In-Situ HLA-A2 MMr StainingIn-situ immunohistochemistry staining was performed as described (Culina et al., 2018Culina S. Lalanne A.I. Afonso G. Cerosaletti K. Pinto S. Sebastiani G. Kuranda K. Nigi L. Eugster A. Osterbye T. et al.Islet-reactive CD8+ T cell frequencies in the pancreas, but not in blood, distinguish type 1 diabetic patients from healthy donors.Sci. Immunol. 2018; 3: eaao4013Crossref PubMed Scopus (31) Google Scholar). Unfixed, frozen sections were dried for 2 h, loaded with 1 μg of PE-labeled MMrs overnight at 4°C, washed gently with PBS and fixed in 2% paraformaldehyde for 10 min. After a further wash, endogenous peroxidase activity was blocked with 0.3% H2O2. Sections were then incubated serially with a rabbit anti-PE Ab (Abcam), horseradish peroxidase-conjugated swine anti-rabbit Ig (Dako) and 3,3′-diaminobenzidine tetrahydrochloride substrate (ThermoFisher). After a final wash, sections were counterstained with hematoxylin, dehydrated via sequential passages in 95%–100% ethanol and xylene, mounted, and analyzed using a Nikon Eclipse Ni microscope with NIS-Elements D software v4.40.In-situ immunofluorescence staining was performed similarly, but non-specific reactions were blocked with 5% goat serum for 2 h at room temperature before serial incubations with rabbit anti-PE Ab (1:250, 1.5 h at room temperature) and Alexa Fluor 594-conjugated goat anti-rabbit IgG (ThermoFisher; 1:500, 1 h at room temperature). After a further wash, sections were incubated for 1 h at room temperature with rat anti-CD8 mAb (Abcam; 1:100) together with mouse anti-CD45RO mAb (BioLegend; 1:200) followed, after one wash, by one final incubation for 1 h at room temperature with Alexa Fluor 488-conjugated goat anti-rat IgG together with Alexa Fluor 647-conjugated goat anti-mouse IgG (1:500/each; both from ThermoFisher). After DNA counterstaining with DAPI, sections were mounted and analyzed using a Leica TCS SP5 confocal laser scanning microscope with LAS software v2.6.0.7266.

HLA-A2 MMr AssaysAll peptides were synthesized at >90% purity (Synpeptides). HLA-A2 MMrs were produced and used as described (Culina et al., 2018Culina S. Lalanne A.I. Afonso G. Cerosaletti K. Pinto S. Sebastiani G. Kuranda K. Nigi L. Eugster A. Osterbye T. et al.Islet-reactive CD8+ T cell frequencies in the pancreas, but not in blood, distinguish type 1 diabetic patients from healthy donors.Sci. Immunol. 2018; 3: eaao4013Crossref PubMed Scopus (31) Google Scholar). Each pHLA-A2 complex was used at a final concentration of 8-27 nM and conjugated with fluorochrome-labeled streptavidin at a 1:4 ratio. The combinatorial MMr panel was first set up by staining HLA-A2+ PBMCs with the same set of fluorescent streptavidin-labeled MMrs, all loaded with the Flu MP58-66 epitope. Compensations were set using fluorescence-minus-one samples (i.e. omitting one streptavidin at a time). The concentration of each fluorescent MMr was corrected for the variable staining index of each streptavidin, in order to obtain a distinct double-MMr+ population for each fluorochrome pair. The identification of the same MMr+ population by each pair validated the panel. PBMCs were isolated by density gradient centrifugation using 50 ml Leucosep tubes (Greiner/Dominique Dutscher), washed twice in RPMI medium supplemented with AB human serum (Sigma), counted on a ThermoFisher Countess II automated counter and frozen in pre-chilled 10% DMSO solution in AIM-V medium (ThermoFisher) using CoolCell containers (Biocision) stored overnight at −80°C prior to transfer into liquid nitrogen. At thawing, PBMCs were immediately diluted in pre-warmed AIM-V medium. Following centrifugation and one additional wash in AIM-V, PBMCs were counted and rested in the presence of 50 nM dasatinib for 30 min at 37°C before magnetic depletion of CD8– cells (StemCell Technologies). Staining was performed for 20 min at 20°C in 20 μl PBS-dasatinib for 107 cells with the combinatorial double-coded MMr panels (Culina et al., 2018Culina S. Lalanne A.I. Afonso G. Cerosaletti K. Pinto S. Sebastiani G. Kuranda K. Nigi L. Eugster A. Osterbye T. et al.Islet-reactive CD8+ T cell frequencies in the pancreas, but not in blood, distinguish type 1 diabetic patients from healthy donors.Sci. Immunol. 2018; 3: eaao4013Crossref PubMed Scopus (31) Google Scholar) detailed in Figure 3, followed, without washing, by mAb and Live/Dead Aqua staining at 4°C for 20 min. After one wash, cells were acquired using a FACSAria III cytometer configured as detailed in Table S6. Candidate epitopes binding to HLA-A2 (Figure S2) that did not yield any appreciable MMr staining provided negative controls for each panel. Data were analyzed with FlowJo software as described in Figure 3. Cells were sequentially gated on small lymphocytes, singlets, live cells (Live/Dead Aqua–), CD3+CD8+ T cells and total PE+, PE-CF594+, APC+, BV650+, BV711+ and BV786+ MMr+ T cells. Using Boolean operators, these latter gates allowed to selectively visualize each double-MMr+ population by including only those events positive for the corresponding fluorochrome pair. For example, UCN31-9 MMr+ cells (PE+PE-CF594+) were visualized by gating on events that were PE+PE-CF594+APC−BV650−BV711−BV786−. Events negative for all MMr fluorochromes (PE−PE-CF594−APC−BV650−BV711−BV786−) were represented in the same PE/PE-CF-594 dot plot to set the double-MMr+ gate, as shown in Figures 4A–4F. CD45RA and CCR7 staining was subsequently visualized by gating on these MMr+ cells. Each dot plot of Figures 3B and 3C displays a color-coded overlay of each double-MMr+ fraction and of the MMr− population to visualize the separation of each epitope-reactive CD8+ T-cell fraction relative to the others.