T cells are the major components of the immune infiltrate in LCH lesions. (A-B) viSNE plots of live CD45⁺ cells from a representative sample of an LCH lesion (A) and a biopsy specimen of control healthy skin (B). Each dot in the viSNE plots represents a single cell, and the axes represent arbitrary units based on the t distributed SNE (t SNE) algorithm. The general phenotype in multidimensional space is represented by the relative position of a cell on the plot. (C) Quantification of LCH lesion-infiltrating cells shown in panel A (n = 15) in comparison with the normal skin specimens shown in panel B (n = 5). (D) Relative T-cell subtypes in LCH. Analyses of CyTOF data were performed with the FlowSOM algorithm and visualized as pie charts, showing the mean values for the indicated markers. The background colors of the nodes result from metaclustering corresponding to an automatic gating procedure to avoid bias in the analyses. The size of the nodes is relative to the percentage of cells present in each cluster. Shown are representative FlowSOM minimal spanning tree (MST) plots for the indicated T-cell types. (E) Quantification of the relative abundance of CD4⁺ and CD8⁺ T cells in the LCH lesion infiltrate (n = 5), healthy tonsil control (n = 5), or normal skin (n = 3). (F) CD4/Treg (C25⁺FOXP3⁺) ratios in the indicated specimens (n = 5 each for LCH and tonsil; n = 3 for skin control samples). For all panels, LCH lesions were derived from liver, bone, or spleen. For panels C, E, and F, data are mean ± SD. **P < .01; ****P < .0001, ns, not significant (P > .05). WMW test in panel C, 1-way ANOVA with Tukey’s post hoc test in panels E and F. NK cell, natural killer cell; SNE, stochastic neighbor embedding; viSNE, visualization of t distributed SNE.