Figure 5.

Analysis of PD-1 nanoclusters from STED microscopy data. (a) Point cloud and corresponding cluster density maps. Representative cluster density maps from each flat sheet condition, with intensity corresponding to the local density of localized clusters (number of clusters per Gaussian kernel with bandwidth = 200 nm). Scale bar = 3 μm. (b) Example image of k-nearest neighbor analysis for a cluster of interest (middle blue cross) with the four nearest neighbors (surrounding blue crosses) and the distances d_n between them (yellow lines). (c) Example image showing cluster size analysis of an individual cluster with a line profile marked and plotted (white circles), providing a full width at half-maximum (FWHM) of 68 nm from a Lorentzian fit to the data (yellow line). (d–f) Results from respective analysis; (d) peak intensity distributions from cluster density maps, (e) nearest neighbor distance distributions and (f) cluster size distributions shown as mean kernel density estimation (solid lines) of the resulting distribution from each cell (N = 24). Corresponding 83.4% confidence intervals of the mean are indicated by shaded areas per condition, FS-PD-L1–13 (orange), FS-PD-L1–200 (blue) and FS-α CD3-CD28 (gray). Regions of significant difference between conditions are marked by colored bars (bottom, FS-PD-L1–13/FS-α-CD3-CD28 (orange/gray), FS-PD-L1–200/FS-α-CD3-CD28 (blue/gray) and FS-PD-L1-13/FS-PD-L1-200 (orange/blue)). N = 24 cells for each flat sheet condition from 2 glasses from 2 days of experiment (N₁ = 15, N₂ = 9 for each condition), a total of 72 cells were analyzed (24 cells × 3 conditions).