168P - Characterization of the lung tumor microenvironment upon anti-PD-L1 therapy reveals an ambiguous role for TNF-α

Background

Immune checkpoint blockade (ICB) of the PD-1 pathway revolutionized the survival forecast for advanced non-small cell lung cancer (NSCLC). Yet, the majority of PD-L1⁺ NSCLC patients are refractory to anti-PD-L1 therapy. Recent observations indicate a pivotal role for PD-L1⁺ tumor-infiltrating myeloid cells in therapy failure.

Methods

We evaluated the abundance, phenotype and function of 11 different myeloid subsets within an orthotopic squamous Lewis lung carcinoma (LLC) model over the course of anti-PD-L1 monoclonal antibody (mAb) treatment.

Results

We confirm that LLC represents an anti-PD-L1 therapy refractory model, despite increased levels of PD-1⁺ lymphocytes and PD-L1⁺ monocytes, macrophages, and type 2 dendritic cells. Furthermore, we show that anti-PD-L1 therapy significantly increased serological level of TNF-α, while it reduced the fraction of tumor-infiltrating MHC-II^low macrophages and monocytes. Notably, the latter were transcriptionally characterized by an increased responsiveness to TNF-α, suggesting a direct link between anti-PD-L1 therapy and TNF-α. However, co-blockade of PD-L1 and TNF-α did not reduce LLC tumor growth as quantified by 3D whole-lung imaging. Mechanistically, we show that the net impact of TNF-α on anti-PD-L1 therapy is immune cell specific: while TNF-α increases IFN-γ secretion by lymphocytes, their killing capacity is significantly reduced when co-cultured with TNF-α-treated monocytes. In line, we were able to show that TNF-α alone or combined with anti-PD-L1 mAb, resulted in elevated expression of the following alternative immune checkpoints: LAG-3 or VISTA, TIM-3 and SIRPα respectively. Currently we are deciphering if combined targeting of one or more of the latter checkpoints can revert the immunosuppressive impact of monocytes on tumor-specific T cells under anti-PD-L1 treatment.

Conclusions

To conclude, this study shows that anti-PD-L1 treatment of lung-tumor bearing mice results in a TNF-α-related increment of immunosuppressive monocytes. While this study further argues against TNF-α and PD-L1 co-blockade to improve therapy effectiveness, it warrants more research into strategies that block monocyte-mediated resistance to anti-PD-L1 therapy.

Legal entity responsible for the study

Prof. Cleo Goyvaerts.

Funding

This research was performed with the financial support of the Research Foundation-Flanders, Wetenschappelijk Fonds Willy Gepts of the UZ Brussel, Kom op tegen Kanker (the Flemish Cancer Society), Universidad Pública de Navarra, Gobierno de Navarra, and The Research Council of the Vrije Universiteit Brussel.

Disclosure

Q. Lecocq: Non-Financial Interests, Institutional, Ownership Interest, granted patent on “Human PD-L1 binding immunoglobulins”: Vrije Universiteit Brussel. M. Keyaerts: Non-Financial Interests, Institutional, Ownership Interest: AbScint; Non-Financial Interests, Institutional, Funding: Precirix; Non-Financial Interests, Institutional, Ownership Interest, granted patent on “Human PD-L1 binding immunoglobulins”: Vrije Universiteit Brussel. K. Breckpot: Non-Financial Interests, Institutional, Ownership Interest, granted patent on “Human PD-L1 binding immunoglobulins”: Vrije Universiteit Brussel. All other authors have declared no conflicts of interest.