Radiolabeled adenoviral sub-unit proteins for molecular imaging and therapeutic applications in oncology

Full text: Our group has initiated investigations on the use of radiolabeled adenoviral (Ad) sub-unit proteins for delivering suitable radionuclides into tumor cells for molecular imaging as well as for combined gene/radionuclide therapy of cancer. A number of issues involved in developing combined gene/radionuclide delivery into tumors mediated by Ad vectors have been identified and are being addressed. Whereas current clinical trials of gene therapy using Ad vectors involve non-systemic delivery of therapeutic genes, the delivery of radionuclides preferably would involve systemic (i.v.) administration. The distribution and delivery of Ad sub-unit proteins following i.v. administration is not understood and must be studied and optimized. In addition, retention of the selective binding and internalization into tumor cells of the radiolabeled viral vectors remains an unmet challenge. We used the intact adenovirus (Ad, ∼80 nm diameter), native adenoviral fiber protein (AdFP, 180 kD trimer, purified from infected human cultured cells) and the adenoviral fiber 'knob' protein (recombinant AdFKP, 60 kD, synthesized in E. Coli), all of which interact with the in-vivo cellular receptor, coxsackie and adenovirus receptor (CAR) through the knob domain of the adenovirus fiber protein. Our initial studies were aimed at optimizing the labeling conditions using I-131 and In-111 to maintain CAR binding activity of the radiolabeled preparations. The CAR-binding was retained as determined using reaction with biotinylated CAR followed by chemiluminescence detection. The biodistribution results in mice and rats following i.v. administration (autoradiography, tissue counting) showed that all three vectors localized preferentially in CAR-expressing organs (liver, lung, heart, kidney), as expected. The CAR-binding of Ad-2 wild serotype was better (∼8 x stronger) than Ad-12, in particular following radiolabeling. Based on the above results, we further focused on the recombinant knob protein as well as CAR and its variable type domain CAR D1 for more detailed studies. The knob protein can be bio-engineered to interact with molecular targets, e.g., tumor-associated antigens, using in vitro evolution methods. The optimized iodination procedure was as follows. NaI-131 was incubated with one Iodobead for 5 min in TRIS pH 6.8. The reactive iodine species was then separated from the bead, and the knob protein added to it for 1 minute. This method produced I-131-knob without either SDS-PAGE detectable conformational damage or loss of CAR-binding activity. Following intravenous injection in mice, there was uptake in CAR-binding tissues (liver, kidney, heart, and lung) as expected, but blood and whole body clearance were both quite rapid. When mice were pre-injected with excess unlabeled knob (cold competitor), the level of I-131-knob (% injected dose/g) in blood was increased (19.4 vs. 2.5) with a corresponding decrease in metabolic deiodination (6.5 vs. 33.5 in stomach). Radioiodinated knob mutants, which are devoid of CAR-binding activity showed high levels (% injected dose/g) in blood at 6 h post injection (31.0), whereas wild-type I-131-knob in blood at 6 h had decreased to background levels (2.0). Knob mutants, which are unable to interact with CAR, are retained longer in the blood and have slower metabolism. These preliminary results suggest that engineered ligands based on adenoviral sub-unit proteins (with or without the ability to bind to CAR) may prove useful as a new class of ligands for molecular imaging and possibly for the combined gene/radionuclide therapy of cancer. (author)