QSP Model of a Cytokine mRNA Immunotherapy

Background

This case study features Applied BioMath's collaboration with Sanofi. 
 
Despite the success of checkpoint blockade immunotherapy, most patients do not respond adequately to treatment and tumors effectively evade T lymphocyte and Natural Killer (NK) cell immune surveillance. Within the tumor microenvironment, the activities of immune cells are regulated by a multitude of signals, including the local cytokine milieu. 
 
Intratumoral administration of our cytokine mRNA cocktail therapy, consisting of transcripts encoding for IL-12sc, IFNα, IL-15sushi and GM-CSF, has induced regression in tumor-bearing mice. Leveraging in vitro, in vivo and in silico data, we have developed a Quantitative Systems Pharmacology (QSP) platform that captures the observed preclinical tumor responses and allows for systemic predictions under a variety of conditions.

Results

  1. The developed preclinical mouse melanoma QSP model captures a variety of tumor responses under different treatment conditions. The multiscale platform is readily translatable across species with the potential for sub-module utilization in the construction  of other Immuno-Oncology models.

  2. Human model predictions could be used to inform the implementation of effective therapeutic strategies in the future.

Model Diagram

Model Diagram

Schematic representation of the QSP model structure: Following intra-tumoral (i.t.) injection of the cytokine mRNA cocktail therapy, the mixture is endocytosed by the tumor cells, translated and subsequently secreted into the tumor interstitium.

The secreted cytokines promote tumor killing and trigger the activation and proliferation of CD4+/CD8+ T lymphocytes and NK cells.

Results

Dose-dependent effects of cytokine mRNA on tumor growth: Tumor bearing mice were subjected to intratumoral control or various doses of cytokine mRNA treatment. Simulated responses are aligned with experimental data with more pronounced regression observed with increasing doses. Colored lines of panels A and B show individual and median experimental responses  +/- 1 standard deviation respectively. Solid black lines indicate QSP model output.

 

Results

Download the Case Study