Resident T Cells in Resolved Psoriasis Steer Tissue Responses that Stratify Clinical Outcome  Irène Gallais Sérézal, Cajsa Classon, Stanley Cheuk, Mauricio Barrientos-Somarribas, Emma Wadman, Elisa Martini, David Chang, Ning Xu Landén, Marcus Ehrström, Susanne Nylén, Liv Eidsmo  Journal of Investigative Dermatology  Volume 138, Issue 8, Pages 1754-1763 (August 2018) DOI: 10.1016/j.jid.2018.02.030 Copyright © 2018 The Authors Terms and Conditions

Figure 1 Cytokine-producing T cells in healthy and psoriasis-afflicted skin. (a, b) Representative FACS dot plots of live CD3+TCR γδ– cells producing (a) IL-17A or (b) IFN-γ in healthy and untreated psoriasis skin. (c, d) Pie charts of the composition of (c) IL-17A and (d) IFN-γ–producing T-cell subsets in epidermis and dermis. Gating strategies are depicted in Supplementary Figure S1. (e) Visualization of αCD3 mouse OKT-3 antibody (green) and CD3 expressing cells (red) in healthy skin after 24 hours ex vivo incubation with OKT-3. (f) Representative FACS dot plots of cytokine-producing epidermal live CD45+CD3+ cells after 24 hours incubation of full-thickness healthy skin biopsy samples with OKT-3 or isotype (IgG2a). Derm, dermis; Epi, epidermis; TCR, T-cell receptor. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions

Figure 2 T-cell activation ex vivo elicits production of cytokines and tissue-derived chemokines. (a) Schematic of experimental setup for ex vivo activation of tissue T cells. Skin explants were incubated with OKT-3, IFN-γ, or IL-17 for 16 hours before collection of supernatants and enzymatic separation of epidermis and dermis. (b) Protein concentrations in supernatants were assessed using ProcartaPlex (eBioscience, Waltham, MA). Bar charts depicting mean ± standard deviation from five individual psoriasis or healthy skin biopsy samples. (c–f) Quantitative PCR was performed using B2M as housekeeping gene. Data plotted with the 2–ΔCT method. Each symbol represents one individual donor. Results were considered significant at ∗P < 0.05 between groups. n = 5–11. RQ, relative quantification to the housekeeping gene. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions

Figure 3 OKT-3 stimulation results in psoriasis-independent interferon tissue responses and psoriasis-restricted IL-17 responses. Epidermal responses to OKT-3 stimulation in healthy control individuals (n = 3), untreated psoriasis (n = 3), and resolved psoriasis after anti IL12/23 treatment (n = 3). (a) List of the top four GO pathways (P < 0.05) for differentially expressed genes (DEGs) after OKT-3 treatment in all samples analyzed; n reflects the number of genes that were up-regulated. Alluvial graph illustrates the GO pathways (Benjamini P-value < 0.05) of the DEGs in each group. Line thickness represents the percentage of the group’s DEGs included in each pathway for each group. (b) Venn diagrams of up-regulated DEGs with probability of difference greater than 80% included. (c) Heatmaps of quantitative PCR analysis of gene expression in healthy keratinocytes (biological triplicate from the same patient) or epidermis (median of a group of 5 patients) after incubation with IL-17A or IFN-γ for 16 hours. Heatmaps were log2 transformed and normalized by gene. (d) RNA counts were analyzed with NanosString technologies. Principal component analysis of up-regulated DEGs in untreated, healthy, and treated epidermis after OKT-3 stimulation. (e) Venn diagrams of statistically significant (Wilcoxon P < 0.05) DEGs using NanoString technologies. GO, gene ontology; Trip, triplicate. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions

Figure 4 Tissue response patterns to OKT-3 stimulation after UVB treatment. (a) Heatmap of baseline gene expression with NanoString (generated in MeV, normalized by gene, log2 transformed). (b) RNA counts were obtained by NanoString analysis of epidermal samples after exposure to OKT-3 or IgG2a. Mean ± standard are shown; n = 5 individual donors for psoralen-potentiated UVA and UVB, n = 6 for biologics. Results were considered significant if ∗P < 0.05 (Wilcoxon match-paired signed rank test). (c, d) Confocal microscopy image depicting RORγt- (green) and CD3- (red) expressing cells in (c) untreated psoriasis and (d) after UVB treatment. Scale bar = 100 μm. (e–f) Linear regressions comparing the time until clinical relapse after successful UVB treatment with baseline gene expression at the end of the UVB treatment or fold changes of the gene expression after OKT-3 treatment compared with IgG2a. PUVA, psoralen-potentiated UVA; RQ, relative quantification to the housekeeping gene. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions

Figure 5 Stratified tissue-responses correlate with clinical outcome after UVB therapy. Gene expression was assessed using quantitative PCR. (a, b) Linear regression was used to explore the relationship between time in clinical remission and the ratio of (a) baseline gene expression of CXCL10 over SPRR2A, SPRR2B, and DEFB4A/B or (b) the ratio between the fold changes after OKT-3 of CXCL10 and SPRR2A, SPRR2B, and DEFB4A/B. Example of formula for DEFB4A/B in b: (CXCL10 after OKT-3/CXCL10 after IgG2a)/(DEFB4A/B after OKT-3/DEFB4A/B after IgG2a). Dotted lines show the correlation line, solid lines show the confidence interval. (c, d) Linear regression comparing time in clinical remission with ratios of net protein concentration changes (OKT-3 – IgG2a) in explant supernatants assessed by ProcartaPlex (eBioscience, Waltham, MA). Example of formula for c: (CXCL-10 after OKT-3 – CXCL-10 after IgG2a)/(IL-17 after OKT-3 – IL-17 after IgG2a). RQ, relative quantification to the housekeeping gene. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions

Figure 6 Activation of TRM cells induces disease-specific tissue responses in resolved psoriasis. The functional phenotypes of TRM cells are skewed in resolved psoriasis with accumulation of IL-17–producing TRM cells in resolved psoriasis lesions. Consequently, ex vivo activation of TRM cells results in IL-17–driven tissue responses that are associated with psoriasis pathogenesis. TRM, tissue resident memory T cells. Journal of Investigative Dermatology 2018 138, 1754-1763DOI: (10.1016/j.jid.2018.02.030) Copyright © 2018 The Authors Terms and Conditions