The systematic assessment of the human immune system bears huge potential to guide rational development of novel immunotherapies and clinical decision making. Multiple assays to monitor the quantity, phenotype, and function of Ag-specific T cells are commonly used to unravel patients’ immune signatures in various disease settings and during therapeutic interventions. When compared with tests measuring soluble analytes, cellular immune assays have a higher variation, which is a major technical factor limiting their broad adoption in clinical immunology. The key solution may arise from continuous control of assay performance using TCR-engineered reference samples. We developed a simple, stable, robust, and scalable technology to generate reference samples that contain defined numbers of functional Ag-specific T cells. First, we show that RNA-engineered lymphocytes, equipped with selected TCRs, can repetitively deliver functional readouts of a controlled size across multiple assay platforms. We further describe a concept for the application of TCR-engineered reference samples to keep assay performance within or across institutions under tight control. Finally, we provide evidence that these novel control reagents can sensitively detect assay variation resulting from typical sources of error, such as low cell quality, loss of reagent stability, suboptimal hardware settings, or inaccurate gating.

A plethora of immune assays to study immune cell subsets on a single-cell level are increasingly applied in clinical immunotherapy trials in infectious diseases, autoimmunity, oncology, and gene therapy (15). Immune assay portfolios can consist of simple monoparametric assays that have been used for >20 y (6), up to cutting edge technologies allowing assessment of multiple functions on a single-cell level using classical flow cytometry (7) or mass cytometry (8). Also, the parallel detection of Ag-specific T cells using combinatorial labeling with MHC–peptide multimers in mass cytometry has recently been introduced and represents useful tools for the comprehensive analysis of T cell responses (9, 10).

Cellular immune assays have an intrinsic complexity that results in an increased variability compared with that of assays quantifying soluble analytes. On the one hand, this has been the major stigma preventing the broad adoption of cellular assays in clinical settings. In contrast, the observed challenges have led to coordinated efforts to standardize immunophenotyping and harmonize immune assays prior to using them in clinical settings (1114). The latter efforts have led to the identification of critical variables that impact on assay performance. In addition, consensus has been reached on the notion that the key for state-of-the-art immunomonitoring lies in the control of the critical process steps, the controlled use of standardized assays, and the implementation of appropriate assay controls (13, 15).

The gap between the urgent need to govern immune assay performance and the lack of appropriate assay controls motivated the development of a new, affordable, easy, versatile, robust, and scalable technology for controlled manufacturing of TCR-engineered reference samples (TERS). We applied molecularly and chemically modified RNA to enable efficient transfer of TCR chain pairs into primary lymphocytes. The engineered Ag-specific T cells were spiked into unmodified PBMC at defined numbers and reproducibly led to Ag-specific signals in three commonly used T cell assays, namely, staining with MHC-peptide multimers, cytokine flow cytometry (CFC), and IFN-γ ELISPOT assay. We provide evidence that TERS 1) lead to physiological assay signals that are comparable to endogenous Ag-specific T cells; 2) are stable over a period of at least 12 mo; 3) work in the hands of different investigators using different locally available assay protocols; 4) can be thawed and applied repetitively to control the performance of a series of experiments (application phase); and 5) sensitively capture deviations in assays due to low-quality cell materials, use of reagents with expired shelf life, unintended change of cytometer settings, and in appropriate data analysis. We believe that the technology is fit for purpose to deliver tight control of immune assay performance and thus is ready to support broad use of T cell assays in clinical immunology.

The human HLA-A2–restricted TCR specific for the wild-type (wt) epitope NY-ESO-1157–165 (SLLMWITQC) was provided by T. Schumacher (the Netherlands Cancer Institute, Amsterdam, the Netherlands). The human HLA-A2–restricted TCR specific for the wt epitope tyrosinase368–376 (YMDGTMSQV) was cloned out of the tyrosinase-specific CTL clone IVSB, which was provided by T. Wölfel (Johannes Gutenberg University, Mainz, Germany). The human wt NY-ESO-197–111–specific, HLA-B7–restricted TCR and the human wt NY-ESO-1165–179–specific, HLA-DRB0401–restricted TCR specific for the epitopes ATPMEAELARRSLAQ or CFLPVFLAQPPSGQR, respectively, were extracted from lymphocytes derived from cancer patients following informed consent. Tumor-associated Ag (TAA)–specific CD4+ or CD8+ T cells were expanded and detected in a CD154 (BD, Franklin Lakes, NJ) staining assay or IFN-γ secretion assay (Miltenyi Biotec, Bergisch Gladbach, Germany), respectively, after specific peptide stimulation. For TCR cloning, labeled TAA-specific T cells were single-cell sorted (BD FACSAria II cell sorter). Total RNA was extracted and purified from each single cell using the RNeasy Micro Kit (Qiagen, Hilden, Germany) and transcribed into cDNA via RT-PCR using SuperScript II Reverse Transcriptase (Invitrogen/Life Technologies, Carlsbad, CA). Specific primers encoding the TCR variable and constant domain of the α- or the β-chain were used to select and to amplify Ag-specific TCR cDNA via PCR. Each TCR chain was then cloned into a pST1 expression vector for DNA large-scale production. Finally, the TCR DNA was linearized using EciI (New England Biolabs, Ipswich, U.K.) and purified by ethanol precipitation for the in vitro transcription of RNA (Ambion, Life Technologies).

For the generation of chimeric TCRs, the human wt TCRs were murinized. The combined DNA template of each TCR containing a human variable and a murine constant domain was ordered by GeneArt (Life Technologies, Regensburg, Germany). The chimeric construct was cloned into a modified pST1 expression vector with introduced SmaI (New England Biolabs) and SfoI (New England Biolabs) restriction sites. TCR cloning and RNA production were performed as described above.

For the generation of MHC class I TERS, PBMCs from consenting buffy coat donors (Transfusion Center, University Medical Center, Mainz, Germany) were isolated by Ficoll density gradient centrifugation using Ficoll–Paque Plus (GE Healthcare, Uppsala, Sweden). A total of 20 × 106 PBMCs was electroporated with 30 μg selected TCR-α and 30 μg TCR-β D2-capped IVT-RNA (in-house produced product) in one shot. Each transfection was performed in a total volume of 250 μl BTXpress High Performance Electroporation Solution (BTX/Harvard Apparatus, Holliston, MA) in 0.4-mm BTX cuvettes (BTX/Harvard Apparatus) using the ECM 830 Square Wave Electroporation System (BTX/Harvard Apparatus) set to 500 V, 3 ms, 1 pulse. A total of 10 × 106 of the transfected cells was transferred and cultured in one well of a six-well culture plate (BD) containing 5 ml culture medium (RPMI 1640 + Glutamax; Gibco/Life Technologies), 5% human serum albumin (Lonza, Basel, Switzerland), 1% MEM Non-Essential Amino Acids Solution (Gibco/Life Technologies), 1% Na4P2O7 (Life Technologies/Life Technologies), 0.5% Pen/Strep (Life Technologies), and 1 μg/ml DNase I (Roche, Basel, Switzerland). The cells were then incubated for 18–20 h at 37°C and 5% CO2. The nontransfected PBMCs were divided into 50-ml falcon tubes and stored at room temperature (RT) in a slanted position in a final volume of 35 ml culture medium and at a final concentration of 6–8 × 106 cells/ml.

The TCR-transfected cells (positive fraction) and the untouched PBMCs (negative fraction) were harvested, and a portion was stained with ViaCount Reagent (Merck Millipore, Darmstadt, Germany) to check the cell viability and recovery. The analysis was performed using the Guava EasyCyte5HT (Merck Millipore). In parallel, the transfection efficiency was determined using MHC–peptide multimers to analyze the percentage of TCR+ cells of all CD8+ T cells. Based on these data, the dilution for each specific batch was defined to reach an optimal number of TCR-positive cells following the spiking into untouched PBMCs. TERS incorporating ∼1–2% MHC–peptide multimer+ CD8+ cells were generated for use in the two functional assays, IFN-γ ELISPOT or CFC, and TERS incorporating ∼0.5% MHC–peptide multimer+ CD8+ cells were generated for subsequent MHC–peptide multimer-staining experiments. The diluted cells (TERS batch), as well as a fraction of the negative fraction, were frozen under controlled conditions using the SY-LAB_14S-B freezing machine (SY-LAB, Neupurkersdorf, Austria) and stored in liquid nitrogen. Per aliquot, 1 × 107 cells (for ELISPOT and cytokine staining) or 5 × 106 cells (for MHC–peptide multimer tests) were frozen in aliquots at a final concentration of 1 × 107/ml in freshly prepared freezing medium containing RPMI 1640 + Glutamax, 10% human serum albumin (CSL Behring, King of Prussia, PA), and 10% DMSO (AppliChem, Darmstadt, Germany). For the generation of a HLA class II–restricted TERS, a small fraction of PBMCs from a buffy coat donor was electroporated with the TAA-specific TCR RNA. In parallel, the naive cell fraction was transfected with RNA encoding the restricting HLA chains (DRB04). Then a defined number of the TCR-specific PBMCs was spiked into the HLA-transfected cells.

Single aliquots were subsequently thawed and subjected to quality control experiments (MHC–peptide multimer staining) and functional testing (ELISPOT and CFC). The aliquots were directly thawed in a 37°C water bath and then carefully transferred into 10 ml prewarmed thawing medium containing RPMI 1640 + Glutamax and 10% human serum albumin. After centrifugation at 300 × g for 8 min at RT, the cells were washed in 5 ml prewarmed 1× PBS (Gibco/Life Technologies) and tested for cell viability and recovery.

MHC–peptide multimer staining was carried out using either MHC–peptide tetramers (Beckman Coulter, Brea, CA) or MHC–peptide dextramers (Immudex, Copenhagen, Denmark). Most experiments with MHC–peptide multimers were performed with MHC–peptide dextramers (Immudex), except for one experiment shown in Supplemental Fig. 1 in which MHC–peptide tetramers (Beckman Coulter) were applied. For the staining, 1 × 106 cells were transferred to a well of a 96-well round-bottom plate (Costar, Amsterdam, the Netherlands). After centrifugation at 500 × g for 5 min at 4°C, the MHC–peptide multimer staining was performed in a total volume of 50 μl staining buffer containing 1× PBS, 5% FCS (Invitrogen/Life Technologies), and 5 mM EDTA (Fluka/Sigma-Aldrich). The MHC–peptide multimer mix was prepared by adding 45 μl MHC–peptide multimer solution containing 1× PBS, 50% FCS, 2 mM EDTA, and 5 μl Ag-specific MHC–peptide multimer for each reaction. The MHC–peptide multimer mix was centrifuged at 13,000 × g for 5 min at 4°C. The supernatant was then used to stain the cells for 30 min at 4°C in the dark when using tetramers or for 10 min at RT in the dark when using dextramers. After washing the cells with 200 μl staining buffer, viability staining using 7-aminoactinomycin D (7-AAD; A07704; Beckman Coulter) was performed in combination with staining of CD8 allophycocyanin (clone SK1; BD Biosciences, San Jose, CA) or PE (clone HIT8a; BD Biosciences) and CD4 FITC (clone RPA-T4; BD Biosciences) mAb in a total volume of 50 μl per reaction and incubated for 20 min at 4°C in the dark. After washing in 200 μl staining buffer, the cells were resuspended in 150 μl staining buffer for flow cytometry acquisition with a Canto II (BD Biosciences). All flow cytometry data FCS files were analyzed with FlowJo software (Tree Star, San Carlos, CA) and will be made available upon request. The cells were first gated to exclude doublets (forward light scatter [FSC]-W versus FSC-A) and dead cells (7-AAD versus FSC-A). The living cell gate (7-AAD–negative population) was used to gate on the lymphocytes (side light scatter-A and FSC-A). Live lymphocytes were tested for their expression of CD4 and CD8 (CD4+ versus CD8+). CD8+CD4 cells were used for gating of Dext+ of CD8+ or Tet+ of CD8+ cells.

A total of 1 × 106 cells of thawed TERS aliquots was transferred for a single stimulation in a well of a 96-well round-bottom culture plate with a final volume of 200 μl in X-vivo15 (Lonza) containing 1 μg/ml DNase (Roche). NY-ESO-1157–165–specific, HLA-A2–restricted TERS were stimulated with the peptide SLLMWITQC, NY-ESO-197-111–specific, HLA-B7–restricted TERS with the peptide ATPMEAELARRSLAQ, NY-ESO-1165-179–specific, HLA-DRB*0401–restricted TERS with the peptide CFLPVFLAQPPSGQR and tyrosinase368-375-specific, HLA-A2–restricted TERS with the peptide YMDGTMSQV each at a final concentration of 10 mM. A total of 2 μg/ml staphylococcal enterotoxin B (SEB; Sigma-Aldrich, St. Louis, MO) was used as a positive control. We also used irrelevant peptide stimulation with the particular nonbinding, unspecific peptide (JPT Peptide Technologies, Berlin, Germany), with a purity <80% and at a final concentration of 10 mM as well as a medium control as negative controls. During the peptide stimulation, 1/1000 Golgi Stop (BD) and 1/1500 Golgi Plug (BD) were added to inhibit the Golgi apparatus, and also CD107a Ab as a marker of degranulation. After an incubation at 37°C at 5% CO2 for 6 h, the cells were centrifuged, washed with 200 μl 1× PBS, and stained with 0.5 μl/test Blue Fluorescent Reactive Dye (Invitrogen/Life Technologies) in 200 μl precooled 1× PBS for 30 min at 4°C for dead cell exclusion. After centrifugation, we added CD8a eFluor650NC (clone RPA-T8; eBioscience, San Diego, CA) and CD4 PerCP-eFluor710 (clone SK3, eBioscience) in a total volume of 50 μl with staining buffer containing 1× PBS, 5% FCS, and 0.5% EDTA for 20 min at 4°C in the dark. After a washing step, the cells were permeabilized in 100 μl/test Cytofix/Cytoperm (BD) for 20 min at 4°C. After incubation, the cells were washed in 150 μl 1× Perm/Wash buffer (BD Biosciences) and stained for cytokine production. The intracellular cytokine staining with TNF-α Pacific Blue (clone Mab-b11; BioLegend), IFN-γ PE-Cy7 (clone B27; BD Biosciences), and IL-2 PE (clone MQ1-17H12; BD, Franklin Lakes, NJ) was carried out in 50 μl 1× Perm/Wash buffer and incubated for 30 min at 4°C in the dark. Then the cells were washed three times with 200 μl 1× Perm/Wash buffer and resuspended in 150 μl 1× Perm/Wash buffer per test for flow cytometry acquisition on the LSR Fortessa (BD Biosciences). Analysis was performed, as previously described above. Finally, CD8+CD4 or CD4+CD8 cells were gated against TNF-α, IFN-γ, IL-2, or CD107a.

ELISPOT plates (Merck Millipore) were coated with 50 μl anti-human IFN-γ Ab (Mabtech; clone 1-D1K) diluted at 1/1300 in PBS and stored overnight at RT. Plates were washed three times with 150 μl PBS and blocked with X-vivo15 (Lonza) and 2% human serum albumin (CSL Behring) for 4 h at 37°C and 5% CO2. Between 100,000 and 300,000 cells (depending on the frequency of specific TCR+ cells in the TERS) were plated in triplicates to the wells of the 96-well ELISPOT plate in 100 μl OpTmizer T Cell Expansion medium including supplement (Invitrogen/Life Technologies). Ag-specific T cells were stimulated with 3 μM relevant or irrelevant peptide in a final volume of 200 μl for 20 h at 37°C and 5% CO2. NY-ESO-1 A2–restricted TERS were stimulated with SLLMWITQC (157–165), NY-ESO-1 DRB*0401–restricted TERS with CFLPVFLAQPPSGQR (165–179), NY-ESO-1 B7–restricted TERS with ATPMEAELARRSLAQ (97–111), tyrosinase A2–restricted TERS with YMDGTMSQV (368–375), and SEB as positive control. For ELISPOT development, plates were washed once with 200 μl PBS containing 0.05% Tween, once with 200 μl double distilled water, and five times in PBS/0.05% Tween. Then the biotinylated IFN-γ Ab (Mabtech, Nacka, Sweden) with a final concentration of 1 μg/ml in 60 μl PBS/0.5% BSA (Sigma-Aldrich) was added to each well of the plate and incubated for 2 h at 37°C and 5% CO2. After washing the plates six times with 200 μl PBS/0.05% Tween, 100 μl avidin-alkaline phosphatase (Sigma-Aldrich) diluted 1/1000 with PBS/0.5% BSA was added per well and incubated for 1 h at RT. After washing the plates three times with 200 μl PBS/0.05% Tween and three times with 200 μl PBS, 100 μl substrate 5-bromo-4-chloro-3-indolyl phosphate/NBT (Sigma-Aldrich) was added per well. For one ELISPOT plate, one 5-bromo-4-chloro-3-indolyl phosphate/NBT tablet was dissolved in 10 ml ddH2O. After 8 min of incubation (protected from light), the reaction was stopped by washing the wells under running water for 1 min. The plate was dried overnight at RT and analyzed on ImmunoSpot Analyzer (CTL, Bonn, Germany).

The main goal of the proficiency panel was to study whether the favorable results generated in Mainz could be confirmed by a group of experienced, external investigators using different laboratory-specific MHC–peptide multimer-staining protocols.

Twelve laboratories from five European countries (Germany, U.K., Denmark, France, and the Netherlands) as well as the United States participated in this proficiency panel. All laboratories were experienced users of MHC–peptide multimer technology. Each of them received two sets of TERS (T1 and T2) prepared from PBMCs of a HLA-A*0201–positive healthy donor. HLA-A*0201–restricted NY-ESO-1–specific TERS were generated using a murinized CD8 T cell–derived TCR. One set consisted of three vials, each containing 5 × 106 cells shipped on dry ice (for European laboratories) or in liquid nitrogen (for the U.S. laboratory). Each of the vials in a given set contained cells from three subbatches (SB) of TERS, namely a SB with a high frequency of ∼15% NY-ESO-1–specific T cells (SB#14.1), a SB with a medium frequency of ∼7.5% NY-ESO-1–specific T cells (SB#14.2), and a SB lacking NY-ESO-1–specific T cells (negative fraction only). Distributed sets of TERS were manufactured and quality controlled in Mainz. All participants also received two MHC–peptide dextramers: an irrelevant dextramer (described as negative dextramer) and a NY-ESO-1157–165 dextramer that were shipped separately at 4°C by the manufacturer Immudex (Copenhagen, Denmark). For all required experiments, laboratories were allowed to use their locally established staining protocol with two limitations for the sake of better comparability. The first mandatory parameter was to use 10 μl dextramer and incubate for 10 min at RT. The second was that the staining cocktails should include at least one CD8 Ab, one CD4 Ab, and a dead cell dye. The reason for asking the laboratories for these three markers was to reach a better comparability of TERS performance across laboratories. All laboratories were allowed to add additional markers (e.g., CD3 Ab) in the staining and to perform the analysis of data (gating strategy) depending on their local preferences.

After completion of the experiments, participants had to complete a questionnaire containing basic information regarding the cell quality (counting method, cell recovery, and viability) and the staining and analyzing procedure; a report form with the resulting dot plots; and a second report form for the analyzed data sets (number of lymphocytes, number of CD8+, number of CD8+ Dext+, and percentage of Dext+ of CD8+ cells).

In a second proficiency panel, TERS were applied in CFC by experienced investigators to assess how the TERS technology performs in a functional assay that is more complex than a MHC–peptide multimer staining and thus more susceptible to variation in a multicenter setting. Five laboratories applied centrally provided TERS in a cytokine-staining experiment at two different time points. For this proficiency panel, we generated a TERS batch containing a defined moderate number of tyrosinase368–376-specific CD8+ T cells. For this, a defined cell number of the TCR-engineered fraction of lymphocytes was mixed with autologous PBMCs in a ratio of 1:7 to obtain a defined frequency of Ag-specific CD8+ T cells. The participants received vials of the central TERS shipped on dry ice containing 5 × 106 cells/vial. Two vials had to be thawed for each of the two experimental tasks (T1 and T2). They also received three vials of a tyrosinase368–376 peptide shipped on dry ice. The peptide was dissolved in DMSO and diluted with PBS to 3 mM.

For each experiment, two vials of cells were thawed and stimulated with the tyrsosinase368–376-specific peptide, an irrelevant peptide (NY-ESO-1157–165), and a positive control reagent (SEB). The stimulation and staining procedure was performed after locally established protocols on two independent time points (T1 and T2). The experiments only had to follow two mandatory requirements. First, participants had to use 1 × 106 cells per test, and second, the staining cocktails had to include at least Abs against CD4, CD8, IFN-γ, and TNF-α and a dead cell dye. Optionally, each laboratory added additional markers (e.g., TNF-α, CD107a, CD137, CD3 Ab, etc.) based on their locally established staining panel. The analysis of data (gating strategy) was performed following local preferences, and final frequencies percentage of marker+ of CD8+ were reported.

The primary goal of the study was to develop a technology enabling the fast generation of multiple stable cell aliquots, containing defined numbers of fully functional Ag-specific T cells, covering a broad range of various CD8+ and CD4+ T cell specificities relevant for clinical applications. RNA was chosen as the nucleic acid format to efficiently equip primary lymphocytes with selected TCRs, due to the fast expression of encoded Ags and the favorable safety profile as compared with viral vectors for gene transfer. To reach a physiological level of exogenous TCR expression on the TCR-transduced lymphocytes over a time period of at least 48 h, we introduced molecularly and chemically engineered RNA for the gene transfer. A strategy to overcome the inhibitory interference of exogenous TCR chains with the endogenous TCR repertoire leading to mispaired chimeric TCRs was to murinize the human wt TCR chains using murine TCR constant sequences. To further increase TCR-RNA t1/2 and stability, the 5-prime end cap was modified from the commercially available m7G(5′)ppp(5′)G (m7) cap structure to a chemically modified m (2)(7,2′-O)Gpp(S)pG phosphorothioate (D1) cap analog (16). Engineered RNA gene vectors coding for the TCR α- and β-chains were then electroporated into freshly isolated PBMCs, and we monitored the cell surface expression of correctly paired TCR chains by using MHC–peptide dextramers over 72 h. Fig. 1A shows that the percentage of HLA-A*0201–restricted tyrosinase-specific lymphocytes following transfer of murinized TCR chains is massively increased compared with the wt TCR sequence, with maximum values of up to 40.3% tyrosinase-dextramer+ of CD8+ compared with 1.98% tyrosinase-dextramer+ of CD8+ in the wt. Chemically stabilized RNA equipped with the D1 cap increased the fraction of TCR-positive CD8+ lymphocytes up to 60.6%. As shown in the representative dot plots, the use of modified RNA did not only increase the fraction of Ag-positive T cells, but also markedly enhanced the TCR expression level per cell. Assuming that the area under curve is a fair estimate of the amount of correctly expressed TCR chains on electroporated lymphocytes, the introduction of molecular and chemical RNA engineering achieved a 70-fold increase in TCR expression, bringing TCR levels up to the physiological levels needed to support Ag-specific TCR signaling and subsequent cytokine secretion of activated T cells. Analogous experiments were performed with m7- and D1-capped chimeric NY-ESO-1–specific TCRs (data not shown). The use of a murinized D1-capped NY-ESO-1 TCR increased the transfection efficiency from 81 to 93% TCR+ cells of CD8+ cells (area under the curve for m7: 5314.57 and for D1: 6329.37), which confirms that optimization of RNA chemistry can substantially increase TCR expression, at least in the case of these two different TCRs.

FIGURE 1.

(A) TCR expression following molecular and chemical engineering of RNA-TCRs. Cell surface expression of TCRs following RNA electroporation was assessed using MHC–peptide multimer staining. The graph shows TCR surface expression for the tyrosinase-specific wt TCR using the m7 (open diamonds)- or D1-cap structure (closed diamonds) and the chimeric TCR with the murine constant domain capped using the m7 (open circles)- or D1-cap structure (closed circles). Transfected cells were cultured at 37°C for up to 72 h. After 4, 8, 12, 16, 18, 20, 48, and 72 h, a fraction of electroporated cells was stained with a tyrosinase368–376 dextramer. The inserted table indicates the area under the curve (AUC) for all four tested conditions. The dot plots show the obtained raw flow data at the time point before TCR expression (T0) and of maximum expression (Tmax). (B) Concept for use of TERS to control immune assay performance. The concept for use is divided into three parts (manufacturing, cutoff definition, and application). First, manufacturing of TERS is performed using the optimized and standardized processes. In this phase, MHC–peptide multimers are used to test the TCR surface expression on CD8+ cells. Once all aliquots of the TERS batch are filled and frozen, quality control (QC) is performed to confirm adequate performance in the selected readout assay. QC controlled and released aliquots are transferred to long-term storage in liquid nitrogen. Second, five aliquots are thawed and tested subsequently in independent experiments to define the expected median signal size and acceptable range (set at 20% above and below the median) for the specific assay protocol. Following the cutoff definition phase, TERS batches can then be applied to control immune assay performance over different time points. The concept for application foresees one test prior to and one test after a test campaign with sample specimens from a clinical trial (test items). Failure of the initial or final TERS to generate a signal within the acceptable range should trigger a root cause analysis.

FIGURE 1.

(A) TCR expression following molecular and chemical engineering of RNA-TCRs. Cell surface expression of TCRs following RNA electroporation was assessed using MHC–peptide multimer staining. The graph shows TCR surface expression for the tyrosinase-specific wt TCR using the m7 (open diamonds)- or D1-cap structure (closed diamonds) and the chimeric TCR with the murine constant domain capped using the m7 (open circles)- or D1-cap structure (closed circles). Transfected cells were cultured at 37°C for up to 72 h. After 4, 8, 12, 16, 18, 20, 48, and 72 h, a fraction of electroporated cells was stained with a tyrosinase368–376 dextramer. The inserted table indicates the area under the curve (AUC) for all four tested conditions. The dot plots show the obtained raw flow data at the time point before TCR expression (T0) and of maximum expression (Tmax). (B) Concept for use of TERS to control immune assay performance. The concept for use is divided into three parts (manufacturing, cutoff definition, and application). First, manufacturing of TERS is performed using the optimized and standardized processes. In this phase, MHC–peptide multimers are used to test the TCR surface expression on CD8+ cells. Once all aliquots of the TERS batch are filled and frozen, quality control (QC) is performed to confirm adequate performance in the selected readout assay. QC controlled and released aliquots are transferred to long-term storage in liquid nitrogen. Second, five aliquots are thawed and tested subsequently in independent experiments to define the expected median signal size and acceptable range (set at 20% above and below the median) for the specific assay protocol. Following the cutoff definition phase, TERS batches can then be applied to control immune assay performance over different time points. The concept for application foresees one test prior to and one test after a test campaign with sample specimens from a clinical trial (test items). Failure of the initial or final TERS to generate a signal within the acceptable range should trigger a root cause analysis.

Close modal

Following the molecular and chemical optimization of the TCR RNA gene vector, we systematically developed an optimized manufacturing process for TERS, in which a fraction of lymphocytes is electroporated with a given TCR and cultured until a high level of TCR surface expression is reached. In parallel, a negative (nonelectroporated) fraction of autologous PBMCs is stored until the controlled spike-in of the RNA-engineered lymphocytes can be performed (described in 2Materials and Methods). We optimized the process to support maximum RNA transfer and cell viability, resulting in the best possible functionality of electroporated lymphocytes, namely the ability to secrete cytokines following Ag-specific T cell activation. In summary, we optimized eight process steps reaching beyond the molecular (1) and chemical (2) engineering of the TCR RNA. We further optimized the electroporation settings (3), the RNA concentration (4), the cell number for gene transfer (5), and the electroporation medium (6). Finally, we studied the TCR expression kinetics (7) to define the optimal time point for cell harvesting and established optimal storage conditions of the untouched, negative cell fraction (8). The stepwise optimization of all process steps finally resulted in transfection efficiencies of between 60 and 95% depending on the TCR specificity and molecular structure, along with associated cell viabilities of between 90 and 95%. The final optimal conditions of the manufacturing process with validity for the used electroporation system are indicated in 2Materials and Methods and were applied in all further experiments.

Parallel to the development of an optimized manufacturing process, we delineated a three-part concept for the day-to-day use of TERS to control immune assay performance (Fig. 1B). First, following manufacturing and successful (assay-format specific) quality control for each TERS batch, the released TERS (n = 50–200) aliquots are transferred to a working cell bank. The second step is to define the expected signal size and acceptable range for a given immune assay. We propose to thaw five aliquots from the working cell bank and apply them in the cutoff definition phase. Notably, the resulting ranges for acceptable assay performance strictly apply to a defined assay protocol and the whole defined experimental setup (including the hardware for data acquisition and strategy for analysis of resulting raw data). Lastly, in the application phase, TERS will come to use prior to assessing the test samples (initial TERS control sample) as well as after the last test sample (finishing TERS control sample), whereas the test sample will in most cases be a patient sample specimen from a clinical trial. If the initial TERS sample leads to a result out of the accepted range (failure), the subsequent experiments with the precious test samples should be put on hold. Failed results should lead to a root cause analysis and corrective action to fix the problem. In the case that the initial TERS sample passes, the experiments with the test samples are initiated. Provided that the finishing TERS sample also passes, all results from the test can be released for further use. Whenever the finishing TERS sample fails, this should again lead to a root cause analysis and resolution of problems prior to considering releasing the experimental data from the test samples. Examples from the TERS cutoff definition phase and examples of applied TERS control samples that fail to deliver results falling in the acceptable range are shown below.

Once a robust and standardized manufacturing process was established, we tested how the TERS perform in the three most commonly applied T cell assays, namely MHC–peptide multimer staining, CFC, and IFN-γ ELISPOT assay. In initial experiments, we focused on the signals obtained with the engineered SB following staining with their cognate MHC–peptide multimers or stimulation with the relevant peptides.

We first generated and stored two spiked TERS SB (SB#1.1 and SB#1.2) of HLA-A2–restricted NY-ESO-1157–165–specific TERS containing different percentages of NY-ESO-1–specific T cells. Aliquots from the two SB were subsequently thawed and stained with dextramers refolded with the NY-ESO-1157–165 epitope. SB#1.1 showed a clearly detectable and strong staining of 28.9% Dext+ of CD8+ T cells (Fig. 2A). SB#1.2 showed a staining level of 11.6% Dext+ of CD8+ T cells. Non-TCR–transfected autologous cells (negative fraction) were also stained with the specific dextramer, showing a low background staining of 0.089% Dext+ of CD8+. Representative dot plots for the two SB, as well as of the negative fraction are shown for one time point in Fig. 2. Several aliquots of both SB were independently thawed and serially tested on one day, or on different days, to study the intra- and interassay variation. The TERS were characterized by a low interassay coefficient of variation (CV) of 6.4% for SB#1.1 and 6.1% for SB#1.2. The intra-assay variation as inferred from the triplicate measurement of the 2-mo stored sample was 3.2% for SB#1.1 and 0.7% for SB#1.2. In addition, we found ∼2.3-fold more dextramer-positive cells in the diluted (1:3) SB#1.1 than in the more highly diluted (1:7) SB#1.2, suggesting that linear dilution is feasible. Representative dot plots for NY-ESO-1157–165–specific SB#1.2, as well as for a similar series of MHC–peptide tetramer stainings with tyrosinase368–376-specific SB#2 and the corresponding negative fractions are shown in Supplemental Fig. 1.

FIGURE 2.

Proof-of-concept experiments for TERS. (A) MHC–peptide multimer staining experiments. Overview of the results obtained for a NY-ESO-1157–165–specific, HLA-A2– restricted TERS tested at two dilutions (SB#1.1, closed circles; SB#1.2, open circles). For both TERS SB, we thawed either one, two, or three aliquots at months 3, 1, and 2, respectively. Each circle represents the result from one single independent experiment. Dot plots show the dextramer signals for both TERS dilutions tested at month 1 as well as the results generated by staining of the negative cell fraction. (B) Cytokine flow cytometry. The figure shows the assay variability for NY-ESO-1165–179–specific HLA-DRB1*0401–restricted TERS (SB#3) tested for production of IL-2 (open diamonds), CD107a (closed diamonds), IFN-γ (open circles), and TNF-α (closed circles) in a series of independent experiments. The x-axis shows the number of months between manufacturing of the TERS and testing. The dot plots show the signals for each of the four markers tested at month 2 following stimulation with an irrelevant peptide stimulus or the cognate Ag NY-ESO-1. (C) IFN-γ ELISPOT. Two NY-ESO-197–111–specific HLA-B7–restricted TERS SB at two different dilutions (SB#5.1, closed circles; SB#5.2, open circles) were tested in IFN-γ ELISPOT in independent tests. The CVs were 12.4% for SB#5.1 and 17.0% for SB#5.2. Representative scans of the ELISPOT filters from wells stimulated with an irrelevant control peptide (left panel) or the cognate peptide (right panel) are shown for time point 5 after 5 mo of storage.

FIGURE 2.

Proof-of-concept experiments for TERS. (A) MHC–peptide multimer staining experiments. Overview of the results obtained for a NY-ESO-1157–165–specific, HLA-A2– restricted TERS tested at two dilutions (SB#1.1, closed circles; SB#1.2, open circles). For both TERS SB, we thawed either one, two, or three aliquots at months 3, 1, and 2, respectively. Each circle represents the result from one single independent experiment. Dot plots show the dextramer signals for both TERS dilutions tested at month 1 as well as the results generated by staining of the negative cell fraction. (B) Cytokine flow cytometry. The figure shows the assay variability for NY-ESO-1165–179–specific HLA-DRB1*0401–restricted TERS (SB#3) tested for production of IL-2 (open diamonds), CD107a (closed diamonds), IFN-γ (open circles), and TNF-α (closed circles) in a series of independent experiments. The x-axis shows the number of months between manufacturing of the TERS and testing. The dot plots show the signals for each of the four markers tested at month 2 following stimulation with an irrelevant peptide stimulus or the cognate Ag NY-ESO-1. (C) IFN-γ ELISPOT. Two NY-ESO-197–111–specific HLA-B7–restricted TERS SB at two different dilutions (SB#5.1, closed circles; SB#5.2, open circles) were tested in IFN-γ ELISPOT in independent tests. The CVs were 12.4% for SB#5.1 and 17.0% for SB#5.2. Representative scans of the ELISPOT filters from wells stimulated with an irrelevant control peptide (left panel) or the cognate peptide (right panel) are shown for time point 5 after 5 mo of storage.

Close modal

We manufactured further NY-ESO-1–specific TERS batches containing MHC class I– or II–restricted Ag-specific TCRs and tested them in CFC as a functional readout. In this study, single aliquots of each batch were thawed and stimulated for 6 h using the cognate peptides prior to testing in CFC for CD107a, IFN-γ, IL-2, and TNF-α. Fig. 2B shows the Ag-specific IFN-γ, IL-2, and TNF-α secretion and CD107a surface expression for a MHC class II (DRB1*0401)–restricted, NY-ESO-1165–179–specific TERS batch (SB#3). As shown in Fig. 2B, the stimulation with the cognate peptide CFLPVFLAQPPSGQR induced clearly detectable cytokine signals with mean values of 2.38, 0.66, 0.57, and 0.55% for TNF-α, IFN-γ, CD107a, and IL-2, respectively. Serial thawing of multiple aliquots was performed to study intra- and interassay variation. Interassay CVs were 12.8, 14.8, 12.4, and 20.7% for the same markers, respectively. Representative dot plots for each of the tested activation markers of the SB#3 that were stimulated with the specific NY-ESO-1165–179 peptide or with an irrelevant peptide are also shown in the figure. The unspecific background staining induced by stimulation with the irrelevant peptide was low for all four activation markers. Supplemental Fig. 2A shows dot plots from the stainings performed with SB#3 after 0, 2, 5, and 6 mo of storage.

In addition, multiple aliquots from a TERS batch (SB#5) containing HLA-B7–restricted NY-ESO-197–111–specific T cells were tested in the ELISPOT assay as a second commonly and often used functional assay. Fig. 2C shows results generated with the tested TERS batches SB#5.1 and SB#5.2, respectively. No spots were detected using autologous negative cells depicted by clean filter scans, whereas the TCR-positive TERS produce clearly detectable spots demonstrating the Ag-specific cytokine secretion after stimulation with the cognate peptide. The mean spot counts of 74 per 3 × 105 PBMCs and 48 per 3 × 105 PBMCs were at a level that can be observed in patient samples. Both SB showed clearly detectable spots and consistent triplicates throughout the testing period of 5 mo. The interassay CVs observed across five independent tests performed were 12.4% for SB#5.1 and 17.0% for SB#5.2. The original data sets from all tested time points are shown in Supplemental Fig. 2B.

Following the proof-of-concept experiments in three commonly used T cell assays, we manufactured further TERS batches containing a very low number of Ag-specific T cells, similar in frequency to spontaneous tumor-induced T cell response levels detectable in cancer patients. In this study, we focused on the size and quality of Ag-specific signals in the various readouts.

Single aliquots of a MHC class I–restricted TERS specific for NY-ESO-1157–165 with four diluted SB ranging from a low dilution of 1:34 to a high dilution of 1:4250 (SB#6.1 to SB#6.4) were tested using dextramer staining with expected frequencies between 2.5 and 0.05% Dext+ of CD8+. As depicted in the dot plots of Fig. 3A, the 1:34 diluted SB#6.1 led to a high frequency of 2.51% Dext+ of CD8+, whereas the more highly diluted SB#6.2 (1:170), SB#6.3 (1:850), and SB#6.4 (1:4250) induced reduced signals of 0.47, 0.17, and 0.05% Dext+ of CD8+ T cells, respectively. The appearance of the dextramer-positive populations in all four dilutions resembled the staining patterns typically observed in donors with endogenous T cell responses. Specific T cells were clearly detected at all four levels, supported by the distinct separation of the dextramer-negative and -positive populations and by the absence of the dextramer-positive population in the negative fraction. The variation of the signals over time was analyzed over ∼6 mo and across six independent experiments and was shown to be low with an interassay CVs of <25% for SB#6.1 to SB#6.3 and a CV of 29.71% for SB#6.4 (Fig. 3A).

FIGURE 3.

Signal variability of highly diluted TERS. (A) MHC–peptide multimer staining experiments. The figure shows the percentage of dextramer-positive CD8+ T cells for a NY-ESO-1157–165–specific HLA-A2–restricted TERS manufactured in four different dilutions: 1:34 (SB#6.1, closed circles), 1:170 (SB#6.2, open circles), 1:850 (SB#6.3; closed diamonds), and 1:4250 (SB#6.4; open diamonds), with expected Dext+ of CD8+ frequencies of 2.5, 0.5, 0.2, and 0.05%, respectively. Each symbol represents an independent experiment, i.e., one single dextramer staining of one time point of a single TERS aliquot. The boxes indicate the CV in % for the corresponding series of experiments that were performed 1, 2, and 6 mo after the manufacturing date. On the right, representative dot plots from one time point (2-mo storage) are shown for each of the four TERS SB and for the autologous negative cell fraction stained with a NY-ESO-1157–165 dextramer. (B) Cytokine flow cytometry. Two TERS batches with tyrosinase368–376- and NY-ESO-1157–165–specific HLA-A2–restricted T cells were manufactured (SB#7, upper row; SB#8, lower row). The representative dot plots show results for the markers IFN-γ, TNF-α, IL-2, and CD107a (y-axis) in the CD8+ T cell subset (x-axis). The first and third rows show results generated following stimulation of the TERS batches with the relevant peptide. The second and fourth rows show the nonspecific background signal generated by peptide stimulation of the autologous negative cell fraction. The left side of the figure indicates the percentage of dextramer-positive CD8+ T cells for both TERS. (C) IFN-γ ELISPOT. Three different diluted TERS batches, SB#9, SB#5.1, and SB#10, as well as the three corresponding autologous negative cell fractions, were applied in the IFN-γ ELISPOT assay and stimulated with the corresponding peptide in triplicates using 300,000 PBMCs per well. The table indicates the unique batch ID, the specificity of the tested batches, and the corresponding ELISPOT filter scans.

FIGURE 3.

Signal variability of highly diluted TERS. (A) MHC–peptide multimer staining experiments. The figure shows the percentage of dextramer-positive CD8+ T cells for a NY-ESO-1157–165–specific HLA-A2–restricted TERS manufactured in four different dilutions: 1:34 (SB#6.1, closed circles), 1:170 (SB#6.2, open circles), 1:850 (SB#6.3; closed diamonds), and 1:4250 (SB#6.4; open diamonds), with expected Dext+ of CD8+ frequencies of 2.5, 0.5, 0.2, and 0.05%, respectively. Each symbol represents an independent experiment, i.e., one single dextramer staining of one time point of a single TERS aliquot. The boxes indicate the CV in % for the corresponding series of experiments that were performed 1, 2, and 6 mo after the manufacturing date. On the right, representative dot plots from one time point (2-mo storage) are shown for each of the four TERS SB and for the autologous negative cell fraction stained with a NY-ESO-1157–165 dextramer. (B) Cytokine flow cytometry. Two TERS batches with tyrosinase368–376- and NY-ESO-1157–165–specific HLA-A2–restricted T cells were manufactured (SB#7, upper row; SB#8, lower row). The representative dot plots show results for the markers IFN-γ, TNF-α, IL-2, and CD107a (y-axis) in the CD8+ T cell subset (x-axis). The first and third rows show results generated following stimulation of the TERS batches with the relevant peptide. The second and fourth rows show the nonspecific background signal generated by peptide stimulation of the autologous negative cell fraction. The left side of the figure indicates the percentage of dextramer-positive CD8+ T cells for both TERS. (C) IFN-γ ELISPOT. Three different diluted TERS batches, SB#9, SB#5.1, and SB#10, as well as the three corresponding autologous negative cell fractions, were applied in the IFN-γ ELISPOT assay and stimulated with the corresponding peptide in triplicates using 300,000 PBMCs per well. The table indicates the unique batch ID, the specificity of the tested batches, and the corresponding ELISPOT filter scans.

Close modal

Next, we tested two MHC class I–restricted TERS batches engineered with TCRs of different affinities, namely one TCR specific for tyrosinase368–376 (SB#7) and one TCR specific for NY-ESO-1157–165 (SB#8) in CFC (Fig. 3B). The aim of this study was to test the functional profiles in the readout following the transfer of TCRs with different affinities. The cells were stimulated with the cognate peptides and stained for IFN-γ, TNF-α, IL-2, and CD107a. The 1:26 diluted tyrosinase-specific SB#7, with a confirmed dextramer frequency of 2.9% of the CD8+ cells (data not shown), showed a clear expression of the four tested markers with values of 0.11% for IFN-γ, 0.33% for TNF-α, 0.05% for IL-2, and 0.21% for CD107a (% of CD8+). The NY-ESO-1–specific 1:92 diluted SB#8 with a dextramer signal of 1.0% of the CD8+ cells showed a strong signal of 0.48% for IFN-γ, 0.79% for TNF-α, 0.18% for IL-2, and 0.54% for CD107a. The corresponding autologous negative cell fractions showed only low background staining for all four activation markers in both batches, indicating that the positive signals observed after stimulation were specifically induced by the cognate peptide and were derived from TCR-transduced cells. Notably, the response profiles of the two SB against the specific peptides differed. Although containing a higher frequency of Ag-specific TCR-engineered T cells, as determined by dextramer staining (2.9% Dext+ of CD8+), the lower affinity tyrosinase-specific SB#7 led to a lower induction of cytokine secretion compared with the NY-ESO-1–specific SB#8 that only contain 1% of the TCR-engineered CD8+ T cells. Interestingly, the higher affinity NY-ESO-1–specific TCR used to engineer SB#8 did not only lead to a higher frequency of cytokine-positive cells, but also led to the induction of higher absolute amounts of cytokines, as reflected by the increased mean fluorescent intensities in each cytokine staining. These findings suggest that, depending on the used TCRs, we can engineer TERS with various functional profiles mimicking endogenous T cell responses.

In addition, we tested three additional highly diluted SB (#9, #5.1, #10) in an ELISPOT assay to study whether the spot appearance and intensity vary depending on the inserted Ag-specific TCR. Fig. 3C shows results from the three NY-ESO-1–specific SB that were engineered using TCRs with different peptide specificities and MHC restrictions (DRB01*0401, B*0702, and A*0201). All three TERS batches produced clearly detectable spots, consistent replicates, and a homogenous distribution of spots across the wells. The spot size varied across different Ags, similar to what one can observe when patient samples are tested. The tested cells from the negative cell fraction that are lacking engineered TCR-positive cells induced no spots following Ag-specific stimulation.

Finally, we conducted a direct side-by-side comparison of the assay signals obtained from a TERS batch compared with an endogenous T cell response from a PBMC sample for all three immune assay formats. To this end, we manufactured a TERS batch specific for the HLA-A2–associated peptide tyrosinase368–376 (SB#11) and tested the control sample in MHC–peptide multimer staining, in CFC for IFN-γ, TNF-α, IL-2, and CD107a, as well as in an IFN-γ ELISPOT. In parallel we tested a PBMC sample from a melanoma patient with known ex vivo detectable reactivity against the same tyrosinase368–376 epitope in the same three cellular assays. Supplemental Fig. 3 shows a side-by-side comparison of the signal morphology/quality of assay signals derived either from a tyrosinase368–376-specific batch on one hand and an endogenous population of tyrosinase368–376-specific T cells on the other hand. The experiment included three different assay formats to quantify Ag-specific T cell responses. The resulting raw data confirm that TERS (shown in the upper row) give rise to assay signals that closely resemble those signals obtained by testing clinical sample specimens that harbor endogenous T cell reactivity of interest (shown in the lower row). Notably, the morphology/quality and size of the Ag-specific signal of the TERS and the patient differ between the three assay formats and the respective activation marker assessed, which is not unexpected.

To study the long-term stability of TERS, we tested two MHC class I–restricted SB specific for tyrosinase368–376 (SB#11) and NY-ESO-1157–165 (SB#12) that were stored in liquid nitrogen for 12 mo. Single aliquots were thawed and analyzed independently 2 wk after initial freezing and then repetitively every 2 mo for 1 y using three NY-ESO-1–specific MHC–peptide dextramers labeled with PE and FITC, as well as tyrosinase dextramers labeled with allophycocyanin. The signal from each TERS was clearly detectable with each dextramer stored at 4°C, as shown in Fig. 4 for tyrosinase allophycocyanin, NY-ESO-1 PE, and NY-ESO-1 FITC. Dextramer reagents used in this stability study were stored either at 4°C or −20°C, which led to similar results (data not shown). After 6-mo testing period, fresh dextramers were additionally used for ongoing TERS stability tests. The detected signals using both the fresh and the stored dextramers were comparable for the same thawed TERS aliquot tested on each time point (data not shown). The tyrosinase-specific reactivity of SB#11 detected with the cognate allophycocyanin-labeled dextramer was stable over time, with a mean response level of 9.23% Dext+ of CD8+ T cells and a low CV of 8.56%. The NY-ESO-1–specific SB#12 shows a mean response level of 10.91% Dext+ of CD8+ stained with the NY-ESO-1 PE dextramer and 10.94% Dext+ of CD8+ stained with the FITC-labeled dextramer. We observed high signal stability over time, characterized by a low variation of 5.08% for PE and 8.23% for FITC. We also stained the NY-ESO-1–specific SB#12 with an allophycocyanin-labeled dextramer, showing stable signals over time up to 12 mo of storage (data not shown). In summary, this study shows the high stability of the TERS, even after 12 mo of storage in liquid nitrogen, and also documents the potential of TERS to be used as a tool for testing reagent stability.

FIGURE 4.

Long-term stability of TERS and dextramer reagents. The figure shows MHC–peptide multimer experiments for TERS that were stored in liquid nitrogen for up to 12 mo. Two HLA-A2–restricted TERS batches specific for NY-ESO-1157–165 (SB#12; NY PE and NY FITC) and tyrosinase368–376 (SB#11; Tyr allophycocyanin) were tested. Single aliquots were thawed and analyzed independently (represented by each data point) 2 wk after initial freezing and then repetitively every second month for 1 y, using two NY-ESO-1157–165 dextramers labeled with PE (closed circles), FITC (open circles), and tyrosinase368–376 labeled with allophycocyanin (closed diamonds). The boxes indicate the CV %. On the right, representative dot plots for each TERS, as well as the corresponding autologous negative fraction at the initial staining at month 0, are shown.

FIGURE 4.

Long-term stability of TERS and dextramer reagents. The figure shows MHC–peptide multimer experiments for TERS that were stored in liquid nitrogen for up to 12 mo. Two HLA-A2–restricted TERS batches specific for NY-ESO-1157–165 (SB#12; NY PE and NY FITC) and tyrosinase368–376 (SB#11; Tyr allophycocyanin) were tested. Single aliquots were thawed and analyzed independently (represented by each data point) 2 wk after initial freezing and then repetitively every second month for 1 y, using two NY-ESO-1157–165 dextramers labeled with PE (closed circles), FITC (open circles), and tyrosinase368–376 labeled with allophycocyanin (closed diamonds). The boxes indicate the CV %. On the right, representative dot plots for each TERS, as well as the corresponding autologous negative fraction at the initial staining at month 0, are shown.

Close modal

Once the TERS technology was established in our laboratory, we wanted to see how TERS perform in the hands of other investigators. To this end, we organized two proficiency panels with experienced laboratories that applied centrally provided TERS in MHC–peptide multimer-staining experiments (Fig. 5A) and CFC (Fig. 5B). For the first proficiency panel, 12 participants received the same set of TERS specific for the HLA-A2–restricted NY-ESO-1157–165 epitope and consisting of the negative cell fraction, one SB containing a 1:7 dilution of RNA-engineered cells (SB#14.1) and a second SB containing a 1:14 dilution (SB#14.2). They were asked to quantify the number of dextramer-positive CD8+ T cells at two independent time points (T1 and T2) following their own assay protocol. All laboratories could recover sufficient number of cells after thawing to perform the required experiments. The overview of the immune assay results from all 12 laboratories led to a colorful picture of dot plots for both TERS batches, which derived from the diversity of staining protocols and applied strategies to analyze the generated raw data. Supplemental Fig. 4A shows representative dot plots of the first time point (T1) for SB#14.2 and the negative fraction. The staining results of the negative fraction with the NY-ESO-1 dextramer led to background signals between 0.00 and 0.26%. All 12 participants successfully detected clearly clustered dextramer-positive populations in SB#14.1 and SB#14.2 with the provided NY-ESO-1 dextramer in all experiments performed.

FIGURE 5.

TERS performance across institutions and assay protocols. (A) MHC–peptide multimer staining. The figure shows the overview of the dextramer staining results from all 12 laboratories that participated in the proficiency panel. The x-axis of the figure shows the unique ID of each participating laboratory. The y-axis shows the percentage of Dext+ of CD8+ T cells, as reported by the participants after staining of both TERS batches and the negative fraction at two independent time points (T1 and T2). Each symbol represents the result from one independent dextramer staining. Experiments were performed with SB#14.1 at T1 (closed circles) and T2 (indicated by x), with SB#14.2 at T1 (open circles) and T2 (indicated by +) or with the negative fraction at T1 (gray circles) and T2 (indicated by ж). (B) Cytokine flow cytometry. The figure shows the overview of the CFC results from the five laboratories that participated in the proficiency panel testing a centrally provided tyrosinase368–376-specific TERS SB#11. The left part of the figure summarizes the results for IFN-γ, and the right part of the figure indicates the results for TNF-α secretion. The x-axis shows the unique ID of each participating laboratory, and the y-axis the percentage of cytokine-positive CD8+ T cells, as reported by the participants after stimulation of the TERS batch at two independent time points (T1 and T2). Each symbol represents the results of one independent staining. The experiments were performed following stimulation with the cognate tyrosinase368–376 peptide at T1 (closed circles) and T2 (indicated by x), or following stimulation with an irrelevant peptide at T1 (gray circles) and T2 (indicated by ж).

FIGURE 5.

TERS performance across institutions and assay protocols. (A) MHC–peptide multimer staining. The figure shows the overview of the dextramer staining results from all 12 laboratories that participated in the proficiency panel. The x-axis of the figure shows the unique ID of each participating laboratory. The y-axis shows the percentage of Dext+ of CD8+ T cells, as reported by the participants after staining of both TERS batches and the negative fraction at two independent time points (T1 and T2). Each symbol represents the result from one independent dextramer staining. Experiments were performed with SB#14.1 at T1 (closed circles) and T2 (indicated by x), with SB#14.2 at T1 (open circles) and T2 (indicated by +) or with the negative fraction at T1 (gray circles) and T2 (indicated by ж). (B) Cytokine flow cytometry. The figure shows the overview of the CFC results from the five laboratories that participated in the proficiency panel testing a centrally provided tyrosinase368–376-specific TERS SB#11. The left part of the figure summarizes the results for IFN-γ, and the right part of the figure indicates the results for TNF-α secretion. The x-axis shows the unique ID of each participating laboratory, and the y-axis the percentage of cytokine-positive CD8+ T cells, as reported by the participants after stimulation of the TERS batch at two independent time points (T1 and T2). Each symbol represents the results of one independent staining. The experiments were performed following stimulation with the cognate tyrosinase368–376 peptide at T1 (closed circles) and T2 (indicated by x), or following stimulation with an irrelevant peptide at T1 (gray circles) and T2 (indicated by ж).

Close modal

The reproducibility of the NY-ESO-1–specific signal from the TERS batches was assessed by repeated staining at two independent time points (T1 and T2) and is shown in Fig. 5A for all three batches detected by each laboratory. This interassay variation generated by the 12 laboratories was always <20%. Ten laboratories even generated data with a variation <10% (data not shown). These results confirm that the TERS are characterized by high signal stability irrespective of staining protocols and multiuser diversity.

Finally, when focusing on the results generated across institutions, we found a high concordance of reported frequencies. This was unexpected due to the heterogeneous and nonharmonized group and lack of standardization of applied protocols. The analysis of the intralab variation shows that the TERS gave robust results characterized by a low CV for all TERS batches and time points (T1-14.2 = 13.01%; T1-14.1 = 10.76%; T2-14.2 = 12.40%; T2-14.1 = 11.56%). This low variability indicates that TERS may be distributed among collaborating laboratories and could be used to control or normalize results generated across multiple centers.

The second proficiency panel was organized for testing TERS in CFC in a multiple-center setting (Fig. 5B). Five participating laboratories received a centrally generated TERS batch containing a defined, moderate number of tyrosinase157–165-specific CD8+ T cells (9.0% Dext+ of CD8+). Each participating laboratory stimulated the TERS with the tyrosinase157–165-specific peptide, an irrelevant peptide (NY-ESO-1157–165), with SEB as positive control and in medium only as negative control on one time point (T1), followed by intra- and extracellular staining. To assess reproducibility, the experiment was independently repeated with a second aliquot of TERS on a different time point (T2). In Fig. 5B, the tyrosinase-specific induced IFN-γ and TNF-α signals as well as the irrelevant background staining (irrelevant peptide) for both time points are presented. The small boxes highlight results obtained at T1 and T2 for the tyrosinase-specific cytokine signals. Each laboratory used a locally established CFC protocol, and the data analysis was also performed according to local preferences. The centers generated stable results on independent time points, which are characterized by an intralab variation <20% for both cytokine-specific signals. Only two laboratories showed a higher intralab variation >20% (lab09: TNF-α and IFN-γ; lab08: TNF-α). The interlab variation of both the tyrosinase-specific and irrelevant signals is high, as expected due to different protocols and gating strategies applied. The testing clearly shows the influence of different stimulation and staining protocols on the signal strength and the background staining. Two laboratories (09, 11) that observed high background signals also detected increased cytokine signals, especially for TNF-α. In contrast, laboratory 01 had a very low background but also low Ag-induced cytokine signals. This diversity of results reported across institutions is also depicted in Supplemental Fig. 4B, presenting a colorful picture of dot plots. Altogether, all five laboratories were able to detect TCR-specific signals with at least one cytokine in the two independent experimental runs.

In a last series of experiments, TERS were applied to assess the impact of common sources of variation on the results of two T cell assays (Fig. 6). We tested a range of scenarios in dextramer-staining experiments (Fig. 6A) using a NY-ESO-1157–165–specific TERS (SB#10) and in CFC (Fig. 6B) using a tyrosinase368–376-specific TERS (SB#11), showing the influence of different parameters such as the quality of the cell material, the cytometer settings, the gating strategy, and the reagent quality.

FIGURE 6.

Use of TERS to sensitively detect common sources of variation in T cell assays. (A) The figure shows the results of TERS SB#10 specific for NY-ESO-1157–165 tested in MHC–peptide multimer staining using a cognate PE-labeled dextramer. On the y-axis, the percentage Dext+ of CD8+ signal (upper part), representative dot plots (middle part), and the stain index (lower part) are shown. For the batch-specific cutoff definition (T1–T5 on the x-axis), five TERS aliquots were independently tested using standardized protocols for thawing of cells, MHC–peptide multimer staining, cytometer settings, and gating strategy. The median of 0.49% Dext+ of CD8+ cells and 20.28 for the stain index are marked as solid line. The acceptable ranges were set at 0.39–0.59% Dext+ of CD8+ and 16.22–24.34 for the stain index, which is marked as gray area. In the application phase, TERS were tested under intentionally induced suboptimal conditions of the following: 1) cell viability (medium and low viability: Via); 2) cytometer settings (voltages too high/low: Volt); 3) gating (inaccurate gate position: too low/high: Gate); and 4) reagent stability (expired dextramer: stored at 4°C for 11, 21, and 35 mo after production: Exp). (B) TERS SB#11 specific for tyrosinase368–376 was tested in CFC after tyrosinase368–376-specific peptide stimulation. The y-axis shows the IFN-γ signals of CD8+ cells (upper part), the corresponding dot plots (middle part), and the stain index (lower part). By testing five aliquots in independent experiments (T1–T5 on the x-axis), the median value was defined as 1.14% IFN-γ+ of CD8+ cells and a stain index of 17.65 (solid line) was defined. An acceptable range of between 0.91 and 1.37% IFN-γ+ of CD8+ cells and 14.12–21.18 for the stain index (gray area) was determined. Applying TERS under suboptimal conditions led to out of range results, indicated by circles located beyond the gray area.

FIGURE 6.

Use of TERS to sensitively detect common sources of variation in T cell assays. (A) The figure shows the results of TERS SB#10 specific for NY-ESO-1157–165 tested in MHC–peptide multimer staining using a cognate PE-labeled dextramer. On the y-axis, the percentage Dext+ of CD8+ signal (upper part), representative dot plots (middle part), and the stain index (lower part) are shown. For the batch-specific cutoff definition (T1–T5 on the x-axis), five TERS aliquots were independently tested using standardized protocols for thawing of cells, MHC–peptide multimer staining, cytometer settings, and gating strategy. The median of 0.49% Dext+ of CD8+ cells and 20.28 for the stain index are marked as solid line. The acceptable ranges were set at 0.39–0.59% Dext+ of CD8+ and 16.22–24.34 for the stain index, which is marked as gray area. In the application phase, TERS were tested under intentionally induced suboptimal conditions of the following: 1) cell viability (medium and low viability: Via); 2) cytometer settings (voltages too high/low: Volt); 3) gating (inaccurate gate position: too low/high: Gate); and 4) reagent stability (expired dextramer: stored at 4°C for 11, 21, and 35 mo after production: Exp). (B) TERS SB#11 specific for tyrosinase368–376 was tested in CFC after tyrosinase368–376-specific peptide stimulation. The y-axis shows the IFN-γ signals of CD8+ cells (upper part), the corresponding dot plots (middle part), and the stain index (lower part). By testing five aliquots in independent experiments (T1–T5 on the x-axis), the median value was defined as 1.14% IFN-γ+ of CD8+ cells and a stain index of 17.65 (solid line) was defined. An acceptable range of between 0.91 and 1.37% IFN-γ+ of CD8+ cells and 14.12–21.18 for the stain index (gray area) was determined. Applying TERS under suboptimal conditions led to out of range results, indicated by circles located beyond the gray area.

Close modal

Initially, we performed five independent tests following standardized protocols, for defining the batch-specific, expected signal size and for defining the acceptable range of results. The experiments for this cutoff definition phase were performed after standardized conditions using optimized protocols for thawing cells, for MHC–peptide multimer staining, and for CFC (described in 2Materials and Methods), aiming for a maximum cell viability (>90%), highly stable and quality controlled reagents, optimal flow cytometer settings (voltages and compensation values), and a gating strategy that allows the detection of all TCR-specific T cells. Subsequently, new aliquots of the same TERS were tested under intentionally induced suboptimal conditions mimicking commonly known sources of assay variation.

Testing the influence of cell quality, we thawed TERS aliquots under harsh conditions obtaining a medium (∼80%) or a low (∼60%) cell viability. The different cell samples were then tested and analyzed using the standardized assay conditions. Testing the influence of cytometer settings, TERS were processed using the standard procedures. Then these cells were acquired using increased (+150 V) or decreased (−150 V) voltages from the optimal settings (dextramer-PE). The following analysis (gating) was performed adequately and partially adapted to the shifted populations. Finally, TERS were used to assess the influence of inaccurate gating. For the detection of Ag-specific T cells (Dext+ of CD8+ and IFN-γ+ of CD8+), the gates were set too far from the Dext+/IFN-γ+–negative population or too close to the negative population. The strategy was to move the gates not too far to force variation, but we set the gates in such a way as it may be done by users who are not highly experienced in flow cytometry. Finally, the dextramer reagent stability was tested by using highly viable TERS that were stained and analyzed after standard protocols.

For the SB#10 specific for NY-ESO-1157–165, tested in dextramer staining (Fig. 6A), we conducted the cutoff definition phase and defined an expected value (median of T1–T5) of 0.49% Dext+ of CD8+ cells and 20.28 for the stain index, which is calculated by mean fluorescence intensity (MFI)CD8+Dext+ − MFICD8+Dext−/2 × SDCD8+Dext−. The acceptable ranges were defined as 0.39–0.59% Dext+ of CD8+ cells and 16.22–24.34 for the stain index by allowing a 20% variation above and below the median value. TERS that were subsequently tested under suboptimal conditions led to variable results that were partially out of the acceptable range. Notably, the staining with MHC–peptide dextramers leads to such robust signals that the percentage of Dext+ of CD8+ cells in some cases still fell in the acceptable range, even under suboptimal conditions (e.g., when gates were set too low or when long-time stored dextramers were used). Notably, use of the staining index was a more sensitive measure of variation.

In Fig. 6B, TERS SB#11 specific for tyrosinase368–376 was tested in CFC. In the cutoff definition phase, the results of five independently tested aliquots (T1–T5) were used to define the expected value (median) of 1.14% IFN-γ+ of CD8+ cells and 17.65 for the stain index. The acceptable range was set between 0.91 and 1.37% IFN-γ+ of CD8+ cells and 14.12–21.18 for the stain index (20% variation). Using TERS under suboptimal conditions led to cutoff range results. In these artificially induced suboptimal test conditions, low cell viability, incorrect cytometer settings, and inadequate gating strategy caused a measurable reduction of the Ag-specific signal.

Avoidance of immune destruction was recently established as a hallmark of cancer that can be therapeutically targeted, and the key role of the immune system to confer tumor growth control has now been commonly accepted (17). Novel therapies modulating the state of the immune system that exert their effects via Ag-specific T cells have been already introduced to clinical care in oncology (18, 19). In addition, there is growing evidence that Ag-specific T cells confer the antitumor effects of checkpoint inhibitory Abs and adoptive transfer of tumor-infiltrating lymphocytes (20, 21). The next waves of immunotherapies are in the advanced stages of development and may soon reach a larger number of cancer patients (2225). The progress recently made in understanding the mechanisms of action of immunotherapies has led to an immunological and regulatory framework for immuno-oncology that in particular acknowledges the biological peculiarities of cancer immunotherapy. As an integral part of the new framework, use of robust and controlled immune assays has been proposed as a critical component to enhance the development of novel immunotherapies (26). The increasing use of immunotherapeutic approaches that are based on the modulation of Ag-specific T cell response will most probably be accompanied by an increased need for robust and performance-controlled immune assessments. In addition to regulatory and mode of action–driven considerations, the cost pressure in health care systems will probably stimulate the search for robust immunological biomarkers for stratification of patients with increased likelihood to show marked clinical benefits, which is a prerequisite for novel therapies to stay affordable (27). To increase the clinical utility of a novel therapeutic in personalized medicine, the use of biomarkers has been proposed (28).

Applying immune assays to guide the rational development of novel therapies or clinical decision making mandates both a deep understanding, as well as rigid control of the variables impacting on the test results (13). The reasoning behind our work was the increasing scientific, regulatory, medical, and economical need to employ performance-controlled immune assays in correlative biomarker studies as well as the current lack of adequate, easily available assay control reagents.

A number of obstacles reaching beyond controlling the technical assay variation are limiting the broader use of T cell assays. These include open unresolved questions regarding the following: 1) the right Ag format for use in in vitro assays; 2) the issue of assessing the frequency of T cells with a particular functional profile or defined TCR avidity; 3) the association between the frequency of Ag-specific cytokine responses and the quality of the APCs available in the in vitro assays; and 4) the variable quality of human subject blood specimens obtained throughout clinical trials. One of the most highly disputed topics is the question regarding which compartment should be tapped for monitoring of T cell responses, as there is uncertainty whether immune responses measured in the peripheral blood appropriately reflect the in vivo situation in the end organ (e.g., tumor or infected tissue). All of these unresolved questions need to be systematically addressed in subsequent studies prior to implementing immune assays as part of clinical decision making. Use of TERS batches in such studies is recommended as this would facilitate drawing firm conclusions from results as uncertainty introduced by assay variation could be efficiently reduced.

TERS with a defined number of functional Ag-specific T cells to control immune assay performance over time, or across institutions, are not currently readily available. To date, PBMCs with confirmed presence of Ag-specific responses against a narrow selection of known microbial recall Ags such as HLA-A2–restricted epitopes or peptide pools derived from CMV, EBV, or Influenza (Flu) are regularly used as controls in immune assays testing CD4+ or CD8+ T cell subsets. Currently, there is a lack of control reagents for T cell assays that 1) have a specificity against rarer microbial Ags, auto-associated Ags, or TAAs; 2) are directed against peptide epitopes binding to less prevalent HLA alleles; 3) recognize HLA class II–presented Ags; 4) reflect functional properties of TCRs with medium to low affinity; and 5) are available without restrictions. Recently, we confirmed that TERS equipped with a tumor-specific TCR following retroviral gene transfer, cell sorting, and in vitro expansion and then spiked into PBMCs constitute a tool to control immune assay performance (29). However, manufacturing of TERS using retroviral vectors requires a specific expertise and is rather laborious. Also, due to the increased biosafety level of gene-modified cells, the transport and distribution of transduced cells are difficult. Consequently, upscaling of the approach for broad adoption of retroviral engineered reference samples is difficult.

The RNA-based TERS technology presented in this work and the concept for its use were developed to enable continuous control of immune assay performance. Using chemically and molecularly optimized TCR RNA and a systematic optimization of the manufacturing process, we established a simple and scalable technology suited for the three most commonly used CD4+ and CD8+ T cell assays. TERS deliver robust assay results that quantitatively and qualitatively resemble physiological signals observed in sample specimens from patients. The quality of signals obtained in all three commonly used immune assays varies depending on the TCR and can therefore mimic response levels as obtained in patient material containing high- or low-affinity TCRs. TERS batches show low inter- and intraday variability, irrespective of the Ag specificity and the number of Ag-specific T cells. A high specificity of each TERS is proven by a low background staining of the negative fraction, as documented for all TERS batches used in this study. TERS are stable for at least 12 mo and perform equivalently across different protocols in multiple laboratories in Europe and the United States.

As part of the immune monitoring of a clinical study, we propose a three-part concept for use of the TERS technology that is based on a standardized manufacturing process with quality controls, a protocol-specific definition of acceptable cutoffs, and the use prior to and after measurement of sample specimens from patients. We show that by use of cutoffs, TERS can be applied to sensitively capture low quality of starting cell material, quality issues of critical reagents, wrong cytometer settings, as well as incorrect data analysis.

TERS are therefore ready for broader use to qualify either reagents or whole experimental workflows and individual operators, prior to initiation of test campaigns or concomitant use during the analysis of patient material. We have now started the regular use of TERS technology in clinical biomarker programs in our laboratory. Although primarily meant to control immune assay performance within institutions, the data generated in the presented proficiency panel suggest that TERS can be implemented in research consortia to efficiently control and possibly normalize results generated across institutions. Knowing that, in addition to variability, the comparability of results generated across institutions is another hampering factor in the field, the potential impact of the TERS technology to enhance immunotherapy is enormous.

The presented studies focused on TCRs specific for known TAAs that are known to be of low to moderate affinity at best. Most importantly, the described findings and the TERS concept are fully transferrable from cancer immunotherapy applications to any other area of immunology. In addition, RNA technology for generation of TERS is principally suited for kit-based approaches that would increase the ease of and would allow serving an even larger group of investigators worldwide. Following manufacturing of TERS batches at peripheral sites, these could be shared among collaborating institutions. We have recently published a novel process for functional retrieval of Ag-specific TCRs from single cells (30). This opens the opportunity to generate a large collection of TCRs with different affinities directed against a versatile collection of Ags for customized TERS kits. The biggest advantage of kit-based approaches arises from the virtually unlimited availability of quality-tested RNA for lymphocyte electroporation. A first test panel of a kit-based approach in seven European laboratories has just been initiated.

Thus, TERS technology offers an original, robust, and versatile tool to overcome one of the major hurdles for broad adoption of T cell assays in systematic human immunology approaches and clinical development of immunotherapies. The new technology comes with the hope of accelerating translation of basic immunology findings into better diagnostics and treatments for patients.

Immudex (Copenhagen, Denmark) provided dextramers for TERS testing and for the proficiency panels. We thank Dr. T. Schumacher for providing the wild-type NY-ESO-1157–165–specific TCR, the whole Association for Cancer Immunotherapy Immunoguiding Program team for support, and Dirk Schröder for mathematical know-how. The tyrosinase368–376-specific CTL clone IVSB was provided by T. Wölfel. Further thanks to all participants of the proficiency panels, namely Pia Kvistborg and M. Toebes (National Cancer Institute, Amsterdam, the Netherlands); P. Palluch and M. Subklewe (Helmholtz-Zentrum München, München, Germany); R. Mendrzyk and S. Walter; D. Paduch and D. Maurer (Immatics Biotechnologies, Tübingen, Germany); T. Jakobsen and H. Petersen (Immudex, Copenhagen, Denmark); M. Hasan, C. Alanio, V. Libri, and M. Albert (Center for Human Immunology, Institut Pasteur, Paris, France); M.J.P. Schoenmaekers-Welters and N. Loof (Leiden University Medical Center, Leiden, the Netherlands); A. Cazaly and R. Challis; D. Joseph-Pietras and K. McCann (Southampton General Hospital, Southampton, U.K.); J. de Vries and T. Duiveman-de Boer (Radboud University Nijmegen Medical Centre, Nijmengen, the Netherlands); J. Matsuzaki and A. Beck (Roswell Park Cancer Institute, Buffalo, NY); S. Reker Hadrup and T. Seremet (University Hospital Herlev, Herlev, Denmark); and S. Heidu, K. Laske, and C. Gouttefangeas (University of Tübingen, Tübingen, Germany).

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • 7-AAD

    7-aminoactinomycin D

  •  
  • CFC

    cytokine flow cytometry

  •  
  • CV

    coefficient of variation

  •  
  • FSC

    forward light scatter

  •  
  • RT

    room temperature

  •  
  • SB

    subbatch

  •  
  • SEB

    staphylococcal enterotoxin B

  •  
  • TAA

    tumor-associated Ag

  •  
  • TERS

    TCR-engineered reference sample

  •  
  • wt

    wild-type.

1
Haynes
B. F.
,
Gilbert
P. B.
,
McElrath
M. J.
,
Zolla-Pazner
S.
,
Tomaras
G. D.
,
Alam
S. M.
,
Evans
D. T.
,
Montefiori
D. C.
,
Karnasuta
C.
,
Sutthent
R.
, et al
.
2012
.
Immune-correlates analysis of an HIV-1 vaccine efficacy trial.
N. Engl. J. Med.
366
:
1275
1286
.
2
Kenter
G. G.
,
Welters
M. J.
,
Valentijn
A. R.
,
Lowik
M. J.
,
Berends-van der Meer
D. M.
,
Vloon
A. P.
,
Essahsah
F.
,
Fathers
L. M.
,
Offringa
R.
,
Drijfhout
J. W.
, et al
.
2009
.
Vaccination against HPV-16 oncoproteins for vulvar intraepithelial neoplasia.
N. Engl. J. Med.
361
:
1838
1847
.
3
Hanekom
W. A.
,
Dockrell
H. M.
,
Ottenhoff
T. H.
,
Doherty
T. M.
,
Fletcher
H.
,
McShane
H.
,
Weichold
F. F.
,
Hoft
D. F.
,
Parida
S. K.
,
Fruth
U. J.
.
2008
.
Immunological outcomes of new tuberculosis vaccine trials: WHO panel recommendations.
PLoS Med.
5
:
e145
.
4
Roep
B. O.
,
Buckner
J.
,
Sawcer
S.
,
Toes
R.
,
Zipp
F.
.
2012
.
The problems and promises of research into human immunology and autoimmune disease.
Nat. Med.
18
:
48
53
.
5
Mingozzi
F.
,
High
K. A.
.
2013
.
Immune responses to AAV vectors: overcoming barriers to successful gene therapy.
Blood
122
:
23
36
.
6
Czerkinsky
C.
,
Andersson
G.
,
Ekre
H. P.
,
Nilsson
L. A.
,
Klareskog
L.
,
Ouchterlony
O.
.
1988
.
Reverse ELISPOT assay for clonal analysis of cytokine production. I. Enumeration of gamma-interferon-secreting cells.
J. Immunol. Methods
110
:
29
36
.
7
Chattopadhyay
P. K.
,
Hogerkorp
C. M.
,
Roederer
M.
.
2008
.
A chromatic explosion: the development and future of multiparameter flow cytometry.
Immunology
125
:
441
449
.
8
Bendall
S. C.
,
Simonds
E. F.
,
Qiu
P.
,
Amir
A. D.
,
Krutzik
P. O.
,
Finck
R.
,
Bruggner
R. V.
,
Melamed
R.
,
Trejo
A.
,
Ornatsky
O. I.
, et al
.
2011
.
Single-cell mass cytometry of differential immune and drug responses across a human hematopoietic continuum.
Science
332
:
687
696
.
9
Hadrup
S. R.
,
Bakker
A. H.
,
Shu
C. J.
,
Andersen
R. S.
,
van Veluw
J.
,
Hombrink
P.
,
Castermans
E.
,
Thor Straten
P.
,
Blank
C.
,
Haanen
J. B.
, et al
.
2009
.
Parallel detection of antigen-specific T-cell responses by multidimensional encoding of MHC multimers.
Nat. Methods
6
:
520
526
.
10
Newell
E. W.
,
Sigal
N.
,
Nair
N.
,
Kidd
B. A.
,
Greenberg
H. B.
,
Davis
M. M.
.
2013
.
Combinatorial tetramer staining and mass cytometry analysis facilitate T-cell epitope mapping and characterization.
Nat. Biotechnol.
31
:
623
629
.
11
Janetzki
S.
,
Britten
C. M.
,
Kalos
M.
,
Levitsky
H. I.
,
Maecker
H. T.
,
Melief
C. J.
,
Old
L. J.
,
Romero
P.
,
Hoos
A.
,
Davis
M. M.
.
2009
.
“MIATA”-minimal information about T cell assays.
Immunity
31
:
527
528
.
12
Maecker
H. T.
,
McCoy
J. P.
,
Nussenblatt
R.
.
2012
.
Standardizing immunophenotyping for the Human Immunology Project.
Nat. Rev. Immunol.
12
:
191
200
.
13
Maecker
H. T.
,
McCoy
J. P.
 Jr.
the FOCIS Human Immunophenotyping Consortium
.
2010
.
A model for harmonizing flow cytometry in clinical trials.
Nat. Immunol.
11
:
975
978
.
14
van der Burg, S. H., M. Kalos, C. Gouttefangeas, S. Janetzki, C. Ottensmeier, M. J. Welters, P. Romero, C. M. Britten, and A. Hoos. 2011. Harmonization of immune biomarker assays for clinical studies. Sci. Transl. Med. 3: 108ps44
.
15
Britten
C. M.
,
Janetzki
S.
,
Butterfield
L. H.
,
Ferrari
G.
,
Gouttefangeas
C.
,
Huber
C.
,
Kalos
M.
,
Levitsky
H. I.
,
Maecker
H. T.
,
Melief
C. J.
, et al
.
2012
.
T cell assays and MIATA: the essential minimum for maximum impact.
Immunity
37
:
1
2
.
16
Kuhn
A. N.
,
Diken
M.
,
Kreiter
S.
,
Selmi
A.
,
Kowalska
J.
,
Jemielity
J.
,
Darzynkiewicz
E.
,
Huber
C.
,
Türeci
O.
,
Sahin
U.
.
2010
.
Phosphorothioate cap analogs increase stability and translational efficiency of RNA vaccines in immature dendritic cells and induce superior immune responses in vivo.
Gene Ther.
17
:
961
971
.
17
Hanahan
D.
,
Weinberg
R. A.
.
2011
.
Hallmarks of cancer: the next generation.
Cell
144
:
646
674
.
18
Kantoff
P. W.
,
Higano
C. S.
,
Shore
N. D.
,
Berger
E. R.
,
Small
E. J.
,
Penson
D. F.
,
Redfern
C. H.
,
Ferrari
A. C.
,
Dreicer
R.
,
Sims
R. B.
, et al
IMPACT Study Investigators
.
2010
.
Sipuleucel-T immunotherapy for castration-resistant prostate cancer.
N. Engl. J. Med.
363
:
411
422
.
19
Hodi
F. S.
,
O’Day
S. J.
,
McDermott
D. F.
,
Weber
R. W.
,
Sosman
J. A.
,
Haanen
J. B.
,
Gonzalez
R.
,
Robert
C.
,
Schadendorf
D.
,
Hassel
J. C.
, et al
.
2010
.
Improved survival with ipilimumab in patients with metastatic melanoma.
N. Engl. J. Med.
363
:
711
723
.
20
Gubin
M. M.
,
Zhang
X.
,
Schuster
H.
,
Caron
E.
,
Ward
J. P.
,
Noguchi
T.
,
Ivanova
Y.
,
Hundal
J.
,
Arthur
C. D.
,
Krebber
W. J.
, et al
.
2014
.
Checkpoint blockade cancer immunotherapy targets tumour-specific mutant antigens.
Nature
515
:
577
581
.
21
Tran
E.
,
Turcotte
S.
,
Gros
A.
,
Robbins
P. F.
,
Lu
Y. C.
,
Dudley
M. E.
,
Wunderlich
J. R.
,
Somerville
R. P.
,
Hogan
K.
,
Hinrichs
C. S.
, et al
.
2014
.
Cancer immunotherapy based on mutation-specific CD4+ T cells in a patient with epithelial cancer.
Science
344
:
641
645
.
22
Grupp
S. A.
,
Kalos
M.
,
Barrett
D.
,
Aplenc
R.
,
Porter
D. L.
,
Rheingold
S. R.
,
Teachey
D. T.
,
Chew
A.
,
Hauck
B.
,
Wright
J. F.
, et al
.
2013
.
Chimeric antigen receptor-modified T cells for acute lymphoid leukemia.
N. Engl. J. Med.
368
:
1509
1518
.
23
Wolchok
J. D.
,
Kluger
H.
,
Callahan
M. K.
,
Postow
M. A.
,
Rizvi
N. A.
,
Lesokhin
A. M.
,
Segal
N. H.
,
Ariyan
C. E.
,
Gordon
R. A.
,
Reed
K.
, et al
.
2013
.
Nivolumab plus ipilimumab in advanced melanoma.
N. Engl. J. Med.
369
:
122
133
.
24
Topalian
S. L.
,
Hodi
F. S.
,
Brahmer
J. R.
,
Gettinger
S. N.
,
Smith
D. C.
,
McDermott
D. F.
,
Powderly
J. D.
,
Carvajal
R. D.
,
Sosman
J. A.
,
Atkins
M. B.
, et al
.
2012
.
Safety, activity, and immune correlates of anti-PD-1 antibody in cancer.
N. Engl. J. Med.
366
:
2443
2454
.
25
Rosenberg
S. A.
,
Yang
J. C.
,
Sherry
R. M.
,
Kammula
U. S.
,
Hughes
M. S.
,
Phan
G. Q.
,
Citrin
D. E.
,
Restifo
N. P.
,
Robbins
P. F.
,
Wunderlich
J. R.
, et al
.
2011
.
Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy.
Clin. Cancer Res.
17
:
4550
4557
.
26
Hoos
A.
,
Britten
C. M.
,
Huber
C.
,
O’Donnell-Tormey
J.
.
2011
.
A methodological framework to enhance the clinical success of cancer immunotherapy.
Nat. Biotechnol.
29
:
867
870
.
27
Sobrero
A.
,
Bruzzi
P.
.
2009
.
Incremental advance or seismic shift? The need to raise the bar of efficacy for drug approval.
J. Clin. Oncol.
27
:
5868
5873
.
28
Hodgson
D. R.
,
Wellings
R.
,
Harbron
C.
.
2012
.
Practical perspectives of personalized healthcare in oncology.
New Biotechnol.
29
:
656
664
.
29
Singh
S. K.
,
Tummers
B.
,
Schumacher
T. N.
,
Gomez
R.
,
Franken
K. L.
,
Verdegaal
E. M.
,
Laske
K.
,
Gouttefangeas
C.
,
Ottensmeier
C.
,
Welters
M. J.
, et al
.
2013
.
The development of standard samples with a defined number of antigen-specific T cells to harmonize T cell assays: a proof-of-principle study.
Cancer Immunol. Immunother.
62
:
489
501
.
30
Simon
P.
,
Omokoko
T. A.
,
Breitkreuz
A.
,
Hebich
L.
,
Kreiter
S.
,
Attig
S.
,
Konur
A.
,
Britten
C. M.
,
Paret
C.
,
Dhaene
K.
, et al
.
2014
.
Functional TCR retrieval from single antigen-specific human T cells reveals multiple novel epitopes.
Cancer Immunol. Res.
2
:
1230
1244
.

U.S., A.K., and S.K. are coinventors of patented RNA technology that may be used to improve the functional characteristics of TCR-engineered reference samples.

Supplementary data