bert vogelstein  (Addgene inc)

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: Activation-induced marker assays are inherently variable and detect heterogeneous subsets of activated CD4 + T cells (A–D) Representative flow cytometry data and quantification of paired raw unstimulated and pp65-stimulated AIM + frequencies among CD4 + T cells (A and B) and CD8 + T cells (C and D). CV (%) is displayed above the responses for each donor. (E and F) Source(s) of variance in CD4 + (E) and CD8 + (F) AIM + cell frequencies calculated after subtracting the corresponding unstimulated AIM + frequencies, using mixed effects analysis, with donor and experiment (technical replicate) as random effects. Graphs show variance components as the percentages of total variance. (G and H) Uniform manifold approximation and projection (UMAP) analysis was performed on CD4 + (G) and CD8 + (H) total AIM + T cells, defined by co-expression of at least one AIM pair. (I and J) FlowSOM clustering of total AIM + CD4 + T cells. (I) Each circle represents one cluster and its relative marker expression; similar clusters are grouped into 8 metaclusters (colored rings); cluster/metacluster spatial proximity indicates similarity. (J) Heatmap showing relative marker expression in the 8 FlowSOM metaclusters. (K) AIM + CD4 + T cell percentages within each FlowSOM metacluster across individual donors. Replicate AIM assays were performed on cryopreserved PBMCs from 5 donors, with 4–8 technical replicates per donor across 2–4 independent experiments; PBMCs were stimulated for 20 h with CMV pp65 peptides. Total AIM + populations from pp65-stimulated samples were pooled from all donors and technical replicates (G–J) or all technical replicates within each donor (K) prior to analysis.

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Activation Assay, Marker, Flow Cytometry, Expressing

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: Box-Cox transformation-based correction for background enhances signal detection and reduces variability (A) Re-analysis of AIM assay data from SARS-CoV-2 spike peptide-stimulated PBMCs from solid organ transplant recipients ( n = 33) vaccinated with three doses of a COVID-19 mRNA vaccine (PREVENT cohort), or (B) new analysis of CMV pp65-stimulated PBMCs from 6 healthy donors across 3 independent experiments at each of 4 research centers in the CAN-ASC cohort ( n = 52 total). Shown are correlation and linear regression analyses of unstimulated AIM + frequencies with raw antigen-stimulated AIM + frequencies, net responses after subtracting or dividing by the unstimulated condition. CD134 + CD25 + of CD4 + T cells and CD137 + CD69 + of CD8 + T cells are presented. p values are from Spearman’s correlation tests. Solid lines represent linear regression with error bars indicating 95% confidence intervals (CIs). For the CAN-ASC cohort, regression and correlation analyses were performed after correcting for between-donor variations. (C) Box-Cox transformation workflow to correct for AIM + cells in unstimulated controls, using a user-defined parameter λ ϵ[0,1], yielding a stimulation index (SI). (D and E) Three statistical methods to estimate optimal λ values by minimizing correlation between the unstimulated and net stimulated AIM values. Left: linear regression to identify λ giving the highest probability of zero slope ( β ˆ = 0 ), expressed as a posterior probability distribution for λ. Center: Spearman’s correlation to identify the value of λ producing an estimated zero correlation. Right: likelihood profile for λ based on linear regression to estimate the probable optimal value of λ. Values were estimated for (D) PREVENT or (E) CAN-ASC cohorts. (F) ICC values as a function of λ in the CAN-ASC cohort, using the SI for CD4 + and CD8 + AIM responses to CMV pp65. Gray dotted lines show ICC values for raw (untransformed) AIM + frequencies. (G) Comparison of bootstrap-adjusted CVs calculated between technical replicates within donors for raw AIM + frequencies, net subtraction-/division-corrected AIM + frequencies/ratios, or Box-Cox-corrected SIs. Each point is the bootstrap-adjusted CV of technical replicates from one donor. p values represent Dunnett’s multiple comparisons test following one-way repeated-measures ANOVA. (H and I) Re-analysis of AIM responses for the PREVENT cohort throughout a three-dose COVID-19 mRNA vaccination schedule calculated by subtracting or dividing by unstimulated AIM + frequencies, or Box-Cox-corrected SI values. p values represent Dunnett’s multiple comparisons test following mixed-effects analysis of log 2 -transformed data. Bold p values denote significant differences between time points detected only with Box-Cox-corrected SI values.

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Transformation Assay, Comparison

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: AIM responses are reliably detectable across various cell sources (A) Percentage viable of total CD3 + T cells in AIM assays. Error bars represent SD. p values represent comparisons using Dunn’s multiple comparisons test following Friedman test. (B–E) Representative flow cytometry (B and C) and SI quantification (D and E) for each AIM relative to the unstimulated control among CD4 + (B and D) and CD8 + (C and E) T cells from each cell source. (F) Ratio of AIM SIs between CMV pp65 protein- and peptide-stimulated CD4 + T cells in relation to cell source. Positive or negative ratios indicate greater detection of AIM responses with whole protein or peptide stimulation, respectively, while 95% CIs overlapping zero indicate no significant difference between protein and peptide stimulation. AIM responses using fresh whole blood, fresh PBMCs, or cryopreserved PBMCs (PBMC-C), with or without a 24-h delay in processing (−24 h) are compared. Cells were rested overnight and stimulated for 20 h with CMV pp65 peptides, CMV pp65 whole protein, or Infanrix vaccine. Error bars are 95% CIs calculated on a log 2 scale (D, E and F), and data are from n = 6 healthy donors (A–F).

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Flow Cytometry, Control

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: A 20-h stimulation period reliably detects most CD4 + and CD8 + T cell AIMs Fresh whole blood or cryopreserved PBMCs (PBMC-C) were stimulated with CMV pp65 peptides for 6, 20, or 44 h. (A) Representative CD134 + CD25 + frequencies of CD4 + T cells. (B) SI quantification of AIM + CD4 + T cells. (C) Representative CD137 + CD69 + frequencies of CD8 + T cells. (D) SI quantification of AIM + CD8 + T cells. p values represent one-sample t tests (B and D), with AIM signals considered present when the mean log 2 -transformed SI significantly differed from zero. Paired samples are from n = 4 healthy donors (A–D).

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Transformation Assay

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: CD134 and CD137 optimally detect antigen-specific Tregs (A) Representative flow cytometric gating strategy for FOXP3 + HELIOS + Tregs. (B) Effect of stimulation with CMV pp65 peptides on FOXP3 + and FOXP3 + HELIOS + frequencies in CD4 + T cells. p values represent Tukey’s multiple comparisons test following two-way ANOVA. (C–E) SI and representative flow cytometry for AIM pairs and Boolean combinations in FOXP3 + HELIOS + Tregs following stimulation of cryopreserved PBMCs with pp65 peptides or whole protein, SARS-CoV-2 spike peptides, or Infanrix combination vaccine for 20 h. Boolean AND/OR gating combinations comprise CD134/CD69, CD134/CD137, and CD137/CD69 (3xAIM); or 3xAIM plus CD154/CD69, CD134/CD154, and CD137/CD154 (6xAIM). Mean log 2 -transformed Box-Cox-corrected SIs are shown. (F) Gating strategy to evaluate CD39 + frequencies among CD134 + CD137 + and non-CD134 + CD137 + FOXP3 + HELIOS + Tregs and FOXP3/HELIOS − Tconvs. (G) Median percentages of CD39 + among CD134 + CD137 + or non-CD134 + CD137 + CD4 + FOXP3 + HELIOS + Tregs or conventional FOXP3/HELIOS − T cells. p values represent Dunnett's multiple comparisons test following repeated measures one-way ANOVA on log2-transformed data. (H) Positive predictive value (PPV) of CD39 or CD39, CD134, and CD137 as markers for antigen-specific Tregs (CD134 + CD137 + FOXP3 + HELIOS + ). Error bars represent 95% confidence intervals of the mean (C, E, and H) or interquartile ranges (G). Healthy donors: n = 9 (pp65 and Infanrix 20-h stimulations) and n = 4 (all others).

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Flow Cytometry, Transformation Assay

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: CMV- and SARS-CoV-2-specific AIM assays are reproducible within and between research centers (A) Experimental design for AIM assay inter-site testing. Replicate aliquots of PBMCs from n = 6 healthy donors assayed in three independent experiments (individual points) by four research groups (coded A–D). Replicate AIM assay experiments ( n = 3 per donor, per site) were performed by stimulating PBMCs for 20 h with CMV pp65 or SARS-CoV-2 spike peptides. (B and C) Box-Cox-corrected SIs of CD134 + CD25 + frequencies among CD4 + T cells and CD137 + CD69 + among CD8 + T cells. Means and SD of log 2 -transformed SIs are shown. Colors indicate the CV within each set of three replicates at each site, with overall CV between sites shown at the top for each donor. (D) Percent CV in CD4 + and CD8 + T cell AIM SIs calculated between technical replicate AIM assays within each site for unstimulated, pp65-stimulated, and spike-stimulated conditions. Each point represents the CV of three technical replicates within one site; bars represent individual donors. CV means with SD are shown. (E and F) ICCs comparing technical variability among CD4 + and CD8 + AIM responses to (E) CMV pp65 and (F) SARS-CoV-2 spike, using data from replicate AIM assays from different sites and donors. Bars represent the mean ICC for each AIM across all sites. (G and H) Percent CV in CD4 + and CD8 + T cell AIM responses to (G) CMV pp65 and (H) SARS-CoV-2 spike between the sites, calculated after averaging three technical replicate AIM assays per site. Each point represents one donor. Error bars represent 95% confidence interval of the mean. (I) ICCs comparing calculated technical variability for replicate AIM assays between sites to biological variability between donors for CD4 + and CD8 + T cells. Dotted lines represent CV = 30% (in D, G and H) and indicate ICC = 0.5 (equal biological and technical variability) (E, F, and I).

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Transformation Assay

Journal: Cell Reports Methods

Article Title: Reproducible detection of antigen-specific T cells and Tregs via standardized and automated activation-induced marker assay workflows

doi: 10.1016/j.crmeth.2026.101458

Figure Lengend Snippet: An automated gating pipeline performs comparably to manual analysis for AIM assays (A) Individuals were trained on identical gating protocols and analyzed the same data files from three technical replicates of 20-h AIM assays with CMV pp65 or SARS-CoV-2 spike peptides on healthy donor PBMCs ( n = 6) collected at four sites designated A–D. Raw flow cytometric data from each replicate at each site were analyzed independently by 3 individuals. Data represent CVs between analysts as the mean CV (with error bars showing 95% CIs) for the three technical replicates of each donor-site combination. (B and C) An automated AIM gating pipeline was created; flow cytometric gating approaches comparing manual and automated analyses are shown for (B) CD4 + CD134 + /CD25 + T cells and (C) CD8 + CD137 + /CD69 + T cells. The CytoStim condition was used by the automated software to set donor-specific AIM gates, which were then applied to other stimulations. (D–F) Automated and manual analyses of multi-center CMV and SARS-CoV-2 AIM assay data. A central analyst (C1) defined the gating strategy, provided instructions to the manual analysts (C2 and O1–O4), and was the reference for comparisons. C1, C2, and the automated software each analyzed all data from all sites, while O1–O4 each analyzed data from a single site (O1: A; O2: B; O3: C; O4: D). Box-Cox-corrected AIM SI values are shown for (D) CMV and (E) SARS-CoV-2 AIM assays. p values represent paired comparisons between each analyst and C1 from post-hoc Dunnett’s multiple comparisons test following mixed-effects analysis and were used to compute (F) total numbers of comparisons for C2 and automated analyses that significantly differed from C1 for CD4 + AIMs, with each of the 32 combinations of CD4 + AIM, site, and antigen (CMV or SARS-CoV-2) considered one comparison. Pooled comparisons were analyzed using Fisher’s exact test. (G–I) Spearman’s correlations of AIM SI values for C1, C2, O1–O4 (pooled), and automated analyses for CMV and SARS-CoV-2. Comparisons are shown against (G and H) the reference C1 and (I) for each analyst against all others in a correlation matrix. (J) F1 scores calculated for C2 manual vs. C1 reference (gray) or automated vs. C1 reference (green) analysis. Each point represents the F1 score from a unique donor-site combination, with p values from paired Wilcoxon signed-rank test after averaging F1 scores from technical replicates.

Article Snippet: Cells were incubated for 6, 20 or 44 h with media, 1.5 μg/mL PepTivator CMV pp65 (Miltenyi), 3.3 μg/mL CMV pp65 recombinant protein (Miltenyi), 1 μg/mL PepTivator SARS-CoV-2 Prot_S (Miltenyi), or 1/100 Infanrix hexa (GSK) or 1/400 CytoStim (Miltenyi).

Techniques: Software, Comparison