David Hodson (1), Hitesh Mistry (1), Leon Aarons (1), Michael Davies (2), Sofia Guzzetti (2), Kayode Ogungbenro (1).
(1) University of Manchester, (2) AstraZeneca - Cambridge, UK.
Objectives: Preclinical studies have shown that that it is possible to synergistically increase the efficacy of radiotherapy (RT) using additional therapeutic interventions such as DNA Damage Response Inhibitors (DDRI) (Riches et al, 2020) or immune checkpoint inhibitors (ICI). This is attributed to inducing a consistent immune response where cytotoxic T lymphocytes successfully clear the tumour. Population based modelling of tumour immune response to RT/ICI combinations have been described (Kosinsky et al, 2018). However, successes are typically limited specifically to certain tumour subsets and specific therapeutic regimens. The long term aims of this project are to build a population based model that captures the impacts of RT in combination with DDRIs and ICIs, and assess this model in different syngeneic tumour models with vastly different baseline levels of immunogenicity.
Methods: This poster describes a mathematical model, which attempts to incorporate the impacts of combination therapies and how they prolong the activity of the adaptive immune response. This model integrates the intrinsic aspects of the immune system on the tumour with drugs that increase the overall immunogenicity of the tumour such as DDRI or drugs which attempt to maintain a population of active CD8 cells such as ICI. The Model was fitted to populations of mice challenged subcutaneously with MC38, a syngeneic tumour model of colorectal cancer, in which, 8 cohorts of 12 mice were given either no therapy, monotherapy, or combination therapies of RT/ICI/DDRI. Mixed effects modelling was performed using NONMEM v 7.4.3 using FOCE-SAEM. Parameters estimated with IIV were the initial tumour diameter, tumour growth rate, and rate of CD8 cell exhaustion. IIV was assumed to be log-normally distributed. Parameters without IIV which were estimated were the rate of APC influx by RT, as well as the impacts of DDRI and ICI on T cell activity. Diagnostic plots, individual fits and visual predictive checks (VPCS) were performed in R version 3.6.3.
Results: Model fits predicted a population T cell exhaustion rate of approximately 0.278 d-1 in both control and RT monotherapy based cohorts. RT in conjunction with DDRIs were able to reduce the rate of T cell exhaustion by 66% while RT in conjunction with ICIs predicted a 92% reduction in T cell exhaustion rate. VPCS confirmed a successful internal model validation that effectively captures how ICIs in conjunction with RT induce a prolonged CD8 T cell response.
Conclusion and implications: This mathematical model successfully captures how PD-L1 upregulation is one of the major sources of immune escape after RT-mediated CD8 activation and how concurrent blockade of PD-L1 is capable of increasing the likelihood of cure in MC38 syngeneic models. With the addition of PK based parameters, as well as the incorporation of CD8 mediated negative feedback, other potential drug dosage regimens may be tested in order to inform how MC38 and other models with similar immunophenotypes to MC38 may respond to various doses of RT in combination with DDR or IO. This may have some potential for clinical translation.
References:
[1] Kosinsky., Y. et al. (2018). Radiation and PD-(L)1 treatment combinations: immune response and dose optimisation via a predictive systems model. Journal for Immunotherapy of Cancer. 6. DOI: 10.1186/s40425-018-0327-9
[2] Riches., L. et al. (2020). Pharmacology of the ATM Inhibitor AZD0156: Potentiation of Irradiation and Olaparib Responses Preclinically. Small Molecule Therapeutics. 19. DOI: 10.1158/1535-7163.MCT-18-1394
Reference: PAGE 29 (2021) Abstr 9631 [www.page-meeting.org/?abstract=9631]
Poster: Drug/Disease Modelling - Oncology