Fig 2: IcAR-PD1 expresses functional hPD1 receptors.(A) Schematic representation of the IcAR system, in which the ectodomain of the human PD1 receptor is fused to the full murine CD3ζ chain. (B) Flow cytometry shows surface expression of hPD1 in the IcAR-PD1 cell line (top) versus hPD1 expression in wild-type BW-5147 cells (middle) and isotype control (bottom). (C) Schematic representation of the functional activation of the IcAR-PD1 cell line in culture plates with wells precoated with anti-PD1 antibodies. Interaction between IcAR-PD1 cells and the anti-PD1 antibody resulted in the secretion of murine IL-2. (D) Stimulation of IcAR-PD1 cells with coated and soluble anti-PD1 antibodies, pembrolizumab (Pembro) and nivolumab (Nivo), and coated anti-PDL1 antibodies, durvalumab (Durva) and avelumab (Avelu), or isotype control (hIgG4). (E) Flow cytometry analysis of IcAR-PD1 cells with rhPDL1-FC (rhPDL1-FC). (F) Flow cytometry analysis of IcAR-PD1 cells with rhPDL2-FC (rhPDL2-FC). FITC, fluorescein isothiocyanate. (G) Schematic representation of the response of IcAR-PD1 cells to different ligands. (H) Response of IcAR-PD1 to different amounts of plastic-bound rhPDL1-FC and rhPDL2-FC (I) Schematic representation showing IcAR-PD1 cells blocked by different antibodies. Ab, antibody. (J) IcAR-PD1 cells were incubated with equimolar concentrations of rPDL1 and PDL2 without antibodies (unblocked) or with blocking antibodies: anti-PD1 antibody (pembrolizumab), anti-PDL1 antibody (durvalumab), anti-PDL2 antibody, and combination of anti-PDL1 and anti-PDL2 antibodies.

Fig 3: IcAR response to fresh and fixed PDXs and generation of an IcAR score.(A) Schematic representation of the IcAR assay protocol. Following flow cytometry determination of surface expression levels of PDL1 and PDL2 and coexpression of PDL1 and PDL2 (double) in fresh PDXs (extracted from mice), the IcAR functional assay was performed on both unblocked ligands and ligands blocked using anti-PDL1, anti-PDL2, and anti-PD1 antibodies. (B) PDL1 and PDL2 levels were determined in five PDXs, and the assay results are presented as percent of surface expression of each ligand and of coexpression (double). (C) Coculture with IcAR-PD1 cells was performed for each PDX using anti-PD1 (pembrolizumab), anti-PDL1 (durvalumab), and anti-PDL2 antibodies; results are displayed as percent of response to each ligand out of the total IcAR response (unblocked). (D) Coculture with IcAR-PD1 cells was performed for each FFPE tissue sample using anti-PD1, anti-PDL1, and anti-PDL2 antibodies; results are displayed as percent of response to each ligand out of the total IcAR response (unblocked). (E) Correlation between IcAR response (namely, IL-2 levels) from fresh and fixed PDXs (n = 5). (F) Example of surface coverage of a well, measured with a JuLI Stage cell recorder, to produce normalized response per surface area. (G) Different PDXs give different scores, depending on the functionality of binding of each ligand (**P < 0.01 and ****P < 0.001). n.s., not significant.

Fig 4: PD1 and PDL1 score can distinguish between responders and nonresponders.(A) Correlation between PD1 score (X) and PDL1 score (Y) for each patient (Pearson R = 0.9596, P < 0.0001, n = 42); dotted lines indicate the 95% confidence intervals (CIs) for the regression line. (B) PD1 score in responder and nonresponder groups (****P < 0.0001, n = 42). (C) PDL1 scores in responder and nonresponder groups (****P < 0.0001, n = 42). (D) Survival curve of patients with PD1 score ≥ 4 and patients with an PD1 score < 4 (log-rank P = 0.0008, n = 42). (E) Logistic regression of the predictive value of PD1 and PDL1 scores for a patient’s clinical response is shown as a ROC curve: Red line depicts a random model, green line depicts PD1 score (AUC = 0.990, P < 0.0001, n = 42), and blue line depicts PDL1 score (AUC = 0.9863, P < 0.0001). (F) Comparison between PD1 score (≥4 and <4) and PDL1 staining. (G) Ligand distribution and prognosis in four patients, one from each of the groups shown in (F). The outer circle shows IcAR response patterns of four patients, and the inner circle shows the patient’s clinical response, where PR is the partial response and PD is the progressive disease.

Fig 5: IcAR-PD1 can discriminate between PDL1 and PDL2 in cancer cell lines.(A) Expression levels of PDL1 and PDL2 were measured for each cell line (basal and cytokine-induced) using flow cytometry; the IcAR functional assay was performed on both unblocked and blocked ligands using anti-PDL1, anti-PDL2, and anti-PD1 antibodies. (B) The IcAR response to nine different cell lines increased upon cytokine induction (paired t test, P = 0.002). (C) The IcAR response to PDL1 (in the presence of anti-PDL2 mAb) increased upon cytokine induction in most cell lines (paired t test, P = 0.0039) (D) The IcAR response to PDL2 (in the presence of anti-PDL1 mAb) was increased upon cytokine induction in most cell lines (paired t test, P = 0.0156). (E) Correlation between the IcAR-PDL1 response and PDL1 surface expression (background subtracted) on tumor cells (n = 18). (F) Correlation between the IcAR-PDL2 response and PDL2 surface expression (background subtracted) on tumor cells (n = 18). (G) Prediction of response to PDL1 calculated by subtracting the IcAR response post PDL1 blockade from the overall IcAR response, without the use of an anti-PDL2 antibody (n = 18). Differences were considered to be statistically significant at a two-sided *P < 0.05 and **P < 0.01. PE, phycoerythrin; APC, allophycocyanin; GeoMean, geometric mean fluorescence intensity.