000 rpm, washed with PBS, and resuspended in 500 μL of cell medium. A constant amount (5 ng) of 211At-labeled and biotinylated poly-L-lysine was then added to all dilutions and incubated for 1 hour, after which the cells were once again centrifuged and washed with PBS. Finally, the cells were measured for bound activity and the bound fraction was estimated from the collected data.

Nonspecific Binding

Nonspecific binding was assessed for the avidin-trastuzumab conjugate and for the astatinated and biotinylated polymer. For the pretargeting molecule, the antigens on the SKOV-3 cells were saturated with an excess of trastuzumab before adding the modified monoclonal antibody. The effector molecule was evaluated by single administration directly to the cells, which were measured for bound activity after 1-hour incubation.


Pretargeting Molecule

In this study, the aim was to evaluate the properties of the produced pretargeting molecule with regard to size, affinity for biotin, and cell-binding capacity. Figure 2A shows a typical chromatogram resulting from the purification of the product by FPLC. A large portion of the MoAb appeared to not react, and a small fraction of unwanted, larger complexes also formed. The desired fraction of monomeric avidin-trastuzumab constituted 17%-26% of the initially added antibody, as determined by dividing the area under the avidin-MoAb peak with the total MoAb-area in the chromatogram. However, the overall yield was smaller because a considerable fraction of the product was lost in the final purification by iminobiotin chromatography. The overall yield, calculated as the extracted avidin-trastuzumab conjugate divided by initial amount of MoAb, was 2%-11%. Figure 2B depicts the purified avidin-trastuzumab conjugate, showing that a fraction (17%) of larger complexes remain in the final solution.

Figure 2.

Chromatograms of the UV absorbance at 280 nm from fast protein liquid chromatography results from (A) the reaction of avidin activated by 4-(N-maleimidomethyl)cyclohexane-1-carboxylic acid 3-sulfo-N-hydroxysuccinimide ester sodium salt (sulfo-SMCC-avidin) (10:1) and thiolated MoAb (2IT-trastuzumab) (10:1) and (B) purified avidin-trastuzumab conjugate. The peaks at 25.84 minutes and 23.84-23.96 minutes represent unreacted MoAb and the desired monomeric avidin-MoAb fraction, respectively. The peaks at 20.95-21.40 minutes correspond to unwanted larger conjugates, and the peak at 29.59 minutes represents avidin.

Figure 3 shows the MW determination of the pretargeting molecule by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The molecular weight was estimated to be approximately 220 kDa, which is in agreement with that of the expected avidin-trastuzumab monomer. The resulting gel also confirms the presence of larger conjugates, as a barely visible band above the monomer fraction.

Figure 3.

Coomassie stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis of molecular mass standards are shown in lane 1, unmodified trastuzumab in lane 2, and avidin-trastuzumab in lane 3. The arrow points at the avidin-trastuzumab monomer.

For 211At-labeling of the pretargeting molecule, the radiochemical yield was in the range of 43%-62% with a radiochemical purity of 96%-98%, as determined by methanol precipitation. The cell binding of the 211At-labeled pretargeting molecule was 64% including a nonspecific binding of less than 10%. The biotin binding capacity was excellent with 86%-95% of the labeled pretargeting molecule bound to the iminobiotin beads after 1 hour.

Effector Molecule

For 211At-labeling of the biotinylated poly-L-lysine, the radiochemical yield was in the range of 26%-77%. For 125I-labeling, the corresponding value was 74%-90%. The radiochemical purity of the 211At-labeled effector molecule was in the range of 92%-97%. The trichloroacetic acid precipitation resulted in a corresponding value of 94%-99% for the iodinated product. The avidin binding ability was also good; after 1 hour, 91%-93% of the added effector molecule had bound to immobilized avidin.

Size Exclusion Chromatography

Figure 4 illustrates a FPLC chromatogram of 211At-labeled and biotinylated poly-L-lysine with the activity distribution attained from the gathered fractions. In the chromatogram, the activity peak overlaps well with the poly-L-lysine peak at around 17 minutes, which underlines the results from the methanol precipitation concluding that the radiochemical purity of the labeled product is good. Note that succinylation of the poly-L-lysine increased the molecular weight by approximately 70%. The chromatogram also depicts a peak at 38 minutes, which most likely corresponds to succinic acid, a residue from the succinylation after labeling the biotinylated poly-L-lysine.

Figure 4.

Chromatogram of the UV absorbance at 225 nm from fast protein liquid chromatography of 211At-labeled and biotinylated poly-L-lysine is accompanied by the activity distributed in the collected fractions. Each dot represents the total activity in 1 fraction.

Pretargeting Assay

After incubation for 1 hour, 75.3 ± 6.2% (SD, n = 5) of the added effector molecules had bound to the avidin-trastuzumab pretargeted SKOV-3-cells. The binding is illustrated in Figure 5. The unspecific binding of the effector molecule was less than 1%.

Figure 5.

Illustrated is the binding of 211At-labeled and biotinylated poly-L-lysine to human carcinoma cells (SKOV-3) pretargeted with avidin-trastuzumab after incubation for 1 hour (mean ± SD, n = 5).


Pretargeting takes many different forms, all aimed at lowering the irradiation of healthy tissue while still delivering high doses to tumors, thus enhancing the therapeutic index of conventional radioimmunotherapy. The advantages of pretargeted radioimmunotherapy are many: in addition to the benefits from faster localization of radioactivity to the targeted tumor, smaller-sized effector molecules penetrate the tumor tissue more efficiently, and molecules that are not bound will clear rapidly from the circulation. Faster diffusion is especially important when using radionuclides with short half-life, such as 211At and 213Bi. The shift from direct antibody radiolabeling to labeling a second molecule will also spare the targeting agent (ie, the antibody) from radiation damage. By separating the targeting molecule and radionuclide-carrying vehicles, higher specific radioactivity can be achieved without compromising the antigen binding of the antibody.

The immunoreactivity of the pretargeting molecule can be significantly decreased by chemical modification with avidin if the conjugation takes place close to the antibody's antigen binding site. Nevertheless, in this study, the results from the analysis of the labeled pretargeting molecule and the full pretargeting system suggest that there were no major adverse effects on binding ability.

Avidin/streptavidin and biotin are widely used for pretargeting purposes, but there are a few issues that speak against the avidin-biotin system in some in vivo situations. The role of endogenous biotin has been debated, but it seems to be a concern mainly in preclinical mice models and can be controlled by the use of biotin-deficient diet.1, 6, 28-31 Another possible problem is biotinidase-induced catabolism of biotin. In response to that, Foulon et al developed biotinidase-resistant compounds that are stable in murine serum, human serum, and cerebrospinal fluid.32 Furthermore, questions have been raised regarding the immunogenicity of avidin/streptavidin, which may hinder fractionated treatment for patients that develop an immunogenic response with the primary pretargeting treatment.3, 4, 33

The choice between avidin and streptavidin depends on the intended route of administration. Streptavidin is a nonglycosylated bacterial analog of avidin and does not follow the same course of clearance as avidin. While avidin rapidly accumulates in the liver, streptavidin remains in the blood for a longer time and is primarily cleared by the kidneys.6 This difference in in vivo behavior makes avidin more suited for intracavitary treatments, such as intraperitoneal treatments, and streptavidin more suited intravenously. Ultimately, we are aiming for systemic treatment with streptavidin-MoABs, but the results from these in vitro experiments with avidin-conjugated MoAbs are valid for both avidin and streptavidin.

The route of administration also determines the need for an additional injection of a clearing agent between the delivery of the pretargeting and effector molecules. For the intraperitoneal treatment of ovarian carcinoma using a 2-step pretargeting system with avidin-conjugated MoAb as the pretargeting molecule, no clearing agent will probably be needed because avidin directs the MoAb-conjugate clearance via the liver once it enters the circulation. An interesting alternative for the systemic administration of avidin/biotin conjugates is the use of extracorporeal affinity adsorption.34-36

The polymer-based effector molecule described in this study was synthesized from poly-L-lysine with an average molecular weight of 130 kDa. This effector molecule size is probably too large for most in vivo applications, and it was chosen mainly as a model to simplify synthesis and purification. The methods developed will be used with minor modifications in further studies focusing on evaluating effector molecules based on much smaller, well-defined poly-L-lysine molecules with sizes ranging from approximately 1500 Da to 4400 Da. Concerning the astatination, an experimental technique was used in which the biotinylated polymer was labeled by a direct procedure. Although biotin is thought to be sensitive to oxidative agents,37 it appears that the avidin binding capacity is intact even after exposing the biotinylated polymer to N-iodosuccinimide, as well as N-bromosuccinimide, as proven by the good binding to immobilized avidin and cells pretargeted with avidin-trastuzumab.

To summarize, avidin-coupled trastuzumab and biotinylated and 211At-labeled poly-L-lysine conjugates were synthesized. The objective was not to maximize the yields of either the conjugated pretargeting molecule or the radiolabeled effector molecule. Instead, the study aimed for a proof of concept concerning this particular system of pretargeting and effector molecule and to determine the tumor cell-bound fraction that could be reached. The achieved avidin-antibody conjugation yields were in the same order as those previously obtained by Foulon et al,9 and the estimated size of the conjugate corresponded to that of the expected avidin-trastuzumab monomer. However, for future studies that demand a larger amount of the pretargeting molecule, the iminobiotin chromatography procedure needs to be improved to increase yield of the purification of the product.

Concerning the pretargeting assay, the high binding of the effector molecule to the pretargeted cells clearly demonstrates a proof of concept for the synthesized molecules and pretargeting system. The system will be further evaluated in vivo in a preclinical nude mice model, including the biodistribution of the pretargeting and effector molecule, as well as therapeutic efficacy of the full-scale system.


The authors thank Ingela Claesson and Helena Kahu at the Department of Oncology, University of Gothenburg, Sweden, for culturing the tumor cells. Thanks are also due to Elin Haglund, Jörgen Elgqvist, Tom Bäck, and Lars Jacobsson at the Department of Radiation Physics at the University of Gothenburg, Sweden. This work was supported by grants from the Swedish Cancer Society, the King Gustaf V Jubilee Clinic Research Foundation in Gothenburg, Sweden, and the Swedish Research Council.


The articles in this supplement represent proceedings of the “12th Conference on Cancer Therapy with Antibodies and Immunoconjugates,” in Parsippany, New Jersey, October 16-18, 2008. Unrestricted grant support for the conference was provided by Actinium Pharmaceuticals, Inc., Bayer Schering Pharma, Center for Molecular Medicine and Immunology, ImClone Systems Corporation, MDS Nordion, National Cancer Institute, NIH, New Jersey Commission on Cancer Research, and PerkinElmer Life & Analytical Sciences. The supplement was supported by an unrestricted educational grant from ImClone Systems Corporation, a wholly owned subsidiary of Eli Lilly and Company, and by page charges to the authors. This study was supported by the Swedish Cancer Society, the King Gustaf V Jubilee Clinic Research Foundation in Gothenburg, Sweden, and the Swedish Research Council.