High-affinity T cell receptor ImmTAC® bispecific efficiently redirects T cells to kill tumor cells expressing the cancer-testis antigen PRAME

Abstract

Background: PRAME (Preferentially expressed Antigen in Melanoma) is a cancer-testis antigen expressed in several tumor indications, representing an attractive anticancer target. However, its intracellular location limits targeting by traditional methods. PRAME peptides are presented on the surface of tumor cells by human leukocyte antigen (HLA) molecules, indicating that a T cell receptor (TCR)-based strategy that redirects T cells to kill PRAME⁺ tumors could be a novel immunotherapeutic option. We confirm that PRAME protein is expressed in cutaneous melanoma, including rare subtypes with limited treatment options, as well as primary and metastatic lung, breast, endometrial, and ovarian tumors. Furthermore, PRAME is expressed homogeneously across tumors with distinct oncogenic mutations, mutation burden, PD-L1 expression, immune infiltration, and features of immune checkpoint resistance. Immunopeptidomic analysis of primary tumors detected HLA class I-restricted PRAME peptides.

Methods: A TCR recognizing PRAME peptide SLLQHLIGL was engineered to high affinity and fused to a CD3 engaging domain to create a TCRxCD3 bispecific molecule (Immune-mobilizing monoclonal TCR Against Cancer, ImmTAC®) with the ability to redirect polyclonal T cells to efficiently kill PRAME⁺ cells.

Rs: The degree of T cell activation was positively correlated with peptide-HLA abundance, with as few as 10 epitopes per cell sufficient for target cell killing. Impaired ImmTAC®-redirected cytotoxicity of exhausted T cells was rescued using an anti-PD-1 antibody, supporting the use of a combination strategy to treat tumors with active PDL1-PD1 axes.

Conclusions: Our data demonstrate selective and efficient T cell activation and killing by a PRAME-directed TCRxCD3 bispecific, supporting further investigation in multiple cancer indications.

Keywords: PRAME; TCR bispecific; immunotherapy; melanoma.

Graphical Abstract

Figure 1.

PRAME protein is homogeneously expressed in patient tumors of multiple types and retained in metastasis. (A) Representative IHC staining images of primary tumor types. PRAME prevalence in (B) primary tumor tissues, (C) primary and metastatic melanoma and lung tumor, and (D) acral and mucosal cutaneous melanoma subsets. The number of samples analyzed is indicated for each tumor type.

Figure 2.

PRAME transcript is widely expressed in melanoma independently of oncogenic driver mutations, mutation burden, immune infiltration status, and checkpoint therapy response. (A) Analyses of PRAME expression in individual melanoma patient tumors within the depicted groups (BRAF: serine/threonine protein kinase B-Raf; NRAS: GTPase N-Ras; TMB: tumor mutational burden; PD-L1). The Wilcoxon test was used to compare PRAME expression between sample groups. (B) Pathway association analyses after unsupervised clustering of bulk melanoma tumor RNAseq, comparing PRAME expression with the presence of infiltrating immune populations. (C) scRNAseq data from Jerby-Arnon et al. [26] relating low and high immune resistance states to PRAME expression across cutaneous melanoma tumors.

Figure 3.

SLLQHLIGL PRAME-derived peptide is presented in the context of HLA class I across different cancer indications. Mirror plots of LC-MS/MS spectra obtained for native peptide (SLLQHLIGL, top) and stable isotope-labeled peptide (SLLQHL*IGL, bottom) in representative examples of cutaneous melanoma, ovarian carcinoma, and NSCLC samples.

Figure 4.

An affinity-enhanced bispecific CD3/TCR ImmTAC® molecule mediates T cell activation and redirected killing of melanoma, lung, and ovarian tumor cells expressing PRAME. HLA-relevant PRAME⁺ melanoma (A), NSCLC (B), and ovarian (C) tumor cells were incubated with PBMC effector cells in the presence of increasing concentrations of the PRAME ImmTAC® IMC-F106C. Ovarian TYK-nu (PRAME^–/HLA-A*02:01⁺) and lung NCI-H1693 (PRAME⁺/HLA-irrelevant) cell lines were included as controls. Effector cells were also cultured with 2 nM IMC-F106C in the absence of target cells, as control. ImmTAC®-mediated T cell activation was determined by IFNγ ELISpot (left panels). $ represents spots too numerous to count by the software and attributed the highest counted value. IMC-F106C-mediated T cell killing was determined as % tumor cell cytolysis in xCELLigence impedance-based assays (right panels). Data were obtained using three healthy PBMC donors and representative data are shown.

Figure 5.

A dose-dependent requirement for intact PRAME antigen provides evidence for the specificity of ImmTAC®-mediated T cell redirection. (A) IFNγ ELISpot analysis of biallelic and tri-allelic deleted MEL624 cell clones K05 and A08 versus wild-type (parental). (B) Representative microscopy data quantitating PRAME-pHLA epitopes for the tri-allelic deleted A08 clone with or without transfection of titrated PRAME mRNA sequence (IVT). (C) Cell lines used in (B) were tested in IFNγ ELISpot to demonstrate rescued ImmTAC®-mediated T cell activation. (D) Example fluorescence microscopy images of PRAME-pHLA epitope detection on three cancer cell lines with varying antigen levels; (E) comparison of PRAME pHLA epitope number with the degree of ImmTAC®-dependent T cell activation (IFNγ ELISpot), area under the curve (AUC), relative to MEL624 (set to 100). Data are representative of at least two experiments per cell line. ELISpot assays were performed with cells from the same healthy PBMC donor.

Figure 6.

PRAME-targeting ImmTAC® redirects PD-1⁺ TILs to kill tumor cells, an effect that is reduced against PD-L1-expressing tumor cells, but alleviated by anti-PD-1 Ab. (A) CD8⁺ TILs from a melanoma tumor sample were FACS sorted into programmed cell death protein 1 (PD-1)⁺ and PD-1^– populations. (B) ImmTAC®-mediated redirection of PD-1^–TILs (left panels) and PD-1⁺ TILs (right panels) resulted in the killing of PD-L1^- tumor cells (MEL624, top panels) and PD-L1⁺ tumor cells (PD-L1 transduced MEL624 cells, bottom panel). The experiment was performed over 4 days in the absence (white circles) or presence (black circles) of blocking anti-PD-1 antibody. (C) % cytolysis area under the curve (AUC) was calculated for the conditions presented in (B), using increasing ImmTAC® concentrations in the presence (black) or absence (white) of anti-PD-1 antibody. Data are shown as mean ± SEM.