Antibody−Dendrimer Conjugates - American Chemical Society

Mar 25, 2008 - Radioimmunotherapy using antibodies with favorable tumor targeting properties and high binding affinity is increasingly applied in canc...
0 downloads 0 Views 1MB Size
Bioconjugate Chem. 2008, 19, 813–820

813

Antibody-Dendrimer Conjugates: The Number, Not the Size of the Dendrimers, Determines the Immunoreactivity C. Wängler, G. Moldenhauer, M. Eisenhut, U. Haberkorn, and W. Mier* Universitaetsklinikum Heidelberg, Abteilung Nuklearmedizin, Germany. Received August 16, 2007; Revised Manuscript Received February 21, 2008

Radioimmunotherapy using antibodies with favorable tumor targeting properties and high binding affinity is increasingly applied in cancer therapy. The potential of this valuable cancer treatment modality could be further improved by increasing the specific activity of the labeled proteins. This can be done either by coupling a large number of chelators which leads to a decreased immunoreactivity or by conjugating a small number of multimeric chelators. In order to systematically investigate the influence of conjugations on immunoreactivity with respect to size and number of the conjugates, the anti-EGFR antibody hMAb425 was reacted with PAMAM dendrimers of different size containing up to 128 chelating agents per conjugation site. An improved dendrimer synthesis protocol was established to obtain compounds of high homogeneity suitable for the formation of defined protein conjugates. The quantitative derivatization of the PAMAM dendrimers with DOTA moieties and the characterization of the products by isotopic dilution titration using 111In/natIn are shown. The DOTA-containing dendrimers were conjugated with high efficiency to hMAb425 by applying Sulfo-SMCC as cross-linking agent and a 10- to 25fold excess of the thiol-containing dendrimers. The determination of the immunoreactivities of the antibody-dendrimer conjugates by FACS analysis revealed a median retained immunoreactivity of 62.3% for 1.7 derivatization sites per antibody molecule, 55.4% for 2.8, 27.9% for 5.3, and 17.1% for 10.0 derivatization sites per antibody but no significant differences in immunoreactivity for different dendrimer sizes. These results show that the dendrimer size does not influence the immunoreactivity of the derivatized antibody significantly over a wide molecular weight range, whereas the number of derivatization sites has a crucial effect.

INTRODUCTION Radioimmunotherapy is an expanding field in cancer therapy and shows very promising treatment results (1–4). However, the potential of this treatment modality has not yet been completely exploited. Besides heterogeneity in antigen expression, tumor load, perfusion defects, and tumor hypoxia, a further reason for this restraint resides in the low specific activity of the radiolabeled antibodies, which does not enable the delivery of optimal radioactive doses to tumor tissues and results in ineffective tumor treatment. A solution for this problem is to introduce more than one radionuclide per protein to enhance the specific activity and by this the radioactive dose that can be applied to the tumor. However, it has been shown already that a few randomly distributed chelators or iodine atoms introduced into an IgG lead to a significant decrease of the immunoreactivity (5, 6). Consequently, an alternative method of introducing many chelators into an antibody molecule that does not lead to a drastic alteration of the properties of the protein is warranted. Recently, some studies about dendritic systems associated with antibodies were undertaken to solve this problem. The synthesized antibody-dendrimer conjugates showed preserved immunoreactivities of the antibodies and provided high specific activities (7–10). In this study, we systematically investigated the influence of introduced dendritic structures with respect to size and the number of derivatization sites per protein. Using the anti-EGFR antibody hMAb425, we assessed the influence of DOTAderivatized PAMAM dendrimers of various dendrimer genera* Universitaetsklinikum Heidelberg, Abteilung Nuklearmedizin, INF 400, Heidelberg, 69120, Germany. E-mail [email protected].

tions and conjugation load on the immunoreactivity of antibodies. This study gives insights into a favorable derivatization pattern for antibody-PAMAM-dendrimer conjugates with a minimal loss of immunoreactivity, suited for improved radioimmunotherapy and diagnostic imaging.

EXPERIMENTAL PROCEDURES General. All commercially available chemicals were of analytical grade and used without further purification. HPLC solvents containing TFA were prepared using 0.1% TFA (v/v). The analytical HPLC system used was an Agilent 1100 system together with a Chromolith Performance (RP-18e, 100– 4.6 mm, Merck, Germany) column. Semipreparative HPLC purifications were performed using a Gyncotech P-580 system (Germering, Germany) equipped with a variable SPD 6-A UV detector and a C-R5A integrator (both Shimadzu, Duisburg, Germany). The column applied for semipreparative purification was a Chromolith (RP-18e, 100–10 mm, Merck, Germany). MALDI-TOF spectroscopy was performed using a Kratos Analytical Compact Maldi III system. ESI spectra were obtained using a Triple-Quadrupole-mass spectrometer TSQ 7000 (Thermo Fisher Scientific, Bremen). NMR spectra were taken using a Varian Mercury Plus 300 MHz and a Varian NMR System 500 MHz, respectively. Thin layer chromatography was performed using Polygram SIL G/UV254 TLC-plates (Macherey-Nagel, Düren, Germany). The PBS buffer consisted of 0.05 M phosphate, 0.15 M NaCl, at pH 7.2. Size-exclusion gel chromatography was carried out using NAP-5 columns (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) and size-exclusion FPLC was performed on a Superdex 200 10/30 GL column (Amersham Biosciences AB, Uppsala, Sweden). 111In was purchased from Perkin-Elmer (Rodgau, Germany).

10.1021/bc700308q CCC: $40.75  2008 American Chemical Society Published on Web 03/25/2008

814 Bioconjugate Chem., Vol. 19, No. 4, 2008

Wängler et al.

Figure 1. HPLC chromatograms of commercially available PAMAM dendrimers with 8 amino functions (A) and 16 amino functions (B). C and D show chromatograms of dendritic structures with 8 and 16 amino functions synthesized as described in the experimental procedures. Chromatograms were recorded at 214 nm and obtained using a Chromolith Performance column and a water/acetonitrile (+0.1% TFA (v/v)) gradient from 0% to 40% acetonitrile over 5 min in A and B. For C and D, a gradient from 20% to 45% acetonitrile over 5 min was used, as the compounds show a higher lipophilicity than the nonderivatized PAMAM dendrimers in A and B.

Figure 2. Synthesis scheme of the penta(ethylene glycol) linker.

Ellman′s Assay. A 2 mM solution of MESNA (sodium 2-mercaptoethanesulfonate) in PB (0.1 M, pH 8) was diluted to obtain solutions of the following concentrations: 125 µM, 64 µM, 32 µM, 16 µM, 8 µM, 4 µM, and 2 µM. To 1 mL of each of these solutions, 200 µL of a solution of Ellman′s reagent (5,5′-dithio-bis-(2-nitrobenzoic acid), 4 mg/mL in PB (0.1 M, pH 8)) was added and the absorption of the obtained mixtures was measured at 412 nm in a UV/vis spectrometer to give the calibration curve. The thiol concentrations of the samples were determined analogously by adding 200 µL of the solution of Ellman′s reagent to 1 mL of sample in PB (0.1 M, pH 8) and by measuring the absorption at 412 nm.

The number of maleimides per dendrimer was determined by reacting it with an access of MESNA (5 equiv) and subsequent titration of free thiol with Ellman′s assay. S-Trityl-mercaptoethyl-tetra(ethylene glycol)-tosylate. To a solution of tetra(ethylene glycol)-bistosylate (22 g, 45 mmol), S-trityl-mercaptoethanol (10 g, 31 mmol), and TBAB (1.5 g, 4.5 mmol) in toluene (200 mL), powdered KOH (∼3 g) was added. After two hours, the mixture was filtered and the resulting solution was evaporated. The remaining oil was purified via column chromatography on silica with n-hexane/acetone 2:1 as the eluent (Rf ) 0.4). The product was obtained as a light yellow oil (11 g, 17 mmol, 55%). S-Trityl-mercaptoethyl-penta(ethylene glycol)-ethylamine (G0) (1). To a solution of S-trityl-mercaptoethyl-tetra(ethylene glycol)tosylate (28.5 g, 59 mmol), and 2-(2-aminoethoxy)ethanol (31 g, 296 mmol) in toluene (100 mL) and DMF (10 mL), powdered KOH (∼5 g) was added. After five hours, the mixture was filtered and the resulting solution was evaporated. The remaining oil was purified via chromatography on silica with CHCl3/ ethanol 5:2 as the eluent (Rf ) 0.37). The product was obtained as a yellow oil (15.5 g, 27 mmol, 45%). Half-Generation PAMAM Dendrimers G0.5 to G6.5 (2, 4, 6, 8, 10, 12, and 14). A solution of the respective fullgeneration (GX) dendrimer in methanol (20 mL) was added to methyl acrylate (200 equiv per amine function) and stirred for 3 to 31 days at room temperature. The volatile components were evaporated and the crude products were purified via chromatography on silica with n-hexane/acetone 4:5, ethanol/acetone

Antibody-Dendrimer Conjugates

Bioconjugate Chem., Vol. 19, No. 4, 2008 815

Figure 3. Synthesis scheme of the PAMAM dendrimers.

5:6, or 55% MeCN in water containing 0.1% TFA for smaller dendrimers and via semipreparative HPLC using a gradient of 0–100% MeCN + 0.1% TFA over 7 min as the eluent for larger dendrimers. The products were obtained as light yellow to orange oils in yields of 30% to 83%. Full-Generation PAMAM Dendrimers G1 to G7 (3, 5, 7, 9, 11, 13, and 15). A solution of the respective half-generation dendrimer (GX.5) in methanol (20 mL) was added dropwise to a cooled solution of ethylenediamine (400 equiv per methyl ester) over a half-hour and stirred for 5 to 41 days at 4 °C until the reaction was complete. Subsequently, the ethylenediamine was evaporated. In the case of smaller dendrimers, the products were found to be clean enough for further use. Larger dendrimers were purified using semipreparative HPLC using a gradient of 0–100% MeCN + 0.1% TFA over 7 min. The products were obtained as yellow to orange oils in yields of 70% to 99%. G0 to G7 Maleimides (16, 19, 22, 25, 28, 31, 34, and 37). A solution of PyBOP (3.9 equiv per amine function of the respective dendrimer) and DIPEA (4 equiv per amine function) in DMF (300 µL) was added to maleimido caproic acid (4 equiv per amine function) and reacted for 2 min. Subsequently, this

mixture was added to a solution of the respective fullgeneration dendrimer (GX) (50 µmol) in DMF (200 µL). After one hour, 1 M NaOH (300 µL) and MeCN (1 mL) were added, and the volatile components were evaporated. After acidification with 1 M HCL (500 µL), the products were purified using semipreparative HPLC using a gradient of 0–100% MeCN + 0.1% TFA over 12 min. The products were isolated as colorless to brown oils in yields of 29% to 84% after lyophilization. G0 to G7 S-trityl-Protected DOTA Dendrimers (17, 20, 23, 26, 29, 32, 35, and 38). To a solution of the respective maleimide-derivatized dendrimer (GX maleimide) (10 µmol) in MeCN (200 µL) was added a solution of thiol-DOTA (4 equiv per maleimide) in 250 µL, and the mixture was reacted for 10 min. After acidification with 1 M HCL (200 µL), the products were purified using semipreparative HPLC using a gradient of 20–100% MeCN + 0.1% TFA over 8 min. The products were isolated as white to ocher solids in yields of 51% to 96% after lyophilization.

816 Bioconjugate Chem., Vol. 19, No. 4, 2008

Wängler et al.

Figure 4. Structures of the studied DOTA-derivatives: tris-tBu-DOTA, tris-allyl-DOTA, tris-benzyl-DOTA, tris-4-nitrobenzyl-DOTA, DOTA-PFP-ester, DOTA–PNP-ester, and thiol-DOTA.

G0 to G7 Deprotected DOTA Dendrimers (18, 21, 24, 27, 30, 33, 36, and 39). The respective S-trityl-protected DOTA dendrimer (10 µmol) was dissolved in a mixture of TIS (200 µL) and TFA (2 mL) and reacted for 5 min. The volatile components were evaporated, and the products were purified using semipreparative HPLC using a gradient of 0–100% MeCN + 0.1% TFA over 7 min. The products were isolated as colorless oils or light ocher solids in yields of 58% to 98% after lyophilization. Dervatization of hMAb425 with Sulfo-SMCC. A freshly prepared solution of Sulfo-SMCC (42 µg, 9.7 × 10-8 mol) in DMF/water 1:1 (10 µL) was added to a fresh solution of hMAb425 (2 mg, 1.39 × 10-8 mol) in 500 µL PBS. After one hour of incubation, the derivatized antibody was purified via size-exclusion chromatography using NAP-5 columns. The derivatized antibody was obtained in high yields (95–98%). The number of maleimide functions per antibody molecule was determined using MESNA and Ellman′s assay and was found to be 1.7. The antibody derivatizations with more equivalents of Sulfo-SMCC were carried out according to this procedure. hMAb425-DOTAX-Dendrimers. To a solution of hMAb425, derivatized with Sulfo-SMCC (1.6 mg, 1.11 × 10-8 mol) in 500 µL PBS, was added a freshly prepared solution of the respective DOTAX-dendrimer (25 equiv, 2.78 × 10-7 mol) in 100 µL PBS. The reaction was complete after 5 min and the products were analyzed using size-exclusion FPLC with PBS as the eluent (Rt between 13.7 and 16.1 min). The conjugation of DOTAX-dendrimers containing a lower number of equivalents of were carried out according to this procedure. Immunoreactivities of the Antibody-Dendrimer Conjugates. Cells. The human colon adenocarcinoma cell line HT29 expressing considerable amounts of epidermal growth factor receptor was maintained in RPMI 1640 medium supplemented with 2 mM L-glutamine, 1 mM pyruvate, and 10% heatinactivated fetal calf serum. Prior to use, cells were detached from the surface of the tissue culture flask by treatment with 0.25% trypsin plus 0.25% EDTA in PBS and counted. Flow Cytometry. Serial dilutions of dendrimer-conjugated matuzumab were compared with unmodified matuzumab for binding to HT29 cells. For this, antibody dilutions ranging from 5 µg/mL to 1 ng/mL were prepared in FACS buffer (Dulbecco’s

PBS with 2% fetal calf serum and 0.1% sodium azide). Staining was performed in a 96-well microtiter plate (U-bottom). 100 µL/well antibody was mixed with 50 µL cell suspension containing 5 × 105 HT-29 cells and incubated for 1 h on ice. Cells were washed twice with FACS-buffer by centrifuging the plate and siphoning off the supernatant. Subsequently, 100 µL/ well F(ab)′2 goat-antihuman IgG (H+L)-FITC (Jackson ImmunoResearch, West Grove, PA) diluted 1:100 in FACS buffer was added as second-step reagent. The plate was incubated for 1 h on ice in the dark, and cells were washed again twice as above. Dead cells were discriminated by propidium iodide staining. Analysis was carried out on a FACScan cytometer (Becton Dickinson, Heidelberg, Germany) using the CELLQUEST software. For each dendrimer conjugate as well as for unconjugated matuzumab, a binding curve was obtained by plotting the mean fluorescence intensity (MFI) versus the antibody concentration.

RESULTS AND DISCUSSION Improvement of PAMAM Dendrimer Synthesis. Dendrimers have shown to be effective drug delivery vehicles for a variety of applications (11–13) and are also the basis for approved pharmaceuticals as, for example, for the diagnostic cardiac marker Stratus and the microbicide Vivagel. In general, a major advantage of dendrimers over polymers is their low dispersity. However, the homogeneity of commercially available dendrimers was found to be low (Figure 1A,B). Similar results have recently been described for commercially available PAMAM dendrimers which were analyzed with HPLC and MALDI-TOF spectrometry showing high polydispersities of the compounds (14, 15). Therefore, these compounds were not chosen as building blocks for the syntheses of the DOTA-multimers. Attempts to synthesize PAMAM dendrimers according to literature methods (16–18) resulted in products with higher quality which, however, did not demonstrate sufficient homogeneity. Therefore, the synthesis protocols had to be improved. The substantial modification was the strict limitation of the reaction temperatures to below 20 and 4 °C for the synthesis of half and full generations, respectively. Reactions conducted above the indicated temperatures led to highly fragmented products and a high amount of intramolecular ring formations, with the consequence of highly heterogeneous dendritic

Antibody-Dendrimer Conjugates

Bioconjugate Chem., Vol. 19, No. 4, 2008 817

Figure 5. Synthesis scheme for the derivatization of the amino groups with maleimides and subsequently with thiol-DOTA as exemplified for the G2 dendrimer (5). Table 1. Characterization Results of the DOTA-Functionalized Dendrimers (n ) 5 and 3 for the Determination of the Number of Maleimides and Functional Chelators, Respectively) dendrimer generation (number of amino number of maleimides number of functional functions and determined by mass chelators determined chelators)/molecular spectrometry/ by mass spectrometry/ weight back-titration isotopic dilution titration G0 (1)/998 g/mol G1 (2)/1883 g/mol G2 (4)/3653 g/mol G3 (8)/7187 g/mol G4 (16)/14261 g/mol G5 (32)/28405 g/mol G6 (64)/55733 g/mol G7 (128)/112309 g/mol

1/– 2/– 4/– 8/– 16/– –/33.1 ( 3.9 –/62.9 ( 6.8 –/120.2 ( 10.5

1/– 2/– 4/– 8/9.9 ( 2.8 –/17.2 ( 0.4 –/32.2 ( 0.3 –/69.1 ( 6.7 –/155.1 ( 33.7

structures and highly demanding purification steps. However, with the modifications of the synthesis protocol described, the dendrimers were obtained in good yields and purity (Figure 1C,D).

Table 2. Parameters of the Antibody-Sulfo-SMCC and -DOTA-Dendrimer Reactions amount of Sulfo-SMCC applied relative to antibody

number of introduced maleimides per antibody

excess of DOTAX-dendrimers applied relative to antibody

7 equiv 12 equiv 22 equiv 45 equiv

1.7 2.8 5.3 10.0

10 equiv 12.5 equiv 15 equiv 25 equiv

A penta(ethylene glycol) linker containing a thiol function was used as the starting structure element. This linker was introduced to enhance the accessibility of the reactive site of the dendritic system, thereby allowing the effective and selective introduction of the dendritic molecule into proteins or peptides. The linker system was synthesized starting from tetra(ethylene glycol)-bistosylate (19), S-trityl-mercaptoethanolamine (20) and 2-(2-amino-ethoxy)-ethanol as shown in Figure 2.

Figure 6. Schematic depiction of the conjugation of the DOTA-dendrimers to the antibody.

818 Bioconjugate Chem., Vol. 19, No. 4, 2008

Wängler et al.

Table 3. Results of the Derivatization Reactions and Determination of the Immunoreactivities of the Antibody-Dendrimer Conjugatesa immunoreactivity retained (number of dendrimers per antibody) antibody conjugate

1.7 maleimides per antibody

2.8 maleimides per antibody

5.3 maleimides per antibody

10.0 maleimides per antibody

hMAb425-DOTA1 hMAb425-DOTA2 hMAb425-DOTA4 hMAb425-DOTA8 hMAb425-DOTA16 hMAb425-DOTA32 hMAb425-DOTA64 hMAb425-DOTA128

60.3% (0.4) 68.4% (0.5) 66.2% (0.5) 57.8% (1.0) 63.2% (0.8) 60.6% (0.8) 58.2% (1.7) 63.9% (1.0)

66.3% (0.5) 64.3% (0.6) 64.7% (0.6) 55.3% (1.2) 49.0% (1.5) 40.8% (1.1) 46.6% (2.6) 56.1% (1.9)

35.7% (1.0) 31.3% (0.9) 28.5% (1.1) 26.7% (2.7) 21.6% (2.5) 27.9% (1.5) 23.3% (3.6) 28.5% (2.5)

30.0% (1.9) 24.4% (1.9) 17.6% (2.1) 13.0% (6.9) 12.6% (5.8) 14.2% (2.7) 8.4% (7.3) 16.4% (4.4)

a

Immunoreactivities are given relative to the immunoreactivity of the underivatized antibody.

It represents the starting generation (G0) of the dendrimer, as it shows no branching. Subsequently, the PAMAM dendrimers were grown on the terminal amino function of the linker by alternating reaction with methyl acrylate and ethylenediamine to give half (GX.5) and full generations (GX+1) of the dendrimer (Figure 3). Using this method, we were able to synthesize PAMAM dendrimers of 1 to 128 amino functions on a penta(ethylene glycol) linker providing a thiol for selective coupling to a carrier protein. The dendrimers obtained were found to be of high homogeneity and are therefore suitable as building blocks for the synthesis of DOTA-multimers. Formation of DOTA-Multimers. For the subsequent derivatization of the PAMAM dendrimer amino functions with DOTA, the following differently protected and activated DOTAderivatives were studied with regard to their coupling yields and the homogeneity of the obtained products: tris-tBu-DOTA (21), DOTA-PFP-ester and DOTA–PNP-ester (22), tris-allylDOTA (23), tris-benzyl-DOTA (24), tris-4-nitrobenzyl-DOTA and thiol-DOTA (25) (Figure 4). Tris-tBu-DOTA and tris-allyl-DOTA showed high coupling yields and very homogeneous coupling products. However, in both cases the deprotection resulted in mixtures that were difficult to purify. DOTA-PFP-ester, DOTA–PNP-ester, trisbenzyl-DOTA, and tris-4-nitrobenzyl-DOTA were found to react incompletely, thus producing inhomogeneous products even when applied in an excess of up to 5 equiv per amine function. This behavior was attributed to the relatively low reactivity of the active esters and the steric hindrance of tris-benzyl-DOTA and tris-4-nitrobenzyl-DOTA. In contrast to the amide linked conjugates, the thiol-DOTA could be introduced quantitatively after preceding derivatization of the amino groups of the dendrimers with maleimido caproic acid (Figure 5). Using these maleimide-bearing dendrimers, the DOTAderivatized compounds were obtained in good yields and high purities. Furthermore, the coupling reaction was completed within a few minutes and did not require an additional deprotection step for the DOTA moieties. The number of maleimides per dendritic structure was determined by reacting an aliquot of the maleimide-containing dendrimer with a 5-fold excess of MESNA (sodium 2-mercaptoethanesulfonate) for 5 min. The excess of thiol was subsequently quantified using the Ellman′s assay. The amount of reacted thiol corresponds to the amount of maleimides. Using this method, it could be shown that the reaction of the dendritic amines with maleimido caproic acid proceeded quantitatively (Table 1). After the derivatization of the dendritic structures with thiolDOTA, the number of functional chelators per dendrimer was determined by isotopic dilution titration (22) using a 10-10 mol/L solution of the respective dendrimer (relating to DOTA) and trace amounts of 111In in the presence of a range of concentrations of natIn (10-5, 10-6, 7.5 × 10-7, 5 × 10-7, 2.5

× 10-7, 10-7, 7.5 × 10-8, 5 × 10-8, 2.5 × 10-8, 10-8, and 10-9 mol/mL) in sodium acetate buffer (0.4 M, pH 5). The complexation was stopped after 20 min at 60 °C by the addition of an excess of EDTA, and the amount of dendrimer-bound 111 In was determined by TLC. The results of the isotopic dilution titrations are shown in Table 1. Apparently, the number of functional chelators per dendrimer molecule provided a slight overestimation, which is possibly due to the formation of radiocolloids. However, the quantitative reaction of thiol-DOTA with the maleimide-containing dendrimers could be shown. Using this technique, the amino functions of the PAMAM dendrimers were derivatized quantitatively with DOTA, yielding homogeneous products in good yields and short reaction times. Furthermore, additional deprotection steps were dispensable. Conjugation of DOTA-Multimers to hMAb425. To determine the influence of the DOTA-containing structures on the immunoreactivity of an antibody, the dendrimers were conjugated to hMAb425. This humanized monoclonal antibody is directed against the EGF receptor and was recently investigated in clinical studies (26–28). For the antibody derivatization, the protein was reacted with 7, 12, 22, or 45 equiv of the heterobifunctional cross-linker Sulfo-SMCC (sulfosuccinimidyl-4-(N-maleimidomethyl)-cyclohexane-1-carboxylate) for one hour and isolated using sizeexclusion gel chromatography in high recovery rates of 95–98%. The antibody concentration in the obtained solution was determined using the Bradford Assay and the number of maleimides per antibody molecule was determined by backtitration with MESNA as described above and found to be 1.7, 2.8, 5.3, and 10.0 for an applied Sulfo-SMCC amount of 7, 12, 22, or 45 equiv, respectively. The maleimide-derivatized antibodies were then reacted with the different DOTA-containing dendrimers for 5 min (Figure 6). A summary of the reaction parameters of the antibody derivatization with Sulfo-SMCC and the following dendrimer conjugation is given in Table 2. The number of dendritic structures per antibody was determined by 111In-labeling of the DOTA-dendrimers and subsequent antibody conjugation as described above. The reaction mixtures were analyzed by size-exclusion FPLC and the amount of dendrimer conjugated was identified by determining the distribution of radioactivity between the protein and dendrimer fractions collected (results shown in Table 3). The immunoreactivities of the antibody-dendrimer conjugates were determined by FACS analysis using HT29-cells. The immunoreactivities relative to the underivatized antibody were calculated from the corresponding IC50 values. The results shown in Table 3 reveal that the dendrimer size does not influence the immunoreactivity of the derivatized antibody significantly. In contrast to this, the number of derivatization sites has a crucial effect. These results show that the synthesis of antibody-dendrimer conjugates containing up to 128 chelating agents per binding site is feasible.

Antibody-Dendrimer Conjugates

For comparison, a conjugate containing single DOTA moieties linked to the antibody hMAb425 was produced using S-2(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane-tetraacetic acid (Macrocyclics Dallas, TX). This preparation contained approximately 5 DOTA molecules per antibody. The affinity measurement described in the experimental procedures showed that it retained 46.8% of its immunoreactivity. The comprehensive study of the influence of the substitution pattern on the immunoreactivity revealed that the size of a conjugated dendritic structure does not significantly influence the immunoreactivity of antibodies over a wide molecular weight range, whereas the number of derivatization sites is the major factor that determines the binding affinity of the conjugates. Therefore, an antibody modification with dendritic structures should consist of a derivatization with few but large dendritic systems to obtain antibodies that result in a minor change of immunoreactivity. In this context, an antibody modification with less than 1.7 derivatization sites per antibody should lead to higher immunoreactivities of the conjugates than those shown above, though they would require a separation of the dendrimerconjugated antibodies to attain high specific activities. Additionally, a site-specific conjugation (29) of the dendritic structures to the constant region of the antibody might be used and allow a further improvement of the concept by less altered immunoreactivities of the antibody-dendrimer conjugates. In these conjugates, only a minor fraction of the chelators in can be occupied by the radiometal and complexes formed with Ca(II), and other metal ions will be the predominant species in the circulation. It has to be investigated whether and to what extent large polar dendritic structures alter the pharmacokinetics and the specific accumulation of these antibody conjugates.

CONCLUSION Several maleimide-reactive DOTA-containing PAMAM dendrimers of different sizes were synthesized for the introduction into antibodies and other macromolecules to enhance their specific activity for radioimmunotherapy and clinical diagnostics. For this purpose, the established synthesis protocols for PAMAM dendrimers were improved to obtain products of high homogeneity that are suited for further derivatization. For the subsequent coupling of DOTA moieties, a synthesis strategy was found that allows the quantitative derivatization of the PAMAM amino functions, making a deprotection of the DOTA moieties dispensable. The DOTA multimers were conjugated to the anti-EGFR-antibody hMAb425, the number of dendrimers per antibody molecule was determined, and immunoreactivity studies of the antibody-dendrimer conjugates were performed. These experiments showed that the dendrimer size over a wide range does not influence the immunoreactivity of the derivatized antibodies significantly, whereas the number of derivatization sites has a crucial effect.

LITERATURE CITED (1) Leahy, M. F., Seymour, J. F., Hicks, R. J., and Turner, J. H. (2006) Multicenter phase II clinical study of iodine-131Rituximab radioimmunotherapy in relapsed or refractory indolent non-hodgkin’s lymphoma. J. Clin. Oncol. 24, 4418–4425. (2) Gordon, L. I., Witzig, T., Molina, A., Czuczman, M., Emmanouilides, C., Joyce, R., Vo, K., Theuer, C., Pohlman, B., Bartlett, N., Wiseman, G., Darif, M., and White, C. (2004) Yttrium-90labeled ibritumomab tiuxetan radioimmunotherapy produces high response rates and durable remissions in patients with previously treated B-cell lymphoma. Clin. Lymphoma 5, 98–101. (3) Akabani, G., Reardon, D. A., Coleman, R. E., Wong, T. Z., Metzler, S. D., Bowsher, J. E., Barboriak, D. P., Provenzale, J. M., Greer, K. L., DeLong, D., Friedman, H. S., Friedman, A. H., Zhao, X. G., Pegram, C. N., McLendon, R. E., Bigner,

Bioconjugate Chem., Vol. 19, No. 4, 2008 819 D. D., and Zalutsky, M. R. (2005) Dosimetry and radiographic analysis of 131I-labeled anti-tenascin 81C6 murine monoclonal antibody in newly diagnosed patients with malignant gliomas: A phase II study. J. Nucl. Med. 46, 1042–1051. (4) Chen, Z. N., Mi, L., Xu, J., Song, F., Zhang, Q., Zhang, Z., Xing, J. L., Bian, H. J., Jiang, J. L., Wang, X. H., Shang, P., Qian, A. R., Zhang, S. H., Li, L., Li, Y., Feng, Q., Yu, X. L., Feng, Y., Yang, X. M., Tian, R., Wu, Z. B., Leng, N., Mo, T. S., Kuang, A. R., Tan, T. Z., Li, Y. C., Liang, D. R., Lu, W. S., Miao, J., Xu, G. H., Zhang, Z. H., Nan, K. J., Han, J., Liu, Q. G., Zhang, H. X., and Zhu, P. (2006) Targeting radioimmunotherapy of hepatocellular carcinoma with iodine (131I) metuximab injection: clinical phase I/II trials. Int. J. Radiat. Oncol. Biol. Phys. 65, 435–44. (5) Brechbiel, M. W., Gansow, O. A., Atcher, R. W., Schlom, J., Esteban, J., Simpson, D. E., and Colcher, D. (1986) Synthesis of 1-(p-isothiocyanatobenzyl) derivatives of DTPA and EDTA. Antibody labeling and tumor-imaging studies. Inorg. Chem. 25, 2112–2181. (6) Nikula, T. K., Boccia, M., Curcio, M. J., Sgouros, G., Ma, Y., Finn, R. D., and Scheinberg, D. A. (1995) Impact of the high tyrosine fraction in complementary determining regions: measured and predicted effects of radioiododination on IgG immunoreactivity. Mol. Immunol. 32, 865–872. (7) Thomas, T. P., Patri, A. K., Myc, A., Myaing, M. T., Ye, J. Y., and Norris, T. B., Jr. (2004) In vitro targeting of synthesized antibody-conjugated dendrimer nanoparticles. Biomacromolecules 5, 2269–2274. (8) Kobayashi, H., Sato, N., Saga, T., Nakamoto, Y., Ishimori, T., Toyama, S., Togashi, K., Konishi, J., and Brechbiel, M. W. (2000) Monoclonal antibody-dendrimer conjugates enable radiolabeling of antibody with markedly high specific activity with minimal loss of immunoreactivity. Eur. J. Nucl. Med. 27, 1334– 1339. (9) Wu, G., Barth, R. F., Yang, W., Chatterjee, M., Tjarks, W., Ciesielski, M. J., and Fenstermaker, R. A. (2003) Site-specific conjugation of boron-containing dendrimers to anti-egf receptor monoclonal antibody cetuximab (IMC-C225) and its evaluation as a potential delivery agent for neutron capture therapy. Bioconjugate Chem. 15, 185–194. (10) Wu, C., Brechbiel, M. W., Kozak, R. W., and Gansow, G. A. (1994) Metal-chelate-dendrimer-antibody constructs for use in radioimmunotherapy and imaging. Bioorg. Med. Chem. Lett. 4, 449–454. (11) Wu, G., Barth, R. F., Yang, W., Kawabata, S., Zhang, L., and Green-Church, K. (2006) Targeted delivery of methotrexate to epidermal growth factor receptor-positive brain tumors by means of cetuximab (IMC-C225) dendrimer bioconjugates. Mol. Cancer Ther. 5, 52–59. (12) Majoros, I. J., Myc, A., Thomas, T., and Mehta, C. B., Jr. (2006) PAMAM dendrimer-based multifunctional conjugate for cancer therapy: synthesis, characterization, and functionality. Biomacromolecules 7, 572–579. (13) Padilla De Jesus, O. L., Ihre, H. R., Gagne, L., and Frechet, J. M., Jr. (2002) Polyester dendritic systems for drug delivery applications: in vitro and in vivo evaluation. Bioconjugate Chem. 13, 453–461. (14) Islam, M. T., Shi, X., Balogh, L., and Baker, J. R., Jr. (2005) HPLC separation of different generations of poly(amidoamine) dendrimers modified with various terminal groups. Anal. Chem. 77, 2063–2070. (15) Wang, X., Inapagolla, R., Kannan, S., Lieh-Lai, M., and Kannan, R. M. (2007) Synthesis, characterization, and in vitro activity of dendrimer-streptokinase conjugates. Bioconjugate Chem. 18, 791–799. (16) Fréchet, J. M. J., Tomalia, D. A. (2002) Dendrimers and other dendritic polymers. (Scheirs, J. Ed.) pp 590–602, Chapter 25, Wiley, Chichester. (17) Tomalia, D. A., Baker, H., Dewald, J., Hall, M., Kallos, G., Martin, S., Roeck, J., Ryder, J., and Smith, P. (1986) Dendritic

820 Bioconjugate Chem., Vol. 19, No. 4, 2008 macromolecules: synthesis of starburst dendrimers. Macromolecules 19, 2466–2468. (18) Peterson, J., Ebber, A., Allikmaa, V., and Lopp, M. (2001) Synthesis and cze analysis of PAMAM dendrimers with an ethylenediamine core. Proc. Estonian Acad. Sci. Chem. 50, 156– 166. (19) Maltese, M. (2001) Reductive demercuration in deprotection of trityl thioethers, trityl amines, and trityl ethers. J. Org. Chem. 66, 7615–7625. (20) Chen, Y., and Baker, G. L. (1999) Synthesis and properties of ABA amphiphiles. J. Org. Chem. 64, 6870–6873. (21) Heppeler, A., Froidevaux, S., Mäcke, H. R., Jermann, E., Béhé, M., Powell, P., and Hennig, M. (1999) Radiometal-labelled macrocyclic chelator-derivatised somatostatin analogue with superb tumour-targeting properties and potential for receptormediated internal radiotherapy. Chem.sEur. J. 5, 1974–1981. (22) Mier, W., Hoffend, J., Krämer, S., Schuhmacher, J., Hull, W. E., Eisenhut, M., and Haberkorn, U. (2005) Conjugation of DOTA using isolated phenolic active esters: the labeling and biodistribution of albumin as blood pool marker. Bioconjugate Chem. 16, 237–240. (23) Wängler, B., Beck, C., Wagner-Utermann, U., Schirrmacher, E., Bauer, C., Rösch, F., Schirrmacher, R., and Eisenhut, M. (2006) Application of tris-allyl-DOTA in the preparation of DOTA-peptide conjugates. Tetrahedron Lett. 47, 5985–5988. (24) Anelli, P. L., Lattuada, L., Gabellini, M., and Recanati, P. (2001) DOTA Tris(phenylmethyl) Ester: A new useful synthon for the synthesis of DOTA monoamides containing acid-labile bonds. Bioconjugate Chem. 12, 1081–1084.

Wängler et al. (25) Mattila, K., Siltainsuu, J., Balaspiri, L., Ora, M., and Lönnberg, H. (2005) Derivatization of phosphopeptides with mercapto- and amino-functionalized conjugate groups by phosphate elimination and subsequent Michael addition. Org. Biomol. Chem. 3, 3039– 3044. (26) Seiden, M. V., Burris, H. A., Matulonis, U., Hall, J. B., Armstrong, D. K., Speyer, J., Weber, J. D. A., and Muggia, F. (2007) A phase II trial of EMD 72000 (matuzumab), a humanized anti-EGFR monoclonal antibody, in patients with platinumresistant ovarian and primary peritoneal malignancies. Gynecol. Oncol. 104, 727–731. (27) Kollmannsberger, C., Schittenhelm, M., Honecker, F., Tillner, J., Weber, D., Oechsle, K., Kanz, L., and Bokemeyer, C. (2006) A phase I study of the humanized monoclonal anti-epidermal growth factor receptor (EGFR) antibody EMD 72000 (matuzumab) in combination with paclitaxel in patients with EGFRpositive advanced non-small-cell lung cancer (NSCLC). Ann. Oncol. 17, 1007–1013. (28) Graeven, U., Kremer, B., Südhoff, Th., Killing, B., Rojo, F., ¨ nal, C., and Schmiegel, W. (2006) Phase Weber, D., Tillner, J., U I study of the humanised anti-EGFR monoclonal antibody matuzumab (EMD 72000) combined with gemcitabine in advanced pancreatic cancer. Br. J. Cancer 94, 1293–1299. (29) Mindt, T. L., Jungi, V., Wyss, S., Friedli, A., Pla, G., NovakHofer, I., Grünberg, J., and Schibli, R. (2007) Modification of different igG1 antibodies via glutamine and lysine using bacterial and human tissue transglutaminase. Bioconjugate Chem. ASAP Article; DOI: 10.1021/bc700306n. BC700308Q