Pharmacokinetic models to evaluate radiolabeling reagents for protein radiopharmaceuticals

Poor outcome with anti-programmed death-ligand 1 (PD-L1) antibody due to poor pharmacokinetic properties in PD-1/PD-L1 blockade-sensitive mouse models

Background

Recently, antiprogrammed cell death protein 1 (aPD-1) and antiprogrammed death-ligand 1 (aPD-L1) monoclonal antibodies (mAbs) have been approved. Even though aPD-1 and aPD-L1 mAbs target the same PD-1/PD-L1 axis, it is still unclear whether both mAbs exert equivalent pharmacological activity in patients who are sensitive to PD-1/PD-L1 blockade therapy, as there is no direct comparison of their pharmacokinetics (PK) and antitumor effects. Therefore, we evaluated the differences between both mAbs in PK and therapeutic effects in PD-1/PD-L1 blockade-sensitive mouse models.

Methods

Herein, murine breast MM48 and colon MC38 xenografts were used to analyze the pharmacological activity of aPD-1 and aPD-L1 mAbs. The PK of the mAbs in the tumor-bearing mice was investigated at low and high doses using two radioisotopes (Indium-111 and Iodine-125) to evaluate the accumulation and degradation of the mAbs.

Results

aPD-1 mAb showed antitumor effect in a dose-dependent manner, indicating that the tumor model was sensitive to PD-1/PD-L1 blockade therapy, whereas aPD-L1 mAb failed to suppress tumor growth. The PK study showed that aPD-L1 mAb was accumulated largely in normal organs such as the spleen, liver, and kidney, resulting in low blood concentration and low distributions to tumors at a low dose, even though the tumors expressed PD-L1. Sufficient accumulation of aPD-L1 mAb in tumors was achieved by administration at a high dose owing to the saturation of target-mediated binding in healthy organs. However, degradation of aPD-L1 mAb in tumors was greater than that of aPD-1 mAb, which resulted in poor outcome presumably due to less inhibition of PD-L1 by aPD-L1 mAb than that of PD-1 by aPD-1 mAb.

Conclusion

According to the PK studies, aPD-1 mAb showed linear PK, whereas aPD-L1 mAb showed non-linear PK between low and high doses. Collectively, the poor PK characteristics of aPD-L1 mAb caused lower antitumor activity than of aPD-1 mAb. These results clearly indicated that aPD-L1 mAb required higher doses than aPD-1 mAb in clinical setting. Thus, targeting of PD-1 would be more advantageous than PD-L1 in terms of PK.

Species difference in radioactivity elimination from liver parenchymal cells after injection of radiolabeled proteins

To elucidate the cause for the different levels of hepatic radioactivity among mammals after injection of protein radiopharmaceuticals, the metabolism of radiolabeled proteins and the fate of their radiometabolites in the parenchymal cells of rat liver were investigated and compared with those of mice. We used galactosyl-neoglycoalbumin (NGA) as a carrier protein, and NGA was labeled with 111In via 1-(4-isothiocyanatobenzyl)ethylenediaminetetraacetic acid (SCN-Bz-EDTA) or 1-[p-(5-maleimidopentyl)aminobenzyl]ethylenediaminetetraacetic acid (EMCS-Bz-EDTA) and with 125I via direct iodination. All radiolabeled NGAs exhibited rapid accumulation in liver parenchymal cells after intravenous injection into rats. Radioactivity was eliminated following NGA-125I injection at similar rates from rat and mouse liver. In contrast, both 111In-labeled NGAs demonstrated much slower elimination of radioactivity in rat when compared with mouse liver. Analyses of radioactivity in bile and liver indicated that both SCN-Bz-EDTA and EMCS-Bz-EDTA rendered mono-amino acid adducts as the final radiometabolites, which were generated in rat liver within 1 h postinjection. Subcellular distribution studies suggested that these radiometabolites were copurified with lysosome in rat liver. Because similar results were observed in mice previously, the difference between rats and mice in radioactivity elimination from liver parenchymal cells would be predominantly attributable to the different efflux rate of the 111In-labeled metabolites from the lysosome between these species. Such differences in the efflux rates of radiometabolites from the lysosome among mammals may also account for the different hepatic radioactivity levels of radiolabeled proteins between animal and clinical studies.

In-vivo evaluation of indium-111-diethylenetriaminepentaacetic acid-labelling for determining the sites and rates of protein catabolism in mice

Pharmacokinetic analyses of protein pharmaceuticals are of prime importance for their clinical application. Because many proteins have pharmacological activity at low concentrations, radiolabelling of proteins is widely used to identify the sites and determine the rates of protein catabolism in-vivo due to the high sensitivity of detection of radioactivity. Recently, a metallic radionuclide, (111)In, has been used to trace the pharmacokinetics of proteins of interest after conjugation of the proteins with diethylenetriaminepentaacetic acid (DTPA). In this study, galactosyl-neoglycoalbumin (NGA) was reacted with the cyclic dianhydride of DTPA and labelled with (111)In to estimate the validity of this radiolabelling procedure for pharmacokinetic analyses. For comparison, we also evaluated direct radioiodination, because directly-radioiodinated proteins are widely used to assess the pharmacokinetics of proteins of interest. The hepatic radioactivity profile after intravenous injection of [131I]NGA or [(111)In]DTPA-NGA into mice was analysed pharmacokinetically, and the first-order rate constant representing the elimination of the respective radiometabolite from hepatic parenchymal cells was determined. The results indicated that direct radioiodination is inappropriate for pursuing the pharmacokinetics of the proteins, because of rapid elimination of the radioactivity from the sites of protein catabolism. These findings also implied that the [(111)In]DTPA label could be used to identify the catabolic sites and determine the rates of catabolism of proteins with relatively short biological half-lives, although characterization of radiolabelled species at the sites of accumulation would be required for accurate determination of the catabolic sites of proteins.