The following antibodies were used: anti-AF700-CD8a (53-6.7), anti-APC-CD154 (MR1), anti-APCCy7-CD8a (53-6.7), anti-PE-CD137 (17B5), anti-PE-CD4 (GK1.5), anti-PerCPcy5.5-CD3 (17A2), anti-BV421-CD62L (MEL-14), anti-BV421-granzyme B (GZMB; QA18A28), anti-BV605-CD4 (RM4-5), anti-BV711-CD44 (IM7), and anti-BV711-interferon gamma (IFN-γ; XMG1.2), purchased from BioLegend (San Diego, CA, United States). Anti-APCCy7-CD19 (1D3) and anti-PE Cy7-FoxP3 (FJK-16s) were purchased from BD Biosciences (San Jose, CA, United States). Nonspecific FcγR binding of labeled monoclonal antibodies (mAbs) was blocked using anti-mouse CD16/32 (2.4G2; BD Pharmingen, Hamburg, Germany). Dead cells were excluded from analysis using the forward Zombie Aqua Fixable Viability Kit (BioLegend), the Zombie NIR Fixable Viability Kit (BioLegend), or 7-aminoactinomycin D (7-AAD; BD Biosciences) staining. For intracellular staining, cells were fixed and permeabilized using the FoxP3/Transcription Factor Staining Buffer Set (BD Biosciences), according to the manufacturer’s instructions. To assess cytokine production, the cells were stimulated using monensin (BD Biosciences) in a culture medium at 37°C in a 5% CO₂ incubator for 4 h prior to staining. The data were collected using LSRFortessa X-20, FACS Canto II, or FACS Celesta (BD Biosciences) and were analyzed using FlowJo v. 10 (Tree Star, Ashland, OR, United States).

C57BL/6 (H-2Db), BALB/c (H-2Dd), and C3H/HeJ (H-2Dk) mice were purchased from CLEA (Osaka, Japan) and maintained in a pathogen-free animal facility of Hiroshima University, Hiroshima, Japan. Female mouse were used at an age of 10–12 weeks. When indicated, the mice were euthanized through cervical dislocation after isoflurane inhalation. All efforts were made to minimize animal suffering [9]. This study was performed in strict accordance with the “Guide for the Care and Use of Laboratory Animals” prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health. All mice received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of the Graduate School of Biomedical Sciences, Hiroshima University (Permit Number: A23-17). A part of this work was performed at the Research Facilities for Laboratory Animal Science, Natural Science Center for Basic Research and Development (N-BARD), Hiroshima University.

Full-thickness skin grafts were transplanted onto the left lateral dorsum of a recipient. Briefly, donor skin tissues were removed from the tails and trimmed into 10 mm × 10 mm strips. Recipient mice were anesthetized using intraperitoneal injection of xylazine (5 mg/kg body weight) and ketamine (100 mg/kg body weight). Skin tissues of the same size were removed from the recipients’ backs and replaced with donor grafts. The skin grafts were covered with bandages for 5 days, and graft survival was evaluated through daily visual inspection. Rejection was defined as destruction of >95% of the skin transplant [10]. An MHC full-mismatch BALB/c into C57BL/6 combination was employed as a rejection model. A BALB/c into C57BL/6 or C3H/HeJ combination previously reported as a tolerance induction model treated with CTLA-4 IgG (abatacept, 200 μg; Bristol-Myers Squibb, Braine-l’Alleud, Belgium) on days 0, 2, 4, and 6, and anti-CD154 antibody (MR1, 250 μg; BioLegend, San Diego, CA, United States) on days 0, 2, and 4 [11] was used for monitoring peripheral tolerance induction.

We prepared mononuclear cell suspensions of BALB/c mouse spleens and purified the B cells via positive selection using CD19 MicroBeads (Miltenyi Biotec, San Diego, CA, United States) in an autoMACS Pro Separator (Miltenyi Biotec), according to the manufacturer’s instructions [9]. The purity of the sorted cells was consistently >95%. Using a cocktail of recombinant mouse CD40L multimer (100 ng/mL; AdipoGen, San Diego, CA, United States) and recombinant mouse IL-4 (10 ng/mL; R&D Systems, Minneapolis, MN, United States), activated B cells were generated by culturing 0.2 × 10⁶ cells/mL at 37°C under 5% CO₂ for 24 h. All cell cultures were performed in complete medium [RPMI 1640 medium (Nacalai Tesque, Kyoto, Japan) supplemented with 5% fetal bovine serum (SERANA, Pessin, Germany), 100 mM sodium pyruvate (Thermo Fisher Scientific, Waltham, MA), 100 U/mL penicillin–streptomycin (Thermo Fisher Scientific), 1% HEPES buffer (Thermo Fisher Scientific), and 50 µM 2-ME] in a 48-well flat-bottom plate. Using activated B cells as stimulators, MLR culture was performed, after which alloreactive T cells were identified. Prior to culturing, the stimulators were irradiated with 40 Gy. Responder T cells were purified from recipient splenocytes via negative selection, using a Pan T-Cell isolation kit (Miltenyi Biotec) in the autoMACS Pro Separator (Miltenyi Biotec), according to the manufacturer’s instructions. The purity of the sorted cells was consistently >95%. Responders and stimulators were co-cultured at a 1:1 ratio (10⁶ cells each) in 96-well U-bottom plates, with 200 µL complete medium containing APC-conjugated anti-CD154-labeled mAbs (MR1, 1 μL; BioLegend) for 18 h. Protein transport inhibitor (monensin, 2 μL; BD Biosciences) was added to the culture medium for the last 4 h of incubation. Alloreactive CD4⁺ and CD8⁺ T cells were identified as CD3⁺CD4⁺CD154⁺ and CD3⁺ CD8⁺CD137⁺ responders, respectively. We collected at least 100,000 counts during flow cytometry acquisition for detecting 0.1% population to keep the coefficient of validation up to 10%.

Recipient splenocytes were labeled with 5 µM carboxy fluorescein succinimidyl ester (CFSE; Molecular Probes) for 5 min prior to culturing. The activated B-cell stimulators were prepared as described in cATD Assay. Responders and stimulators were co-cultured at a 1:1 ratio (2 × 10⁵ cells each) for 4 days in 96-well U-bottom plates with 200 µL medium. Attenuation of CFSE fluorescence intensity was evaluated as proliferating activity gated on CD4⁺ and CD8⁺ T cells. Mitotic index (MI) was calculated as previously described [12, 13].

Statistical analyses were performed using JMP 16 (SAS Institute, Cary, NC, United States). Chi-square or Fisher’s exact test was used to compare categorical variables, and Student’s t-test or Mann–Whitney U-test was used for continuous variables. Comparisons between groups were made using the one-way analysis of variance (ANOVA), and significant differences were examined using Tukey–Kramer’s multiple-comparison post-hoc test. Differences with p < 0.05 were considered statistically significant.

Skin allografts were rejected from 7 to 15 days in the full MHC mismatched rejection model (BALB/c into C57BL/6) (MST 11 days, Supplementary Figure S1). We did not observe a sensitized reaction in peripheral CD4⁺ and CD8⁺ T cells at 3 days after transplantation, as determined using a proliferation assay (syngeneic vs. rejection model, median MI; CD4⁺ 0.18 vs. 0.08, p = 0.53 (upper) and CD8⁺ 0.20 vs. 0.10, p = 0.80 (lower), Figure 1). Seven days after transplantation, we observed a higher proliferation response of both CD4⁺/CD8⁺ T cells in the rejection model than in the syngeneic model (median MI; CD4⁺ 0.18 vs. 0.46, p < 0.05 (upper) and CD8⁺ 0.57 vs. 1.28, p < 0.05 (lower), Figure 1). The cATD assay revealed a sensitized immune response after skin transplantation at the same time point as the proliferation assay, showing a higher proportion of donor-reactive CD4⁺CD154⁺/CD8⁺CD137⁺ T cells than that in the syngeneic model at 7 days after transplantation (syngeneic vs. rejection model, CD4⁺CD154⁺ in total CD4⁺; 1.0% vs. 1.4%, p < 0.05 (upper) and CD8⁺CD137⁺ in total CD8⁺; 0.27% vs. 0.53%, p < 0.05 (lower), Figure 2). Donor-reactive T cells identified in this assay showed an increase in proportion and enhancement in function under antigen-specific stimulation in recipients sensitized with BALB/c mouse graft (Supplementary Figure S2). The multiparametric flowcytometric analyses demonstrated a unique functionality of donor-reactive CD8⁺ T cells in the rejection model; for instance, the production ability of the crucial effectors, GZMB and IFN-γ, was specifically enhanced in donor-reactive CD8⁺ T cells in the rejection model at 7 days after transplantation (syngeneic vs. rejection model, % positive for GZMB 16.1% vs. 54.7%, p < 0.05 (upper), and IFN-γ 1.82% vs. 8.18%, p < 0.05 (lower), Figure 3). As a proof of sensitization, effector memory T (TEM; CD44⁺CD62L⁻) cells were enriched in the donor-reactive population after transplantation (syngeneic vs. rejection model, median % TEM in CD4⁺CD154⁺ 20.2% vs. 32.9%, p < 0.05 (upper), and CD8⁺CD137⁺ 9.2% vs. 19.5%, p < 0.05 (lower), Figure 4).

Permanent engraftment was observed in C3H/HeJ recipients of the full MHC mismatched BALB/c graft treated with CTLA-4 IgG and anti-CD154 antibody (treated tolerance (TT) model, ≥30-day survival was recorded in 16/19 animals, 84.2%), whereas all C57BL/6 recipients with the same immunosuppression eventually experienced allograft rejection within 20 days (treated rejection (TR) model, Figure 5). We investigated the immunological status at 7 and 30 days after transplantation, that is, before and after rejection, respectively. The proportion of FOXP3⁺ Tregs in CD4⁺ T cells was comparable, despite tolerance induction, between 7 and 30 days after transplantation (Supplementary Figure S3). The proliferation assay conducted at 7 days after transplantation showed a significant reduction in response to immunosuppression in both C3H/HeJ and C57BL/6 recipients compared with that in an untreated rejection (UR) model. However, the conventional proliferation readout results did not show the differential immune response at 7 days after transplantation between C3H/HeJ and C57BL/6 recipients, despite a different outcome (TT model vs. TR model, median MI; CD4⁺ 0.41 vs. 0.21, p = 0.44 and CD8⁺ 0.15 vs. 0.19, p = 0.70, respectively, Figure 6A). The cATD assay revealed that the proportion of CD8⁺ donor-reactive T cells in the TT model was lower than that in the TR model at 7 days after transplantation (TT model vs. TR model, median % donor-reactive CD8⁺; 0.18% vs. 0.35%, p < 0.05, Figures 7A, B). The GZMB- and IFN-γ-producing capacity of the CD8⁺ donor-reactive T cells was comparatively low in the three groups (Figures 7C, D). Regardless of the final outcome, models with immunosuppression exhibited impaired memory formation in donor-reactive T cells (Supplementary Figure S4). At 30 days after transplantation, the proliferation assay showed a lower MI of CD8⁺ T cells for the response of the TT model than that for the response of both UR and TR models (TT model vs. UR and TR models, median CD8⁺ MI: 0.18 vs. 0.93 and 0.66, p < 0.05, respectively, Figure 6B). The cATD assay performed at 30 days after transplantation revealed that donor-reactive T cells were detectable in the TT model, similar to those in the UR and TR models (UR vs. TT vs. TR, %CD4⁺CD154⁺ in total CD4⁺ was 2.01%, 1.77%, and 2.35%, %CD8⁺CD137⁺ in total CD8⁺ was 1.00%, 0.7%, and 0.81%, respectively, Figures 8A, B). As expected, the functionality of donor-reactive CD8⁺ T cells in the TT model was lower than that in the UR model (UR vs. TT model, % positive in donor-reactive CD8⁺ T cells, GZMB; 31.9% vs. 11.2%, p < 0.05, and IFN-γ; 33.9% vs. 3.04%, p < 0.05, respectively, Figures 8C, D). However, there were no differences in functionality and memory formation between the TT and TR models (Supplementary Figures S4, S5).