Alberto Ippolito
AstraZeneca
Objectives:
Conditionally activated molecules are designed to deliver a targeted therapy or reduce the off-tumor activity by exploiting specific properties of the tumor micro-environment (TME), such as low pH or overexpression of specific proteases. These molecules are becoming increasingly present in clinical trials and are soon to reach the market. While most of the development focuses on the modality of activation of the molecules, the interplay between these modalities and the PK properties of the molecule that ultimately lead to better efficacy with higher specificity remains elusive. Here, we perform an exploratory study by introducing a generic PKPD framework for the activity of conditional molecules and we utilize published data from existing conditional formats for immune checkpoint inhibition (ICI) to analyse the interplay between the molecule half-life and the modality of action. For a given molecule potency, our results show that there is a sweet spot within these properties which allows for the best trade-off between molecule efficacy and locality of action.
.
Methods:
Here we modify a simplified human PKPD model with IV-injection from [1] composed of 3 compartments: the central or blood, the tumor and the peripheral compartment. For the purpose of this work, we have assumed that the molecules are masked biologics, in particular Fc and non-Fc bearing antibodies, but the same methodologies and conclusions can be extended to other formats of therapeutic molecules. We have assumed that all the PK/PD properties are the same as [1], except for the molecule half-life, which we vary to simulate a range of molecule PK (from order of hours as BiTe-like molecules to weeks as usual IgG1 formats [2]).
While we do not specify the exact mechanism of activation, we consider that the affine molecule sites are activated by a rate which in tumor is multiplied by a factor α=10, representing the ratio between the tumoral and non-tumoral activation rates, as done in [1]. The base affinity, receptor expression and turnover rate remain unaffected by the properties of the molecules.
We use as a proxy of the efficacy the occupancy or engagement of the tumor ligands by the conditional molecule while the specificity of action is the ratio between the percent occupancy of the ligands in the tumor over the percent in the non-target tissue. We evaluate the occupancies as the average receptor occupancy over the therapeutic interval. Results are averages over 210 days of distinct simulated therapies with dosing schedules of Q1W, Q2W and Q3W with fixed doses of 1, 10 and 30 mg/kg.
Results:
For a selected dose schedule and amount, we study the change in occupancy and specificity as a function of the molecule half-life and dynamics of activation. For each dosing schedule, we analyse the respective average occupancy and specificity for each fixed dose. The simulated data clearly shows that fast activation and a long half-life are ideal for maximizing target occupancy while a short half-life with a slow activation are best for improved specificity. In fact, a fast activation rate essentially entails that all the molecules are in their active state and therefore can occupy the receptors indiscriminately. Thus, they maximize the occupancy on both tumor and non-target tissue, leading to high efficacy but low specificity of action. On the other hand, since we have assumed that the activation rate in the tumor is higher than off-tumor, short half-life molecules tend to activate mostly in the tumor and are cleared before they can travel to the rest of the body, leading to high specificity. However, since this differential in activation requires a relatively low activation rate, most molecules remain inactive and thus the efficacy is low. This trend is maintained independent of the dose amount and the same result can be extended for different dosing schedules as well.
Conclusions:
Here, we have shown the interplay between the mechanism of activation and the PK of conditionally activated molecules using a simple 3-compartment PKPD model. The simulated results suggest that maximizing molecule efficacy requires slow clearance with fast activation while the polar opposite molecule properties are required to maximize specificity of action. However, efficacy is obtained not only by the receptor occupancy but also by the potency of the molecule. Therefore, this simple framework suggests that very potent molecules should be designed as short half-life with a moderate to slow activation rate. On the other hand, less potent molecules that require higher occupancies should ideally be designed with a moderate to long half-life and a fast activation mechanism.
References:
- Stroh M, Sagert J, Burke JM, Apgar JF, Lin L, Millard BL, Kavanaugh WM. Quantitative systems pharmacology model of a masked, tumor‐activated antibody. CPT: Pharmacometrics & Systems Pharmacology. 2019 Sep;8(9):676-84.
- Baeuerle PA, Reinhardt C. Bispecific T-cell engaging antibodies for cancer therapy. Cancer research. 2009 Jun 15;69(12):4941-4.
Reference: PAGE 32 (2024) Abstr 10765 [www.page-meeting.org/?abstract=10765]
Poster: Drug/Disease Modelling - Oncology