Fig. 2. Key framework residues on HLA molecules confer allelic interactions with 10LH.

(A) Schematic depicting cross-HLA interactions by 10LH. (B) Sequence identity across 219 HLA-I alleles computationally predicted to present the PHOX2B peptide mapped onto the HLA-A*24:02 surface (PDB ID 8EK5). The α₁ and α₂ helices and HLA framework residues are denoted by the black outlines and dashed boxes, respectively, on the HLA-A*24:02 surface. (C) Schematic depicting the HLA framework residues on the α₁ helix. Below, a sequence motif of the HLA framework residues across common alleles predicted to bind the PHOX2B peptide (n = 219). Created using Seq2Logo. (D) SPR determined K_D values for PHOX2B/HLA-A*24:02/β₂m with site-directed mutations of the framework residues. Data are mean ± SD for n = 3 technical replicates. The blue dashed line depicts the K_D of the WT PHOX2B/HLA-A*24:02/β₂m to 10LH. (E) Scatter plot of sequence similarity of HLA interface residues (BLOSUM62 score) against the SPR binding penalty for the predicted 10LH binders. The SPR binding penalty was calculated by adding the log base 10 ratio of the mean K_D value for each polymorphism in a given allele to the wild-type K_D. The gray dashed line indicates the presence of at least one non-binding polymorphism. (F) SPR determined K_D values for selected HLA allotypes. Polymorphisms among the framework residues are listed above non-binding alleles. Data are mean ± SD for n = 3 technical replicates. NB, no binding. (G) Structural model of PHOX2B/HLA-A*24:03 with polymorphisms relative to HLA-A*24:02 denoted in red sticks. The zoom-in details the steric clash (red curve) between the polymorphic residue (red) and the PHOX2B peptide (blue). (H) A summary of SAB screening, where beads are pre-incubated over weekend (o/w) with PHOX2B or PHOX2B^R6A peptides, and stained by 10LH PE-tetramers for a shift in MFI. (I) Correlation of log₁₀(MFI) for 10LH stained HLA allotypes pre-incubated with PHOX2B and PHOX2B^R6A. Data are mean for n = 3 technical replicates with SD and individual data points are shown in fig. S12. The black dashed line represents a conceptual 1:1 correlation (no difference between PHOX2B- and PHOX2B^R6A-dependent 10LH interactions). (J) In vitro cytotoxic activity of 10LH.BBz CAR T cells against HLA-A*24:03 neuroblastoma cell line SK-N-DZ. 10LH.BBz CAR T cells were co-cultured (black dashed line indicates addition of effector cells to target cells) at a 3:1 E:T ratio with SK-N-DZ target cells. Target cell viability was determined using the cell index of the xCELLigence Real-Time Cell Analysis (RTCA) system. Target cell index was normalized to the timepoint immediately before addition of effector cells. Values represent mean ± SD using effector cells from n = 2 biological donors, in triplicates.