Detection and amplification of epitope-specific T cells hold great promise for diagnosis and therapy of cancer patients. Currently, measurement and retrieval of epitope-specific T cells is hampered by limited availability of patients’ biomaterials and lack of sensitive and easy-to-implement T cell priming and expansion. We have developed an in vitro T cell amplification system starting from healthy donor blood and tested different subsets and ratios of autologous T cells and APCs as well as the resting period between amplification cycles. We demonstrated in 10 different donors significantly enhanced frequency of T cells specific for MelanA/HLA-A2, which relied on coculturing of naive T cells and CD11c+ dendritic cells in a 1:1 ratio followed by three weekly amplification cycles using the effluent of the naive T cell sort as APCs, a 24-h rest period prior to every reamplification cycle, and IFN-γ production as a readout for epitope-specific T cells. Using this system, MelanA/HLA-A2–specific T cells were enriched by 200-fold, measuring up to 20–60% of all T cells. We extended this system to enrich NY-ESO-1/HLA-A2– and BMLF-1/HLA-A2–specific T cells, examples of a cancer germline Ag and an oncoviral Ag differing in their ability to bind to HLA-A2 and the presence of specific T cells in the naive and, in case of BMLF-1, also the Ag-experienced repertoire. Collectively, we have developed a sensitive and easy-to-implement in vitro T cell amplification method to enrich epitope-specific T cells that is expected to facilitate research and clinical utility regarding T cell diagnosis and treatments.

Tcells are key players of the adaptive immune system, shaping the immune response toward tumors and viruses, and in some cases toward transplanted organs or the body’s own organs. Identification and isolation of Ag-specific T cells hold diagnostic and therapeutic value. In fact, T cell isolation systems aid target identification, monitoring, and characterization of immune responsiveness, thereby facilitating our understanding of cancer, autoimmune, and infectious diseases (1). Another timely application would be adoptive T cell therapy, which has shown clinical successes especially in hematological malignancies in the context of chimeric Ag receptor T cells (25) and in melanoma, multiple myeloma, and sarcoma in the context of TCRs (69). In addition, virus-specific T cells can provide immunity against infections such CMV in case of postallogeneic transplantation, and chronic viral infections such as hepatitis B virus and hepatitis C virus (10, 11). Moreover, T cells recognizing conserved coronavirus epitopes were demonstrated to associate with milder SARS-CoV-2 infections (12). Besides cancer and virus infections, T cells that recognize specific targets have been validated in preclinical models for lupus and AIDS, and are now being tested in clinical trials (1315). Despite the recognized clinical value, the current identification and isolation of epitope-specific T cell responses rely heavily on patient material, in silico tools, and protocols, which are seriously hampered by limited availability, accuracy, and simplicity (16, 17).

Numerous studies have attempted to isolate and amplify epitope-specific T cells using in vitro cocultures of APCs and T cells. An overview of key studies and their limitations is shown in Table I. The first limitation is limited availability of starting material. A study by Lin et al. (18) used autologous PBMCs from melanoma patients as a source for Ag-specific T cells. Patient PBMCs are difficult to obtain and require specialized ethical approval, which limits their use in daily practice. A second limitation is inefficient amplification of Ag-specific T cells. For example, a study by Theaker et al. (19) used CD3/CD28 Ab-coated beads instead of APCs to induce T cell expansion. The use of such beads, however, can lead to nonspecific proliferation of T cells. A third limitation is insensitive detection of epitope-specific T cell responses. For example, Lin et al. (18) used epitope-bound MHC tetramers to detect T cell responses. Such complexes miss low-avidity T cells, which in fact often dominate immune responses (20).

In this study, we have revisited and optimized an in vitro amplification system of epitope-specific T cells starting from healthy donor PBMCs to address the above-mentioned limitations. This system was set up using a highly immunogenic epitope of the differentiation Ag MelanA, and its application was extended to epitopes of the cancer germline Ag (CGA) NY-ESO-1 and the oncoviral Ag BMLF-1. In this study, we demonstrated that priming naive T cells using an equal number of epitope-pulsed and autologous CD11c-expressing conventional dendritic cells (DCs), followed by minimally one additional enrichment cycle using epitope-pulsed PBMCs, which was preceded by a rest period, enabled detection of low-frequency epitope-specific T cells. Collectively, we demonstrated that using this system we can easily and sensitively detect epitope-specific T cells with different and low precursor frequencies in healthy donor blood.

PBMCs were isolated from healthy human buffy coats (Sanquin, Amsterdam, the Netherlands) by centrifugation via Ficoll-Isopaque (density = 1.077 g/cm3; Amersham Pharmacia Biotech, Uppsala, Sweden), passed through a cell strainer (70 µm), and frozen down either for direct usage as APCs during amplification cycles 2–4 or for isolation of APC subsets and T cell subsets during amplification cycle 1. To isolate CD11c+ DCs, PBMCs were stained for 10 min at 4°C with Fc Block (10 µl/107 PBMCs, BD Pharmingen, Vianen, the Netherlands), after which cells were stained with CD11c-PE Ab (10 µl/107 PBMCs, B-ly6, BD Pharmingen) for 30 min at 4°C, washed, and incubated with PE beads (Miltenyi Biotec, Bergisch Gladbach, Germany, 10 µl/107 PBMCs) for another 15 min at 4°C. After washing, cells were dissolved in MACS buffer and passed through a MACS LS column (Miltenyi Biotec) to positively select CD11c+ cells. To isolate naive or pan T cells from PBMCs, we used the naive T cell (Miltenyi Biotec) and the pan T cell isolation kits (Miltenyi Biotec) according to the manufacturer’s recommendations. Before and after sort, cells were stained either with CD11c-PE (1:50) or with the following combination of markers: 7-aminoactinomycin D (7AAD)-PerCP (1:50, BD Pharmingen), CD3-BV421 (1:50, SP34-2, BD Horizon), CD4-allophycocyanin Cy7 (1:50, RPA-T4, BD Pharmingen), CD8-BV650 (1:100, RPA-T8, BD Horizon), CCR7-PE (1:5, 150503, R&D Systems, Abingdon, U.K.), and CD45RA-BV510 (1:50, HI100, BD Horizon). Sort efficiencies for CD11c+, naive, or pan T cells were assessed by flow cytometry, and were always >60% for the CD11c sorts and >90% for the T cell sorts.

CD11c+ cells and PBMCs were irradiated at 30 Gy to avoid proliferation of NK cells and other non-naive T cells (21) and cultured (2 × 106/ml) in CD11c medium (RPMI 1640 medium supplemented with 1% human serum [Sanquin, Amsterdam, the Netherlands], 200 mM l-glutamine, and 1% antibiotics). Cells were loaded either with MelanA (ELAGIGILTV, 2.5 µg/ml), NY-ESO-1 (SLLMWITQC, 10 µg/ml), or BMLF-1 epitope (GLCTLVAML, 10 µg/ml) (all three epitopes were from ProImmune, Oxford, U.K., dissolved in 100% DMSO) in CD11c medium supplemented with GM-CSF (ImmunoTools, Gladiolenweg, Germany, 10 ng/ml), IL-4 (R&D systems, 10 ng/ml), LPS (Invitrogen, Göteborg, Sweden, 100 ng/ml), and IFN-γ (PeproTech, London, U.K., 10 ng/ml). The epitope-loaded cells were plated in flat-bottom 24-wells plates (0.5 × 106 cells/well) for overnight incubation. Isolated naive or pan T cells were cultured (3 × 106/ml) overnight in T cell medium (RPMI 1640 medium supplemented with 25 mM HEPES, 5% human serum [Sanquin], 200 mM l-glutamine, and 1% antibiotics) supplemented with 5 ng/ml IL-7 (BD Pharmingen). The effluent of the naive T cell sort (effluent T) was collected and frozen down for usage during amplification cycles 2–4.

To optimize in vitro amplification of epitope-specific T cells, we tested the effect of different sources and ratios of APCs and T cells in cycle 1, a resting day between cycles, and different sources of APCs in cycles 2–4 (Fig. 1). Naive or pan T cells were resuspended in T cell medium supplemented with IL-21 (Miltenyi Biotec, 60 ng/ml). Cycle 1 started by coculturing APCs with T cells (0.5 × 106/well each) in 24-well plates, which were cytokine starved for 72 h and supplemented with 5 ng/ml IL-7 and IL-15 (Miltenyi Biotec) at days 4 and 6, and 10 ng/ml IL-7 and IL-15 at day 8. Cycles 2–4 commenced at day 12 following the start of coculture and were or were not preceded by a 24-h rest period, in which IL-7 and IL-15 (5 ng/ml) were refreshed. During these cycles, irradiated, epitope-loaded, and cytokine-supplemented PBMCs or effluent T were added to all wells in a 1:1 ratio relative to the T cells and cocultures were supplemented with IL-7 and IL-15 (10 ng/ml) at day 15. After cycle 2, IL-7 was replaced with IL-2 (360 IU/ml, aldesleukin [Proleukin]; Chiron, Amsterdam, the Netherlands). T cells were sampled from cycles 2 and 4 for testing epitope reactivity. Cell numbers were measured at the end of cycles 2 and 4 and used to calculate fold change in growth relative to the cell number at the start of coculture. For cycle 4, fold change in growth was corrected for cells used for the stimulation assay.

To test abundance of epitope-specific T cells, IFN-γ production upon stimulation with epitope-loaded T2 cells (LCLxT lymphoblastoid hybrid cell line 0.1743CEM.T2) was measured. T2 cells (2 × 106/ml) were loaded with MelanA, NY-ESO-1, or BMLF-1 epitopes (10 µg/ml) for 30 min at 37°C in T cell medium. Non-loaded T2 cells were used as a negative control. T cells (2 × 106/ml) were stimulated with T2 cells in a 1:1 ratio (100 µl each) overnight in a round-bottom 96-wells plate. Supernatants were collected and IFN-γ levels were determined with an ELISA (Invitrogen) according to the manufacturer’s protocol.

HLA-A2 binding of epitopes was performed as previously described by Wu et al. (22) with minor adjustments. In short, 0.15 × 106 T2 cells were incubated in serum-free medium supplemented with β2-microglobulin (3 µg/ml; Sigma-Aldrich, Amsterdam, the Netherlands) and exposed to titrated amounts of epitope ranging from 30 nM to 30 µM for 3 h at 37°C, after which T2 cells were stained with anti–HLA-A2 (1:20, BB7.2, BD Pharmingen) on ice in the dark. The mean fluorescence intensity was measured on the three-laser FACSCelesta flow cytometer (BD Biosciences) using the FACSDiva 8.x software, after which EC50 values were calculated with GraphPad Prism 5.0, with non-epitope–loaded T2 cells representing baselines.

To determine the frequency of MelanA-specific T cells following in vitro amplification, cells were stained with epitope/HLA multimers. To this end, 5 µl of empty loadable HLA tetramers (Tetramer Shop, Kongens Lyngby, Denmark) were incubated with 0.5 µl of epitope (200 µM) for 30 min on ice. HLA tetramer complexes were centrifuged (3300 × g, 5 min) and incubated with 0.5 × 106 T cells for 15 min at 37°C in the dark. Next, an Ab mixture containing CD3-PerCP (1:10, BD Biosciences) and CD8-allophycocyanin (1:80, eBioscience) was added and incubated for 30 min at 4°C in the dark. Finally, T cells were washed twice, fixed with 1% paraformaldehyde, and events were acquired with FACSCelesta (BD Biosciences) and analyzed using FlowJo software.

MelanA-specific CD8+ T cells enriched from two healthy donors were stained with cocktails consisting of Abs targeting markers of T cell maturation, coinhibition, and costimulation and measured on a FACSCelesta (BD Biosciences). The T cell maturation markers panel consisted of MelanA tetramer-PE, CD3-BV786 (1:100, SP34-2, BD Biosciences), CD8-BV650 (1:100, RPA-T8, BD Horizon), CD4-allophycocyanin-Cy7 (1:50, RPA-T4, BD Pharmingen), 7AAD-PerCp (1:50, BD Pharmingen), CD14-PerCp (1:50, MϕP9, BD Biosciences), CD27-allophycocyanin (1:50, L128, BD Biosciences), CD95-PE-Cy7 (1:100, DX2, eBioscience), CD45RA-BV510 (1:50, HI100, BD Biosciences), and CCR7-BV421 (1:50, 150503, BD Biosciences); the coinhibitory receptors panel consisted of MelanA tetramer-PE, CD3-BV421 (1:200, SP34-2, BD Horizon), CD8-BV650 (1:100, RPA-T8, BD Horizon), CD4-V500 (1:50, RPA-T4, BD Biosciences), 7AAD-PerCp (1:50, BD Pharmingen), CD14-PerCp (1:50, MϕP9, BD Biosciences), CD223-PE-Cy7 (LAG3, 1:20, 3DS223H, eBioscience), CD279-allophycocyanin-Cy7 (PD-1, 1:50, EH12.2H7, BioLegend), and CD366-allophycocyanin (TIM-3, 1:50, F38-2E2, BioLegend Europe); and the costimulatory receptors panel consisted of MelanA tetramer-PE, CD3-BV421 (1:200, SP34-2, BD Horizon), CD8-BV650 (1:100, RPA-T8, BD Horizon), CD4-V500 (1:50, RPA-T4, BD), 7AAD-PerCP (1:50, BD Pharmingen), CD14-PerCP (1:50, MϕP9, BD Biosciences), CD137-allophycocyanin (4-1BB, 1:25, 4B4-1, BD Biosciences), CD154-allophycocyanin-Cy7 (CD40L, 1:20, 24-31, BioLegend), and CD278-PE-Cy7 (ICOS, 1:20, ISA-3, eBioscience) (see Supplemental Fig. 2 for gating strategy). Epitope-specific T cells were identified through staining with 5 µl of empty loadable HLA-A2 tetramer (Tetramer Shop, Kongens Lyngby, Denmark) and incubated with 0.5 µl of MelanA epitope (200 µM) for 30 min on ice. The HLA tetramer complex was centrifuged (3300 × g, 5 min) and incubated with 0.5 × 106 T cells for 15 min at 37°C in the dark. Next, an Ab mixture containing CD3-PerCP (1:10, BD Biosciences) and CD8-allophycocyanin (1:80, eBioscience) was added and incubated for 30 min at 4°C in the dark. Finally, T cells were washed twice, fixed with 1% paraformaldehyde, and events were acquired with a FACSCelesta (BD Biosciences) and analyzed using FlowJo software. All Abs for both donors were normalized for background fluorescence using backbone tubes of fluorescence minus one controls.

Statistical analysis was performed using GraphPad Prism 5.0. The Mann–Whitney U test was performed to compare T cell IFN-γ production and growth per parameter. Differences between test parameters were considered statistically significant when p < 0.05. The parameters with the best outcome regarding T cell responses and growth were included in the final in vitro amplification system of epitope-specific T cells. The Wilcoxon signed-rank test was performed to determine specific T cell responses using the final system (*p < 0.05, **p < 0.01, and ***p < 0001).

To assess the importance of naive T cells in our T cell amplification system, we isolated and compared naive T cells versus pan T cells (consisting of all T cell subsets) from blood from healthy donors, and used autologous MelanA epitope-loaded CD11c+ DCs as Ag presenters (Fig. 1, parameter A, naive T cells = six donors, pan T cells = five donors; Table I). We observed early detection of MelanA-specific T cells (as evidenced by epitope-specific T cell IFN-γ production) already after two cycles of amplification when using naive T cells but not pan T cells (p = 0.0031). Also, later detection of MelanA-specific T cells after four cycles of amplification with naive T cells was accompanied by significantly higher IFN-γ production when compared with pan T cells (Fig. 2A, p = 0.019). Furthermore, we observed significantly enhanced amplification of total T cells when using naive versus pan T cells (Fig. 3A, cycle 2 p = 0.0024).

FIGURE 1.

Critical parameters tested to optimize in vitro amplification of epitope-specific T cells. We investigated the optimal source and ratio of autologous T cells and APCs as well as the introduction of a resting period between amplification cycles to obtain optimal detection of epitope-specific T cells. Critical parameters mentioned (in italics) were tested using MelanA, and the final protocol was validated using the oncoviral Ag BMLF-1 and CGA NY-ESO-1. This figure was created with BioRender.

FIGURE 1.

Critical parameters tested to optimize in vitro amplification of epitope-specific T cells. We investigated the optimal source and ratio of autologous T cells and APCs as well as the introduction of a resting period between amplification cycles to obtain optimal detection of epitope-specific T cells. Critical parameters mentioned (in italics) were tested using MelanA, and the final protocol was validated using the oncoviral Ag BMLF-1 and CGA NY-ESO-1. This figure was created with BioRender.

Close modal
FIGURE 2.

Frequency of MelanA-specific T cells affected by subset and relative starting amount of T cells during first cycle and amount of APCs during every amplification cycle. IFN-γ levels of epitope-stimulated T cells after cycles 2 and 4 were measured using the following conditions. (A) Cycle 1: coculture between CD11c+ DCs and naive versus pan T cells (naive T cells = six donors, pan T cells = five donors); cycles 2–4: PBMCs were used as APCs. (B) Cycle 1: coculture between naive T cells and CD11c+ DCs versus PBMCs (CD11c DCs = six donors, PBMCs = three donors); cycles 2–4: PBMCs were used as APCs. (C) Cycle 1: coculture between CD11c+ and naive T cells in a 1:1 versus 1:5 ratio (1:5 ratio = three donors, 1:1 ratio = six donors); cycles 2–4: PBMCs were used as APCs. (D) Cycle 1: coculture between CD11c+ and naive T cells; cycles 2–4: PBMCs versus effluent T cells were used as APCs (effluent T cells = five donors, PBMCs = four donors). (E) Cycle 1: coculture between CD11c+ and naive T cells; cycles 2–4: preceded or not by a 24-h resting period (rest = five donors, no rest = five donors). Each dot represents a single well (two to six wells per donor were used) and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. Statistical analysis was performed with a Mann–Whitney U test (*p < 0.05, **p < 0.01. ns, not significant).

FIGURE 2.

Frequency of MelanA-specific T cells affected by subset and relative starting amount of T cells during first cycle and amount of APCs during every amplification cycle. IFN-γ levels of epitope-stimulated T cells after cycles 2 and 4 were measured using the following conditions. (A) Cycle 1: coculture between CD11c+ DCs and naive versus pan T cells (naive T cells = six donors, pan T cells = five donors); cycles 2–4: PBMCs were used as APCs. (B) Cycle 1: coculture between naive T cells and CD11c+ DCs versus PBMCs (CD11c DCs = six donors, PBMCs = three donors); cycles 2–4: PBMCs were used as APCs. (C) Cycle 1: coculture between CD11c+ and naive T cells in a 1:1 versus 1:5 ratio (1:5 ratio = three donors, 1:1 ratio = six donors); cycles 2–4: PBMCs were used as APCs. (D) Cycle 1: coculture between CD11c+ and naive T cells; cycles 2–4: PBMCs versus effluent T cells were used as APCs (effluent T cells = five donors, PBMCs = four donors). (E) Cycle 1: coculture between CD11c+ and naive T cells; cycles 2–4: preceded or not by a 24-h resting period (rest = five donors, no rest = five donors). Each dot represents a single well (two to six wells per donor were used) and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. Statistical analysis was performed with a Mann–Whitney U test (*p < 0.05, **p < 0.01. ns, not significant).

Close modal
FIGURE 3.

Total number of T cells is affected by subsets and the ratio of T cells and APCs during the first cycle. (AE) Fold change in total number of T cells after cycles (C) 2 and 4 were measured using the same conditions as explained in the legend to (Fig. 2. Number of donors is as follows. (A) Naive T cells = six donors, pan T cells = five donors. (B) CD11c DCs = six donors, PBMCs = three donors. (C) 1:5 ratio = three donors, 1:1 ratio = six donors. (D) Effluent T cells = five donors, PBMCs = four donors. (E) Rest = seven donors, no rest = four donors). Dot plots display fold changes in total number of T cells at cycles 2 and 4 (each relative to cycle 0). Each dot represents a single well (two to six wells per donor were used), and all wells belonging to an identical condition were collectively used for analysis. Statistical analysis was performed with a Mann–Whitney U test (**p < 0.01, ***p < 0001. ns, not significant).

FIGURE 3.

Total number of T cells is affected by subsets and the ratio of T cells and APCs during the first cycle. (AE) Fold change in total number of T cells after cycles (C) 2 and 4 were measured using the same conditions as explained in the legend to (Fig. 2. Number of donors is as follows. (A) Naive T cells = six donors, pan T cells = five donors. (B) CD11c DCs = six donors, PBMCs = three donors. (C) 1:5 ratio = three donors, 1:1 ratio = six donors. (D) Effluent T cells = five donors, PBMCs = four donors. (E) Rest = seven donors, no rest = four donors). Dot plots display fold changes in total number of T cells at cycles 2 and 4 (each relative to cycle 0). Each dot represents a single well (two to six wells per donor were used), and all wells belonging to an identical condition were collectively used for analysis. Statistical analysis was performed with a Mann–Whitney U test (**p < 0.01, ***p < 0001. ns, not significant).

Close modal
Table I.

Overview of key protocols for in vitro amplification of epitope-specific T cells

ReferenceaAgs TestedParameters TestedSource MaterialReadoutsAdvantagesDisadvantages
Butler et al. (40MelanA
NY-ESO-1
Telomerase
Her-2/neu 
Stimulator cells: K562ABCb
Responding cells: CD8+ T cells
Ratio between stimulator and responder: 1:20
Resting day: 24 h between cycles 
Artificial APCs
PBMCs from healthy donors or melanoma patients 
pMHC binding
Chromium- release assay
Secretion of IFN-γ by ELISPOT 
No DCs required
No beads coated with Abs required 
Little amplification of epitope-specific T cells
Limited to HLA- A2–restricted epitopes
Creation of artificial APCs requires specialized materials (41
Lin et al. (18Flu-MP
MelanA
NY-ESO-1 
Stimulator cells: irradiated PBMCs
Responding cells: PBMCs
Ratio between stimulator and responder: 1:1
Resting day: not indicated 
PBMCs from healthy donors (for Flu-MP)
PBMCs from melanoma patients (for MelanA, NY- ESO-1) 
pMHC binding
Intracellular staining for IFN-γ, IL-2, MIP-1β, and TNF-α 
No cell lines or DCs required
No beads coated with Abs required 
Readout according to insensitive pMHC binding (21, 42)
Limited availability of starting material (when using patient PBMCs) 
Wölfl et al. (19MelanA
STEAP1
gp100
CMV pp65 
Stimulator cells: moDCsc
Responding cells: naive CD8+ T cells
Ratio between stimulator and responder: 1:4
Resting day: not indicated 
PBMCs from healthy donors pMHC binding
Intracellular staining for IFN-γ and TNF-α 
No beads coated with Abs required Use of DCs requires high number of PBMCs
moDCs require preculture of 3–7 d, thereby necessitating that naive T cells are isolated from frozen material limiting their yield 
Theaker et al. (20BMLF-1
Flu-MP/Flu-HA
Ebola virus epitopes
Type 1 diabetes epitopes
Cancer Ags 
Stimulator: beads coated with CD3/CD28 Abs
Responding cells: T cells
Ratio between stimulator and responder: 2:1
Resting day: not indicated 
PBMCs from healthy donors
or type 1 diabetes patient 
Secretion of MIP-1β and IFN-γ by ELISA
pMHC staining
Chromium- release assay 
No pMHC multimers required
No cell lines or DCs required
No repeated exposure to Ag required 
Use of CD3/ CD28 beads lead to amplification of nonspecific T cells
Loss of low- frequency T cell clones 
ReferenceaAgs TestedParameters TestedSource MaterialReadoutsAdvantagesDisadvantages
Butler et al. (40MelanA
NY-ESO-1
Telomerase
Her-2/neu 
Stimulator cells: K562ABCb
Responding cells: CD8+ T cells
Ratio between stimulator and responder: 1:20
Resting day: 24 h between cycles 
Artificial APCs
PBMCs from healthy donors or melanoma patients 
pMHC binding
Chromium- release assay
Secretion of IFN-γ by ELISPOT 
No DCs required
No beads coated with Abs required 
Little amplification of epitope-specific T cells
Limited to HLA- A2–restricted epitopes
Creation of artificial APCs requires specialized materials (41
Lin et al. (18Flu-MP
MelanA
NY-ESO-1 
Stimulator cells: irradiated PBMCs
Responding cells: PBMCs
Ratio between stimulator and responder: 1:1
Resting day: not indicated 
PBMCs from healthy donors (for Flu-MP)
PBMCs from melanoma patients (for MelanA, NY- ESO-1) 
pMHC binding
Intracellular staining for IFN-γ, IL-2, MIP-1β, and TNF-α 
No cell lines or DCs required
No beads coated with Abs required 
Readout according to insensitive pMHC binding (21, 42)
Limited availability of starting material (when using patient PBMCs) 
Wölfl et al. (19MelanA
STEAP1
gp100
CMV pp65 
Stimulator cells: moDCsc
Responding cells: naive CD8+ T cells
Ratio between stimulator and responder: 1:4
Resting day: not indicated 
PBMCs from healthy donors pMHC binding
Intracellular staining for IFN-γ and TNF-α 
No beads coated with Abs required Use of DCs requires high number of PBMCs
moDCs require preculture of 3–7 d, thereby necessitating that naive T cells are isolated from frozen material limiting their yield 
Theaker et al. (20BMLF-1
Flu-MP/Flu-HA
Ebola virus epitopes
Type 1 diabetes epitopes
Cancer Ags 
Stimulator: beads coated with CD3/CD28 Abs
Responding cells: T cells
Ratio between stimulator and responder: 2:1
Resting day: not indicated 
PBMCs from healthy donors
or type 1 diabetes patient 
Secretion of MIP-1β and IFN-γ by ELISA
pMHC staining
Chromium- release assay 
No pMHC multimers required
No cell lines or DCs required
No repeated exposure to Ag required 
Use of CD3/ CD28 beads lead to amplification of nonspecific T cells
Loss of low- frequency T cell clones 

moDC, monocyte-derived DC.

a

Protocols follow the principle of T cell priming as illustrated in (Fig. 1.

b

K562ABC is a human leukemic cell line expressing transgenes for HLA class I, CD54, CD58, CD80, and CD83.

c

moDCs were cultured for 72 h in medium supplemented with GM-CSF and IL-4 following culture in medium supplemented with GM-CSF, IL-4, LPS, and IFN-γ.

Next, we tested the source of APCs to optimally prime MelanA-specific T cells (Fig. 1, parameter B). To this end, we loaded CD11c+ DCs versus PBMCs with the MelanA epitope during the first cycle (i.e., priming cycle) of autologous naive T cells, and compared their ability to induce MelanA-specific T cell responses (CD11c DCs = six donors, PBMCs = three donors). In both cases, we used epitope-loaded PBMCs during the subsequent cycles (i.e., amplification cycles). MelanA-specific T cell responses were already detectable after two cycles when using CD11c+ DCs during priming, and consistently showed higher IFN-γ production when comparing CD11c+ DCs versus PBMCs (Fig. 2B, cycle 2 p = 0.0019, cycle 4 p = 0.023). Again, we also observed enhanced amplification of total T cells when CD11c+ DCs versus PBMCs were used (Fig. 3B, cycle 2 p = 0.0002). Furthermore, we tested different ratios (1:1, 1:5, 1:10, and 1:20) between CD11c+ DCs and naive T cells (Fig. 1, parameter C, 1:5 ratio = three donors, 1:1 ratio = six donors, 1:10 ratio = two donors, and 1:20 ratio = two donors). We observed that using the 1:5 ratio (i.e., lowering the relative presence of CD11c+ DCs during T cell priming) reduced MelanA-specific IFN-γ production (Fig. 2C, cycle 2 p = 0.0074, cycle 4 p = 0.0086) as well as total T cell amplification (Fig. 3C, cycle 2 p < 0.0001, cycle 4 p = 0.0006). The ratios of 1:10 and 1:20 resulted in no T cell amplification (data not shown), thereby excluding assessment of MelanA-specific T cell IFN-γ production.

For cycles 2–4 (i.e., amplification cycles), we tested whether the effluent from the naive T cell sort (effluent T; containing all immune cell populations normally present in PBMCs except for naive T cells) could replace PBMCs as the source of APCs (Fig. 1, parameter D). Because PBMCs are also used as the starting material for both the CD11c+ DC and naive T cell sort, the use of the above effluent would maximize the number of test epitopes from a single donor. To this end, MelanA-loaded CD11c+ DCs were cocultured with naive T cells and amplified for two to four cycles using effluent T versus PBMCs (effluent T cells = five donors, PBMCs = four donors). We observed no significant differences in IFN-γ production and T cell amplification in both conditions (Figs. 2D, 3D, IFN-γ production: cycle 4 p = 0.74; T cell amplification: cycle 2 p = 0.97, cycle 4 p = 0.47). This similarity between effluent T cells and PBMCs warrants the use of effluent T cells in our system.

Finally, we tested whether introducing a 24-h rest period for T cells prior to starting a subsequent amplification cycle further enhances detection of MelanA-specific T cells (Fig. 1, parameter E). We used MelanA-loaded CD11c+ DCs to prime naive T cells, and amplified T cells using PBMCs either in the presence or absence of a 24-h rest prior to every amplification cycle (resting day = five donors, no resting day = five donors). We observed a nonsignificant trend of increased T cell IFN-γ production upon stimulation with MelanA when applying the 24-h rest condition (Fig. 2E, cycle 2 p = 0.07, cycle 4 p = 0.26), whereas amplification of total T cells was not affected (Fig. 3E, resting day = seven donors, no resting day = four donors, cycle 2 p = 0.8, cycle 4 p = 0.21).

When implementing all optimal parameters into a single system (Fig. 4A), that is, coculture of MelanA-loaded CD11c+ DCs and naive T cells in a 1:1 ratio (cycle 1), after which T cells were amplified using MelanA-loaded effluent T cells (cycles 2–4), with a 24-h rest period prior to every cycle, we observed enhanced MelanA/HLA-A2–specific IFN-γ production (Fig. 4B, p < 0.0001) and 2-fold enrichment of total T cells in 10 donors (Fig. 4C). The epitope specificity of T cells was confirmed with stainings for MelanA/HLA-A2 multimers before and after enrichment, and we observed a 200-fold increase in the frequency of MelanA-specific T cells (unpaired Student t test, p = 0.0022, seven donors, (Fig. 4D, 4E).

FIGURE 4.

Detection and characterization of blood T cells specific for MelanA when applying final in vitro amplification system. (A) Timeline of the protocol to enrich epitope specific T cells. (B) IFN-γ levels of MelanA- or no peptide (NoP)–stimulated T cells after cycle 4 of T cell amplification with MelanA epitope; n = 10 donors (two to six wells per donor). Each dot represents a single well, and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. (C) Fold enrichment of MelanA-specific T cells after cycles 2 and 4 (relative to cycle 0). (D) Frequency of MelanA-specific T cells pre- and postenrichment according to % of epitope-MHC-positive CD8+ T cells; n = 7 donors; one well per donor. (E) Representative scatter plot depicting frequencies of MelanA-specific T cells pre- and postenrichment. (F) Stacked bar plot representing different maturation status (Tem, effector memory T cell; Temra, terminally differentiated effector memory T cell; Tm, memory T cell; Tnaive, naive T cell) of MelanA-specific CD8+ T cells after cycles 0, 2, and 4 of enrichment (n = 2 donors; one well per donor). (G) Dot plots representing changes in frequency of different MelanA-specific CD8+ T cell phenotypes after cycles 0, 2, and 4 of enrichment for two donors, one well per donor. Statistical analysis was performed with the Wilcoxon signed-rank test (B and C) and with an unpaired Student t test (D) (**p < 0.01, ***p < 0001. ns, not significant).

FIGURE 4.

Detection and characterization of blood T cells specific for MelanA when applying final in vitro amplification system. (A) Timeline of the protocol to enrich epitope specific T cells. (B) IFN-γ levels of MelanA- or no peptide (NoP)–stimulated T cells after cycle 4 of T cell amplification with MelanA epitope; n = 10 donors (two to six wells per donor). Each dot represents a single well, and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. (C) Fold enrichment of MelanA-specific T cells after cycles 2 and 4 (relative to cycle 0). (D) Frequency of MelanA-specific T cells pre- and postenrichment according to % of epitope-MHC-positive CD8+ T cells; n = 7 donors; one well per donor. (E) Representative scatter plot depicting frequencies of MelanA-specific T cells pre- and postenrichment. (F) Stacked bar plot representing different maturation status (Tem, effector memory T cell; Temra, terminally differentiated effector memory T cell; Tm, memory T cell; Tnaive, naive T cell) of MelanA-specific CD8+ T cells after cycles 0, 2, and 4 of enrichment (n = 2 donors; one well per donor). (G) Dot plots representing changes in frequency of different MelanA-specific CD8+ T cell phenotypes after cycles 0, 2, and 4 of enrichment for two donors, one well per donor. Statistical analysis was performed with the Wilcoxon signed-rank test (B and C) and with an unpaired Student t test (D) (**p < 0.01, ***p < 0001. ns, not significant).

Close modal

When characterizing the phenotype of MelanA-specific T cells (see Materials and Methods and Supplemental Table I for details and Supplemental Fig. 2A–C for gating strategies), we observed a shift from naive T cells (61–77%, CD8+CCR7+CD45RA+) toward the effector memory phenotype (80–85%, CD8+CD45RACCR7), which started at cycle 2 and continued in cycle 4 (90–92%, two donors, (Fig. 4F, Supplemental Fig. 2D). Besides T cell maturation, we evaluated expression of the checkpoint molecules PD1, LAG3, and TIM3, as well as the costimulatory receptors 4-1BB, ICOS and CD40L and observed a 20-fold increase in LAG3+, 13- to 60-fold increase of TIM3+, and 62-fold increase of LAG3+TIM3+ expressing MelanA-specific CD8+ T cells after two cycles of enrichment, which remained at cycle 4 (two donors). In contrast, there was a 28- to 47-fold increase in CD40L expressing MelanA-specific CD8+ T cells at cycle 2 (Fig. 4G). Expression of PD1, 41BB, or ICOS was not affected (Supplemental Fig. 2D).

Finally, we extended this T cell amplification system to two additional epitopes, namely HLA-A2–binding epitopes from the CGA NY-ESO-1 and the oncoviral EBV-derived Ag BMLF-1 (Fig. 1). Notably, these epitopes display a 100-fold difference regarding their binding affinity toward HLA-A2, with NY-ESO-1 being a strong binder (EC50 = 4.6E−06, SD = 1.93E−06), comparable to MelanA (EC50 = 2.2E−06, SD = 8.87E−07), and BMLF-1 being a weak binder (EC50 = 1.2E−04, SD = 0.0E−0.0) (Fig. 5A). With the T cell amplification system, we observed a 3.5- and 2.2-fold enrichment of total T cells (Fig. 5C, 5E) and enhanced epitope-specific IFN-γ production (Fig. 5B, 5D, NYESO p = 0.01, BMLF-1 p = 0.0078) when enriched for NY-ESO-1/HLA-A2 and BMLF-1/HLA-A2 epitopes in four out of five and three out of three donors, respectively.

FIGURE 5.

Detection of blood T cells specific for NY-ESO-1 or BMLF-1 when applying final in vitro amplification system. (A) HLA-A2 binding of MelanA (ELA), NY-ESO-1 (SLL), and BMLF-1 (GLC) epitopes using titrated amounts of epitopes (31 nM to 31 μM); representative titration curves are shown. EC50 values were calculated using GraphPad Prism 5 (n = 3). (B) IFN-γ levels of NY-ESO-1– or NoP-stimulated T cells after cycle 4 of T cell amplification with the NY-ESO-1 epitope; n = 5 donors. (C) Fold enrichment of NY-ESO-1–specific T cells after cycles 2 and 4 (relative to cycle 0). (D) IFN-γ levels of BMLF-1 or NoP-stimulated T cells after cycle 4 of T cell amplification with the BMLF-1 epitope; n = 3 donors. (E) Fold enrichment of BMLF-1–specific T cells after cycles 2 and 4 (relative to cycle 0). Each dot represents a single well (two to six wells per donor were used), and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. Statistical analysis was performed with the Wilcoxon signed-rank test (*p < 0.05, **p < 0.01. ns, not significant).

FIGURE 5.

Detection of blood T cells specific for NY-ESO-1 or BMLF-1 when applying final in vitro amplification system. (A) HLA-A2 binding of MelanA (ELA), NY-ESO-1 (SLL), and BMLF-1 (GLC) epitopes using titrated amounts of epitopes (31 nM to 31 μM); representative titration curves are shown. EC50 values were calculated using GraphPad Prism 5 (n = 3). (B) IFN-γ levels of NY-ESO-1– or NoP-stimulated T cells after cycle 4 of T cell amplification with the NY-ESO-1 epitope; n = 5 donors. (C) Fold enrichment of NY-ESO-1–specific T cells after cycles 2 and 4 (relative to cycle 0). (D) IFN-γ levels of BMLF-1 or NoP-stimulated T cells after cycle 4 of T cell amplification with the BMLF-1 epitope; n = 3 donors. (E) Fold enrichment of BMLF-1–specific T cells after cycles 2 and 4 (relative to cycle 0). Each dot represents a single well (two to six wells per donor were used), and all wells belonging to an identical condition were collectively used for analysis. See Materials and Methods for details. Statistical analysis was performed with the Wilcoxon signed-rank test (*p < 0.05, **p < 0.01. ns, not significant).

Close modal

To investigate the robustness of the optimized parameters, we tested our protocol in an Ag-experienced setting. To that effect, we enriched BMLF-1–specific T cells using buffy coats of healthy donors that successfully resolved an EBV infection (EBV resolvers, two donors) and compared them with EBV-naive healthy donors (nonresolvers, three donors). We observed no significant differences in epitope-specific IFN-γ production by T cells from EBV resolvers compared with nonresolvers (Supplemental Fig. 1A, 1B). In EBV resolvers, we also compared naive and pan T cells (which could contain BMLF-1–specific memory T cells) cocultured with BMLF-1–loaded CD11c+ DCs. We observed higher BMLF-1–specific IFN-γ production after four cycles of enrichment while using naive T cells as the T cell source in EBV resolvers compared with pan T cells (Supplemental Fig. 1C), but total numbers of enriched T cells did not differ (Supplemental Fig. 1D).

In this study, we have revisited and validated an in vitro T cell amplification system to sensitively and easily detect epitope-specific T cells with low and variable precursor frequencies in blood. Key parameters we identified include: 1) equal numbers of autologous, naive T cells and CD11c+ DCs during T cell priming; 2) non-naive T cells as APCs during T cell boosting; 3) a rest period preceding each amplification cycle; and 4) IFN-γ production as a readout for epitope-specific T cell frequencies.

With our final in vitro T cell amplification system, we were able to achieve a 200-fold enrichment of MelanA (ELA)-specific T cells. This amplification system yielded 20–60% of CD8+ T cells mainly comprised of an effector memory phenotype that bound the MelanA/HLA-A2 multimers, where only 0.02–0.1% of CD8+ T cells bound these multimers at the start of the amplification. We validated our system with other epitopes and demonstrated clear detection of NY-ESO-1– as well as BMFL-1–specific T cell responses in multiple donors. NY-ESO-1 is a member of the family of CGAs, and it induces humoral and cellular immune responses in cancer patients (6, 7, 23, 24), whereas BMLF-1 is the product of an early lytic EBV gene that induces clear EBV-specific T cell responses (25). The detection of T cells specific for these epitopes in healthy blood is noteworthy, as HLA-A2 has contrasting binding avidity for the NY-ESO-1 (SLL) and BMLF-1 (GLC) epitope, and the precursor frequency was not recorded for the NY-ESO-1 (SLL)–specific T cells whereas high frequency (0.04–0.47% of CD8+ T cells) for BMLF-1 (GLC) –specific T cells was recorded, respectively (26, 27). We also observed that our amplification system with the selected parameters could perform in both an Ag-experienced and Ag-naive setting. These outcomes point toward the robustness of our in vitro system to detect epitope-specific T cells in blood.

Regarding the required materials, we chose healthy donors as the source for PBMCs due to their relatively easy accessibility and chose an autologous setting to minimize the risk of off-target responses (i.e., yielding false-negative detections). The frequency of epitope-specific T cells in healthy donor blood is estimated to be between 0.6 and 500 cells per million CD8+ T cells (28, 29), justifying the various cycles of T cell amplification to obtain sufficient numbers of T cells for detection. The MelanA epitope, recognized for its high immunogenicity (30) and high precursor frequency of specific CD8+ T cells in blood of melanoma patients and healthy donors (31), provided our reference epitope to set up the in vitro T cell amplification system. The investigation of the optimal choice of APC source, T cell source, and their relative ratio for epitope-specific T cell enrichment in vitro is a prerequisite for seting up the coculture system. Almost all immune cells can act as APCs but they vary in their ability to (cross)present epitopes and provide costimulatory support needed for optimal T cell priming (32). In our setting particularly, the introduction of epitope-loaded CD11c+ DCs and naive T cells in a 1:1 ratio during the first cycle enabled detection of low frequencies of epitope-specific T cells according to epitope-induced production of IFN-γ. A possible explanation is that pan T cells comprise regulatory T cells that inhibit epitope-specific T cells; this in contrast to naive T cells that result in potent epitope-specific T cells to familiar and unfamiliar epitopes. In addition, priming of naive T cells by CD11c+ DCs induces stronger and earlier responses compared with PBMCs. This is in line with CD11c+ DCs being reported as specialized Ag presenters to naive T cells offering superior costimulatory support and epitope presentation (33, 34), whereas PBMCs harbor a mixture of different types of immune cells, thereby diluting the number of specialized Ag presenters. Indeed, when reducing the relative presence of CD11c+ DCs compared with naive T cells, we observed a decrease in the frequency of epitope-specific T cells and lesser T cell growth. Besides the use of CD11c+ DCs and naive T cells, also three subsequent amplification cycles using epitope-loaded effluent T cells, as well as a 24-h rest period prior to every cycle, contributed to the detection of low frequencies of epitope-specific T cells. The use of effluent T cells (the run-through after MACSort of naive T cells) maximized the use of PBMCs to sort CD11c+ DCs and naive T cells, and the testing of more epitopes per donor. The beneficial effect of a resting day prior to every restimulation cycle is in line with a study by Kutscher et al. (35) who showed that a resting day enhanced the functionality and Ag sensitivity of epitope-specific T cells. Furthermore, although formally not tested, the use of both CD4+ and CD8+ T cells in this system is considered advantageous, as multiple studies have shown that CD4+ T cells help potentiate CD8+ T cell responses (36). Interestingly, we observed no significant enhancement in IFN-γ production or growth of total T cells in our setting after cycle 2; additionally, the phenotype (i.e., maturation status and expression of cosignaling molecules) of epitope-specific T cells remained stable after cycle 2, which supports our rationale of using 2–4 enrichment cycles to obtain epitope-specific T cells. Increasing the number of amplification cycles beyond four leads to a decline in T cell proliferation, likely caused by overactivation and anergy of T cells (data not shown). In line with this, we observed increased frequencies of coinhibitory molecules LAG3, TIM3, and LAG3+ TIM3+ or costimulatory molecule CD40L expressing MelanA-specific CD8+ T cells already after two cycles of amplification, which remain unchanged in cycle 4. One could argue to make use of checkpoint blockade Abs targeting LAG3 and TIM3 in the cocultures to improve the fitness and potentially longevity of these epitope-specific T cells. However, the effects of these Abs require a thorough investigation in the future.

Finally, based on all of our observations, we recommend the use of at least four wells per epitope, a minimum of three donors, and a minimum of two stimulation cycles, with the latter establishing a minimum assay time of 20 d.

Even though we have established a working in vitro system to detect low frequencies of epitope-specific T cells, our study currently shows limitations. First, the list of parameters we have tested is not exhaustive. Other parameters, such as the starting cell numbers per well, different DC subsets, or cytokine combinations, may further improve the sensitivity of detection. Second, some studies have shown that high-dose (>2 Gy) irradiation of DCs could dampen T cell proliferation (37). A lower dose of irradiation could be considered for further studies in our coculture system. Third, there may be epitopes, particularly those with very low binding strength toward HLA-A2, that do not yield a detectable frequency of T cells. Although we cannot exclude this limitation, recent studies have indicated that factors modulating the ability of an epitope to induce a T cell response range beyond its binding avidity toward HLA, and include the precursor frequency of epitope-specific T cells in the naive repertoire (38). Along this line, a study by Pogorely et al. (39) demonstrated a higher prevalence of T cells specific for common pathogens such as EBV and influenza in the umbilical cord blood versus PBMCs. In the case of a very weak HLA binder, or in the case of other physical or biochemical properties of an epitope that adversely affect its immunogenicity, the efficacy of our protocol might be questionable. Third, the presented system may work particularly well for HLA-A2 and has not yet been tested for other HLA alleles.

In conclusion, we present an in vitro T cell amplification system to sensitively and easily detect epitope-specific T cells with low precursor frequencies in blood. This T cell amplification system may be used to determine the frequency of T cells for a given epitope that relates to the immunogenicity of the epitope. The T cell amplification system can be used to assess and compare the immunogenicity of putative T cell epitopes, and these immunogenicity scores may facilitate vaccine developments. Another practical applicability of this system relies on the retrieval of the epitope-specific T cells themselves. This enables isolation and characterization of the corresponding TCRs, which, when they pass the necessary assays for efficacy and safety, hold great promise for further development of therapies against tumors and chronic infections. Along the same lines, the T cell amplification system may be used to identify TCRs that can aid the monitoring of epitope-specific T cells in blood or tissues in disease settings (using peptide/MHC complexes or anti-TCR Abs, which can be used in combination with markers for certain T cell subsets), thereby potentially enhancing our understanding of the contribution of epitope-specific T cells to the pathology of certain diseases.

In short, our coculture system presents a valuable tool for preclinical as well as diagnostic and therapeutic applications for different T cell–related studies.

This work was supported by the Health Holland Public-Private Partnership Grant EMC-TKI LSH20020.

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • 7AAD

    7-aminoactinomycin D

  •  
  • CGA

    cancer germline Ag

  •  
  • DC

    dendritic cell

1.
Simon
P.
,
T. A.
Omokoko
,
A.
Breitkreuz
,
L.
Hebich
,
S.
Kreiter
,
S.
Attig
,
A.
Konur
,
C. M.
Britten
,
C.
Paret
,
K.
Dhaene
, et al
2014
.
Functional TCR retrieval from single antigen-specific human T cells reveals multiple novel epitopes.
Cancer Immunol. Res.
2
:
1230
1244
.
2.
Grupp
S. A.
,
M.
Kalos
,
D.
Barrett
,
R.
Aplenc
,
D. L.
Porter
,
S. R.
Rheingold
,
D. T.
Teachey
,
A.
Chew
,
B.
Hauck
,
J. F.
Wright
, et al
2013
.
Chimeric antigen receptor-modified T cells for acute lymphoid leukemia.
N. Engl. J. Med.
368
:
1509
1518
.
3.
Kalos
M.
,
B. L.
Levine
,
D. L.
Porter
,
S.
Katz
,
S. A.
Grupp
,
A.
Bagg
,
C. H.
June
.
2011
.
T cells with chimeric antigen receptors have potent antitumor effects and can establish memory in patients with advanced leukemia.
Sci. Transl. Med.
3
:
95ra73
.
4.
Brentjens
R. J.
,
M. L.
Davila
,
I.
Riviere
,
J.
Park
,
X.
Wang
,
L. G.
Cowell
,
S.
Bartido
,
J.
Stefanski
,
C.
Taylor
,
M.
Olszewska
, et al
2013
.
CD19-targeted T cells rapidly induce molecular remissions in adults with chemotherapy-refractory acute lymphoblastic leukemia.
Sci. Transl. Med.
5
:
177ra38
.
5.
Kochenderfer
J. N.
,
M. E.
Dudley
,
R. O.
Carpenter
,
S. H.
Kassim
,
J. J.
Rose
,
W. G.
Telford
,
F. T.
Hakim
,
D. C.
Halverson
,
D. H.
Fowler
,
N. M.
Hardy
, et al
2013
.
Donor-derived CD19-targeted T cells cause regression of malignancy persisting after allogeneic hematopoietic stem cell transplantation.
Blood
122
:
4129
4139
.
6.
Robbins
P. F.
,
R. A.
Morgan
,
S. A.
Feldman
,
J. C.
Yang
,
R. M.
Sherry
,
M. E.
Dudley
,
J. R.
Wunderlich
,
A. V.
Nahvi
,
L. J.
Helman
,
C. L.
Mackall
, et al
2011
.
Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with NY-ESO-1.
J. Clin. Oncol.
29
:
917
924
.
7.
Robbins
P. F.
,
S. H.
Kassim
,
T. L. N.
Tran
,
J. S.
Crystal
,
R. A.
Morgan
,
S. A.
Feldman
,
J. C.
Yang
,
M. E.
Dudley
,
J. R.
Wunderlich
,
R. M.
Sherry
, et al
2015
.
A pilot trial using lymphocytes genetically engineered with an NY-ESO-1-reactive T-cell receptor: long-term follow-up and correlates with response.
Clin. Cancer Res.
21
:
1019
1027
.
8.
Rapoport
A. P.
,
E. A.
Stadtmauer
,
G. K.
Binder-Scholl
,
O.
Goloubeva
,
D. T.
Vogl
,
S. F.
Lacey
,
A. Z.
Badros
,
A.
Garfall
,
B.
Weiss
,
J.
Finklestein
, et al
2015
.
NY-ESO-1-specific TCR-engineered T cells mediate sustained antigen-specific antitumor effects in myeloma.
Nat. Med.
21
:
914
921
.
9.
D’Angelo
S. P.
,
L.
Melchiori
,
M. S.
Merchant
,
D.
Bernstein
,
J.
Glod
,
R.
Kaplan
,
S.
Grupp
,
W. D.
Tap
,
K.
Chagin
,
G. K.
Binder
, et al
2018
.
Antitumor activity associated with prolonged persistence of adoptively transferred NY-ESO-1c259T cells in synovial sarcoma.
Cancer Discov.
8
:
944
957
.
10.
Kaeuferle
T.
,
R.
Krauss
,
F.
Blaeschke
,
S.
Willier
,
T.
Feuchtinger
.
2019
.
Strategies of adoptive T -cell transfer to treat refractory viral infections post allogeneic stem cell transplantation.
J. Hematol. Oncol.
12
:
13
.
11.
Maldini
C. R.
,
G. I.
Ellis
,
J. L.
Riley
.
2018
.
CAR T cells for infection, autoimmunity and allotransplantation.
Nat. Rev. Immunol.
18
:
605
616
.
12.
Mallajosyula
V.
,
C.
Ganjavi
,
S.
Chakraborty
,
A. M.
McSween
,
A. J.
Pavlovitch-Bedzyk
,
J.
Wilhelmy
,
A.
Nau
,
M.
Manohar
,
K. C.
Nadeau
,
M. M.
Davis
.
2021
.
CD8+ T cells specific for conserved coronavirus epitopes correlate with milder disease in COVID-19 patients.
Sci. Immunol.
6
:
eabg5669
.
13.
Lam
S.
,
C.
Bollard
.
2013
.
T-cell therapies for HIV.
Immunotherapy
5
:
407
414
.
14.
Kuhlmann
A. S.
,
C. W.
Peterson
,
H. P.
Kiem
.
2018
.
Chimeric antigen receptor T-cell approaches to HIV cure.
Curr. Opin. HIV AIDS
13
:
446
453
.
15.
Jin
X.
,
Q.
Xu
,
C.
Pu
,
K.
Zhu
,
C.
Lu
,
Y.
Jiang
,
L.
Xiao
,
Y.
Han
,
L.
Lu
.
2021
.
Therapeutic efficacy of anti-CD19 CAR-T cells in a mouse model of systemic lupus erythematosus.
Cell. Mol. Immunol.
18
:
1896
1903
.
16.
Peters
B.
,
M.
Nielsen
,
A.
Sette
.
2020
.
T cell epitope predictions.
Annu. Rev. Immunol.
38
:
123
145
.
17.
De Simone
M.
,
G.
Rossetti
,
M.
Pagani
.
2018
.
Single cell T cell receptor sequencing: techniques and future challenges.
Front. Immunol.
9
:
1638
.
18.
Lin
Y.
,
H. F.
Gallardo
,
G. Y.
Ku
,
H.
Li
,
G.
Manukian
,
T. S.
Rasalan
,
Y.
Xu
,
S. L.
Terzulli
,
L. J.
Old
,
J. P.
Allison
, et al
2009
.
Optimization and validation of a robust human T-cell culture method for monitoring phenotypic and polyfunctional antigen-specific CD4 and CD8 T-cell responses.
Cytotherapy
11
:
912
922
.
19.
Theaker
S. M.
,
C.
Rius
,
A.
Greenshields-Watson
,
A.
Lloyd
,
A.
Trimby
,
A.
Fuller
,
J. J.
Miles
,
D. K.
Cole
,
M.
Peakman
,
A. K.
Sewell
,
G.
Dolton
.
2016
.
T-cell libraries allow simple parallel generation of multiple peptide-specific human T-cell clones.
J. Immunol. Methods
430
:
43
50
.
20.
Rius
C.
,
M.
Attaf
,
K.
Tungatt
,
V.
Bianchi
,
M.
Legut
,
A.
Bovay
,
M.
Donia
,
P.
Thor Straten
,
M.
Peakman
,
I. M.
Svane
, et al
2018
.
Peptide-MHC class I tetramers can fail to detect relevant functional T cell clonotypes and underestimate antigen-reactive t cell populations.
J. Immunol.
200
:
2263
2279
.
21.
Wölfl
M.
,
P. D.
Greenberg
.
2014
.
Antigen-specific activation and cytokine-facilitated expansion of naive, human CD8+ T cells.
Nat. Protoc.
9
:
950
966
.
22.
Wu
Z. Y.
,
Y. F.
Gao
,
Y. H.
Wu
,
W.
Liu
,
M.
Sun
,
M. X.
Zhai
,
Y. M.
Qi
,
Y.
Ye
.
2011
.
Identification of a novel CD8+ T cell epitope derived from cancer-testis antigen MAGE-4 in oesophageal carcinoma.
Scand. J. Immunol.
74
:
561
567
.
23.
Kunert
A.
,
M.
van Brakel
,
S.
van Steenbergen-Langeveld
,
M.
da Silva
,
P. G.
Coulie
,
C.
Lamers
,
S.
Sleijfer
,
R.
Debets
.
2016
.
MAGE-C2-specific TCRs combined with epigenetic drug-enhanced antigenicity yield robust and tumor-selective T cell responses.
J. Immunol.
197
:
2541
2552
.
24.
Ramachandran
I.
,
D. E.
Lowther
,
R.
Dryer-Minnerly
,
R.
Wang
,
S.
Fayngerts
,
D.
Nunez
,
G.
Betts
,
N.
Bath
,
A. J.
Tipping
,
L.
Melchiori
, et al
2019
.
Systemic and local immunity following adoptive transfer of NY-ESO-1 SPEAR T cells in synovial sarcoma.
J. Immunother. Cancer
7
:
276
.
25.
Selin
L.
,
L.
Watkin
,
K.
Luzuriaga
,
N.
Aslan
.
2012
.
Severity of Epstein-Barr virus (EBV)-induced infectious mononucleosis (IM) correlates with the frequency of crossreactive influenza A virus-M1 and EBV-BMLF-1-specific CD8 T cells.
J. Immunol.
188
(
1 Suppl
):
105.48
.
26.
Berner
B. R.
,
M.
Tary-Lehmann
,
N. L.
Yonkers
,
A. D.
Askari
,
P. V.
Lehmann
,
D. D.
Anthony
.
2005
.
Phenotypic and functional analysis of EBV-specific memory CD8 cells in SLE.
Cell. Immunol.
235
:
29
38
.
27.
Sukdolak
C.
,
S.
Tischer
,
D.
Dieks
,
C.
Figueiredo
,
L.
Goudeva
,
H. G.
Heuft
,
M.
Verboom
,
S.
Immenschuh
,
A.
Heim
,
S.
Borchers
, et al
2013
.
CMV-, EBV- and ADV-specific T cell immunity: screening and monitoring of potential third-party donors to improve post-transplantation outcome.
Biol. Blood Marrow Transplant.
19
:
1480
1492
.
28.
Alanio
C.
,
F.
Lemaitre
,
H. K. W.
Law
,
M.
Hasan
,
M. L.
Albert
.
2010
.
Enumeration of human antigen-specific naive CD8+ T cells reveals conserved precursor frequencies.
Blood
115
:
3718
3725
.
29.
Legoux
F.
,
E.
Debeaupuis
,
K.
Echasserieau
,
H.
De La Salle
,
X.
Saulquin
,
M.
Bonneville
.
2010
.
Impact of TCR reactivity and HLA phenotype on naive CD8 T cell frequency in humans.
J. Immunol.
184
:
6731
6738
.
30.
Valmori
D.
1997
.
Enhanced generation of specific tumor-reactive CTL in vitro by Melan-A/Mart-1 immunodominant peptide analogs.
Immunol. Lett.
56
:
223
224
.
31.
Pittet
M. J.
,
D.
Valmori
,
P. R.
Dunbar
,
D. E.
Speiser
,
D.
Liénard
,
F.
Lejeune
,
K.
Fleischhauer
,
V.
Cerundolo
,
J. C.
Cerottini
,
P.
Romero
.
1999
.
High frequencies of naive Melan-A/MART-1-specific CD8+ T cells in a large proportion of human histocompatibility leukocyte antigen (HLA)-A2 individuals.
J. Exp. Med.
190
:
705
715
.
32.
Berard
M.
,
D. F.
Tough
.
2002
.
Qualitative differences between naïve and memory T cells.
Immunology
106
:
127
138
.
33.
Osugi
Y.
,
S.
Vuckovic
,
D. N. J.
Hart
.
2002
.
Myeloid blood CD11c+ dendritic cells and monocyte-derived dendritic cells differ in their ability to stimulate T lymphocytes.
Blood
100
:
2858
2866
.
34.
Merad
M.
,
P.
Sathe
,
J.
Helft
,
J.
Miller
,
A.
Mortha
.
2013
.
The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting.
Annu. Rev. Immunol.
31
:
563
604
.
35.
Kutscher
S.
,
C. J.
Dembek
,
S.
Deckert
,
C.
Russo
,
N.
Körber
,
J. R.
Bogner
,
F.
Geisler
,
A.
Umgelter
,
M.
Neuenhahn
,
J.
Albrecht
, et al
2013
.
Overnight resting of PBMC changes functional signatures of antigen specific T- cell responses: impact for immune monitoring within clinical trials.
PLoS One
8
:
e76215
.
36.
Shedlock
D. J.
,
H.
Shen
.
2003
.
Requirement for CD4 T cell help in generating functional CD8 T cell memory.
Science
300
:
337
339
.
37.
Cao
M. D.
,
B. G.
Xiao
.
2003
.
Influence of gamma irradiation on phenotype and function of human dendritic cells in vitro.
Zhongguo Shi Yan Xue Ye Xue Za Zhi
11
:
282
286
.
38.
Doytchinova
I. A.
,
D. R.
Flower
.
2006
.
Class I T-cell epitope prediction: improvements using a combination of proteasome cleavage, TAP affinity, and MHC binding.
Mol. Immunol.
43
:
2037
2044
.
39.
Pogorelyy
M. V.
,
A. D.
Fedorova
,
J. E.
McLaren
,
K.
Ladell
,
D. V.
Bagaev
,
A. V.
Eliseev
,
A. I.
Mikelov
,
A. E.
Koneva
,
I. V.
Zvyagin
,
D. A.
Price
, et al
2018
.
Exploring the pre-immune landscape of antigen-specific T cells.
Genome Med.
10
:
68
.
40.
Butler
M. O.
,
J. S.
Lee
,
S.
Ansén
,
D.
Neuberg
,
F. S.
Hodi
,
A. P.
Murray
,
L.
Drury
,
A.
Berezovskaya
,
R. C.
Mulligan
,
L. M.
Nadler
,
N.
Hirano
.
2007
.
Long-lived antitumor CD8+ lymphocytes for adoptive therapy generated using an artificial antigen-presenting cell.
Clin. Cancer Res.
13
:
1857
1867
.
41.
Zhang
D. K. Y.
,
A. S.
Cheung
,
D. J.
Mooney
.
2020
.
Activation and expansion of human T cells using artificial antigen-presenting cell scaffolds.
Nat. Protoc.
15
:
773
798
.
42.
Dolton
G.
,
E.
Zervoudi
,
C.
Rius
,
A.
Wall
,
H. L.
Thomas
,
A.
Fuller
,
L.
Yeo
,
M.
Legut
,
S.
Wheeler
,
M.
Attaf
,
D. M.
Chudakov
,
E.
Choy
,
M.
Peakman
,
A. K.
Sewell
.
2018
.
Optimized Peptide–MHC Multimer Protocols for Detection and Isolation of Autoimmune T-Cells.
Front. Immunol.
9
:
1378
.

The authors have no financial conflicts of interest.

Supplementary data