(662f) Single-Cell Tumor Metabolism of Immune Checkpoint Inhibitors Determines Optimal Dosing for This Class of Antibody Therapeutics

Successful cancer therapeutics must strike a balance between the effective dose and the drug tolerability, termed ‘the therapeutic window’. Typically, the clinical dose is increased until the dose limiting toxicity (DLT) is reached, and then reduced to the maximum tolerated dose (MTD). However, many next-generation therapeutics, such as immune checkpoint inhibitors, frequently lack a relevant MTD, making clinical dose estimation challenging. Overdosing patients results in higher drug costs, the potential for immune reactions, and restricts routes of administration with no benefit to the patient. In the absence of an upper bound due to toxicity, the focus is placed on efficacy. Similar to many non-chemotherapy drugs, the maximum efficacy occurs when the target is saturated. However, the relationship between systemic plasma concentration and receptor occupancy at the site of action within a tumor is complex. This is particularly acute with the dynamic expression of immune checkpoint proteins such as PD-1 and PD-L1 that have widely varying expression levels on multiple cell types that can change dramatically during treatment. In this work, we used near-infrared fluorescence ratio imaging to quantify tumor uptake and metabolism of antibodies targeting PD-1 and PD-L1 with single-cell resolution. The results highlight tumor metabolism as one of the defining parameters for clinical dosing of these drugs. Strikingly, this is not a routine measurement during drug development. Likewise, the results show that systemic clearance and plasma markers can be misleading when determining target saturation within the tumor microenvironment.

There are three key factors that impact receptor occupancy and determine the dose needed for target saturation: plasma clearance, binding affinity, and tumor metabolism. Plasma clearance and binding affinity are well characterized for clinical therapeutics with binding affinity being measured early in the preclinical phase and plasma clearance scaled from multiple animal models. However, tumor metabolism is not routinely investigated. The affinity for many antibodies is in the single digit to sub-nanomolar range, while plasma clearance half-lives are generally greater than 2 weeks. With relatively high affinity and long half-lives, our simulations show that saturation during treatment is commonly limited by tumor metabolism. Therefore, measuring tumor metabolism rates of these biologics is a critical parameter in determining the clinical therapeutic dose.

In this work, we studied the dosing of Programmed cell Death 1 (PD-1) and Programmed Death-Ligand 1 (PD-L1) antibodies. Because of varying target expression on multiple cell types, quantifying PD-1 and PD-L1 expression and internalization rates in vitro may not reflect the metabolism in vivo. To quantify the rates in vivo using a syngeneic mouse model, we used two low toxicity antibodies (anti-PD-1 and anti-PD-L1) in Balb/c mice with CT-26 xenografts. Anti-PD-1 and anti-PD-L1 antibodies were dually labelled with a residualizing near-infrared (NIR) fluorescent dye (that is trapped within the cell following metabolism) and a non-residualizing NIR dye that washes away. By harvesting the treated tumors, generating a single-cell digest, and analysing the relative ratio of the NIR dyes by flow cytometry, we were able to measure single-cell binding and metabolism of the antibodies on different cell populations within the tumor (e.g. tumor cells, T-cells, and macrophages). Once these in vivo metabolism rates were measured, the saturating dose can be approximated by a Thiele modulus. Originally developed to describe the relationship between diffusion and reaction rate in porous catalyst pellets, this characteristic value can describe the relative rate of diffusion for the therapeutic to tumor metabolism based on parameters specific to the tumor environment (e.g. receptor expression and internalization) and the therapeutic (e.g. diffusion/target affinity).

Using our experimental values and tumor parameters from literature, the adapted Thiele modulus predicted a tumor saturating dose to be 0.03mg/kg and 4mg/kg for anti-PD-1 and anti-PD-L1, respectively. These values are similar to the clinical dose for anti-PD-L1 therapies but well below those used for anti-PD-1 therapy. This may explain a lack of dose response in anti-PD-1 clinical trials. Interestingly, the experimental data and simulations indicate that much higher doses of anti-PD-1 are needed to saturate healthy tissue than the tumor, highlighting a potential disconnect between tumor saturation and healthy tissue saturation (as indicated by dose-dependent clearance).

In summary, although many cancer therapies are administered at the MTD, emerging therapeutics, such as immune checkpoint inhibitors, have low toxicity, resulting in unclear dosing strategies for the clinic. According to our in vivo data and computational modeling, tumor metabolism is often the rate limiting target occupancy, and current low toxicity antibody therapies like anti-PD-1/anti-PD-L1 should be administered based on tumor metabolism to achieve efficacy without over-treating patients. Surpassing the saturating dose no longer benefits the patient and may increase the risk of adverse events such as immunogenic responses to the protein (e.g. human anti-human antibodies). Setting the optimal dose based on tumor metabolism will serve as a way to maintain efficacy and reduce over-treatment for this drug class.