Abstract
1000
Objectives: Immune checkpoint blockade (ICB) therapies are often effective for cancer treatment. Atezolizumab (Atz), a fully humanized anti-PD-L1 monoclonal antibody (mAb), is approved for ICB therapy for lung and bladder cancer. Despite dramatic success with Atz in some patients, we cannot yet reliably predict who will respond positively to Atz treatment, but favorable response appears related to PD-L1 levels in the tumor microenvironment. Immunohistochemical assessment of tumor PD-L1 levels is invasive and requires biopsy, which is often impractical, particularly in metastatic disease or for assessing disease over time. With the objective of non-invasive tumor PD-L1 assessment, we generated Atz labeled with the PET isotope, 89Zr (89Zr-Atz), and evaluated its ability to target xenograft tumors in mice that express human PD-L1 (hPD-L1). METHODS: Atz was conjugated with deferoxamine (DFO) at 10:1, 5:1, and 2:1 DFO:mAb ratios and 89Zr chelated. Radiochemical yield, purity, stability and aggregation were assayed by instant thin-layer chromatography (iTLC) and FPLC. 89Zr-Atz immunoreactivity was quantified by cell binding and Kds of Atz and Atz-DFO determined by enzyme-linked immunosorbent assays (ELISA) and of 89Zr-Atz by cell binding. Binding to mouse PD-L1 (mPD-L1) was assayed by flow cytometry and ELISA. SCID mice (n≥4) with bilateral shoulder xenografts of Cho and Cho-PD-L1 cells were injected with 3.7 MBq (~11 µg) of 89Zr-Atz without or with cold Atz (10, 50, 250 or 1250 µg) or 250 µg cold IgG. 89Zr-IgG alone was also injected. Biodistribution was determined 7 days post-injection (p.i.) and standardized uptake values (SUVmax and SUVave) determined from PET/CT images acquired at days 1, 3 and 7 p.i. RESULTS: We obtained 89Zr-Atz radiochemical yields of >260 MBq/mg, with >95% purity and >90% immunoreactivity. Atz, Atz-DFO and 89Zr-Atz exhibited similar hPD-L1 Kds of <2 nM. >95% 89Zr remained bound after 7 days in serum. Aggregation of Atz-DFO and 89Zr-Atz were less than 10% and Atz bound mPD-L1. 89Zr-IgG did not specifically target PD-L1 expressing tumors, as assayed by biodistribution (Table 1) or SUV analysis. Similarly, specific targeting was not observed with 89Zr-Atz alone, with co-injection of cold IgG, or with co-injection of 10, 50, or 1250 µg cold IgG. Blood levels of 89Zr-Atz were low in the absence of co-injected cold Atz, indicating saturable binding to mPD-L1. In contrast, co-injection of 250 µg of cold Atz resulted in specific 89Zr-Atz targeting of PD-L1 expressing tumor cells by biodistribution (Table 1, [asterisk]p<0.05) and SUV analyses (e.g., day 3 SUVave Cho 1.2±0.3 vs. Cho-PD-L1 2.6±0.3, p<0.05). CONCLUSIONS: 89Zr-Atz-DFO was generated with high immunoreactivity and specific activity. Atz bound to both mouse and human PD-L1, and 89Zr-Atz recognized mouse tissues expressing PD-L1. The ability of 89Zr-Atz to specifically target tumor cells expressing PD-L1 was highly dependent on Atz protein mass. These results suggest that the amount of co-injected cold Atz with 89Zr-Atz will be critical for maximal detection of PD-L1 in the tumor microenvironment in future clinical trials. We thank Sridhar Nimmagadda for the Cho-PD-L1 cell line.
Biodistribution (%ID/g)