Targeting glycosylation of PD-1 to enhance CAR-T cell cytotoxicity

To the Editor,

Chimeric antigen receptor T (CAR-T) cells are not satisfying in treating solid tumors [1]. PD-1 limits CAR-T cell therapy within solid tumors. CRISPR/cas9 can downregulate PD-1 [2] but also potentially leads to carcinogenesis, because success of such tools relies on suppressing DNA damage response [3]. Furthermore, CRISPR/cas9 could bring in missense mutations that might exacerbate T cell dysfunction. Hence, we need safer and more precise gene-editing tools to produce better CAR-T cells. N-linked glycosylation can stabilize PD-1 to compromise anti-tumor immunity [4]. As N-linked glycosylation is restricted on asparagine coded by aac or aat, adenine base editors (ABE) can convert a·t to g·c base pair [5], and may be used to diminish such glycosylation. Herein, we explored the potentials of ABE to edit and downregulate PD-1 in CAR-T cells.

Mutated PD-1 at N74 had decreased surface expression (Fig. 1a). Therefore, N74 in PD-1 is a good target for ABE. Three types of amino acids may be produced after base editing at N74 coded by aac (Fig. 1b). All 3 types of mutations into D74 (gac), S74 (agc), and G74 (ggc) comparably downregulated the surface and total PD-1 (P < 0.001) (Fig. 1c and Additional file 1: Figure S1c). Next, we investigated whether ABE was able to decrease PD-1 in CAR-T cells. The delivery of gRNA using lentivirus is efficient [6], so we constructed lentiviral vectors simultaneously expressing mesothelin-directed CAR and gRNA targeting non-specific sites (scramble) or N74 of PDCD1 (gRNA), under 2 independent promoters (Additional file 1: Figure S1a). T cell transduction efficacies were over 85% (Additional file 1: Figure S1b). Then the commercially synthesized ABE proteins were delivered into CAR-T cells by electroporation. Sequencing data showed the conversion to g majorly happened from the first adenine within N74 codon of PDCD1 in CAR-T cells expressing specific gRNA (Fig. 1d). Conversion was also noticed at the second adenine with lower frequencies (Fig. 1d). This editing pattern is consistent with previous report [7].

In following experiments, the ratios of CAR-expressing cells were comparably adjusted to 85%. In gRNA CAR-T cells, PD-1 expression was decreased at protein level but not at mRNA level (Fig. 1e and f). Consistently, surface PD-1 was remarkably decreased in resting and activated gRNA CAR-T cells (P < 0.01) (Fig. 1g). Further analysis suggested that ABE editing did not impair the proliferation and activation of CAR-T cells (P > 0.05) (Fig. 1h–j) when PD-L1 was absent. Then mesothelin-positive cells with high PD-L1 expression were prepared (Fig. 2a). After washing out exogenous cytokines, CAR-T cells and target cells were co-incubated. Upon target cell engagement, CAR-T cells divided efficiently (Fig. 2b). Compared with gRNA counterparts, the proliferations of CAR-T cells expressing scramble RNA were significantly suppressed (P < 0.05) (Fig. 2b). gRNA CAR-T cells had enhanced cytolytic capacities (P < 0.05) and increased secretions of IL-2 and IFN-γ (P < 0.05) after activation by tumor cells (Fig. 2c and d). To further confirm the effectiveness of ABE in relieving T cell inhibition, we checked the anti-tumor functions of CAR-T cells in vivo. Consistently, CAR-T cells expressing N74-targeted gRNA attained greater expansion (P < 0.05) (Fig. 2e and Additional file 2: Figure S2). Decreased surface PD-1 (P < 0.01) and upregulated activation markers (CD69 and CD27) (P < 0.05) were noticed on gRNA CAR-T cells (Fig. 2f and g). gRNA CAR-T cells more efficiently delayed tumor growth and improved overall survival when compared with scramble counterparts (P < 0.05) (Fig. 2h–j) (Additional files 3 and 4).

Single base editing can modulate the stability and function of target protein by changing a single residue [8]. Our work further uncovered the potential of such editing tool in T cells. Compared with CRISPR/cas9, ABE has narrower editing window and much less frequent off-target events [9], representing a safer and more precise approach for gene editing. ABE-mediated point mutation can downregulate the inhibitory PD-1, therefore providing an alternative approach to augment T cell immunotherapy.