Abstract
The two most common antibody targeting principles in oncology are the induction of direct antitumor effects and the release of antitumor T cell immunity by immune checkpoint blockade. These two principles, however, may be overlapping if the targeted checkpoint molecule is not located on the immune cell but on the tumor cell itself. Secondary resistance by epitope escape may therefore remain a challenge in both settings. We previously reported epitope escape through L88S and truncating programmed cell death ligand 1 (PD-L1) gene mutations in colorectal cancer patients on selective pressure with avelumab, a PD-L1-directed checkpoint blocker that-in addition to T cell disinhibition-allows direct tumor cell killing via its unmodified Fc portion. Here, we confirmed this principle by liquid biopsy monitoring in a colorectal cancer patient from an independent clinical trial. In this patient, both PD-L1 L88E and L88fs mutations emerged under selective pressure with avelumab. By ectopically expressing PD-L1 L88E, we show that this mutation leads to a reduction of full-length glycosylated PD-L1 and greatly reduced avelumab surface binding. Further experiments indicated that PD-L1 L88E represents a phosphomimetic variant of PD-L1 L88S leading to loss of protein stability and increased proteasomal degradation. The association of this PD-L1 mutation with the high-affinity FCGR3A single nucleotide polymorphism rs396991 confirms prior evidence that patients harboring this polymorphism experience the strongest selective pressure by avelumab. Together, position 88 of PD-L1 is a hotspot residue critically regulating PD-L1 cell surface expression with clinical significance in the context of immune checkpoint blockade.