Targeting of radiolabeled J591 antibody to PSMA-expressing tumors: optimization of imaging and therapy based on non-linear compartmental modeling.
Academic Article
Overview
abstract
BACKGROUND: We applied a non-linear immunokinetic model to quantitatively compare absolute antibody uptake and turnover in subcutaneous LNCaP human prostate cancer (PCa) xenografts of two radiolabeled forms of the humanized anti-prostate-specific membrane antigen (PSMA) monoclonal antibody J591 ((124)I-J591 and (89)Zr-J591). Using the model, we examined the impact of dose on the tumor and plasma positron emission tomography (PET)-derived time-activity curves. We also sought to predict the optimal targeting index (ratio of integrated-tumor-to-integrated-plasma activity concentrations) for radioimmunotherapy. METHODS: The equilibrium rates of antibody internalization and turnover in the tumors were derived from PET images up to 96 h post-injection using compartmental modeling with a non-linear transfer rate. In addition, we serially imaged groups of LNCaP tumor-bearing mice injected with (89)Zr-J591 antibody doses ranging from antigen subsaturating to saturating to examine the suitability of using a non-linear approach and derived the time-integrated concentration (in μM∙hours) of administered tracer in tumor as a function of the administered dose of antibody. RESULTS: The comparison of (124)I-J591 and (89)Zr-J591 yielded similar model-derived values of the total antigen concentration and internalization rate. The association equilibrium constant (k a) was twofold higher for (124)I, but there was a ~tenfold greater tumoral efflux rate of (124)I from tumor compared to that of (89)Zr. Plots of surface-bound and internalized radiotracers indicate similar behavior up to 24 h p.i. for both (124)I-J591 and (89)Zr-J591, with the effect of differential clearance rates becoming apparent after about 35 h p.i. Estimates of J591/PSMA complex turnover were 3.9-90.5 × 10(12) (for doses from 60 to 240 μg) molecules per hour per gram of tumor (20 % of receptors internalized per hour). CONCLUSIONS: Using quantitative compartmental model methods, surface binding and internalization rates were shown to be similar for both (124)I-J591 and (89)Zr-J591 forms, as expected. The large difference in clearance rates of the radioactivity from the tumor is likely due to differential trapping of residualizing zirconium versus non-residualizing iodine. Our non-linear model was found to be superior to a conventional linear model. This finding and the calculated activity persistence time in tumor have important implications for radioimmunotherapy and other antibody-based therapies in patients.