Validation of Ligand Tetramers for the Detection of Antigen-Specific Lymphocytes

This article is featured in Top Reads, p. 1013

Over the past several decades, the use of fluorochrome-conjugated ligands of all types has become a standard approach to detect Ag-specific lymphocytes by flow cytometry. For Ag-specific B cell detection, the most used tools are fluorochrome-conjugated Ag/peptide tetramers, fluorochrome-conjugated virus-like particles, as well as other types of fluorochrome-conjugated ligand (1–21). Combined, the use of these tools has led to important advances in all fields of immunology, the development of experimentally and/or clinically useful Abs, and a starting point for rationally designed vaccine Ags.

Throughout the course of a study analyzing Ag-specific B cells by flow cytometry, it is often necessary to produce dozens of Ag probe batches to conduct experiments over the long periods of time research often takes. Even with strict production protocols, this often means that different individuals have produced batches of Ags using disparate reagents. Ags can also undergo different storage times and temperatures during the various production steps or multiple freeze/thaw cycles, because protocols often neglect to set strict guidelines for every factor that could influence Ag structure. These issues are magnified across fields, where the same Ag could be produced, purified, and stored using completely different protocols. For example, one researcher may produce an Ag using a bacterial system, whereas another used a mammalian system, which could influence glycosylation, folding, or stability of the Ag. Alternatively, production and purification of an Ag may in one case include multiple freeze/thaw cycles, but another protocol would not, which could alter the stability of the Ag probe. Another difference can occur after probe production, when one researcher may produce an Ag tetramer fresh for each experiment, whereas another researcher stores the tetramer at 4°C or −20°C before use. Despite these sources of variation, there are typically few validation steps to ensure consistent and reproducible Ag probe performance within each experiment. In this article, we describe a rapid and robust approach to compare and validate the quality of Ag and MHC tetramers before and/or within each experiment that can be used to authenticate this key resource.

Six- to fourteen-week-old C57BL/6 male and female mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and maintained in a specific pathogen-free facility in accordance with Fred Hutchinson Cancer Center Institutional Animal Care and Use Committee approval and National Institutes of Health guidelines.

Biotinylated wild-type receptor binding domain (RBD) and biotinylated omicron RBD were purchased from Sinobiological or produced as described previously (22, 23). RBD-HISAVI protein was biotinylated and concentrated in vitro using the BirA500:BirA biotin-protein ligase standard reaction kit (Avidity) according to the manufacturer’s protocol and stored at −20°C. The average number of biotin molecules bound to each Ag molecule was confirmed, and tetramers were produced as described previously (24, 25). Tetramers were stored at 0.1–1 μM in 1× DPBS at 4°C or mixed 1:1 with 100% glycerol and stored at −20°C before use.

The spleen and inguinal, axillary, brachial, cervical, mesenteric, and periaortic lymph nodes from two to four mice were pooled, shredded using forceps, and forced through 100-μm mesh to generate filtered single-cell suspensions. Cell suspensions were then centrifuged at 300 × g for 5 min at 4°C, and pellets were resuspended in 1.2 ml ice-cold FACS buffer (1× DPBS containing 1% heat-inactivated newborn calf serum) and divided into twelve 0.1-ml fractions. Each fraction received 2 μg of anti-Fc receptor Ab 2.4G2 (BioXCell) and 0.75 pmol control allophycocyanin-DyLight755 tetramer before incubation on ice for 5 min. Next, 0.25 pmol RBD allophycocyanin tetramer, 0.2 μg anti-CD3 BV510 (145-2C11; BD Biosciences), 0.2 μg anti-F4/80 BV510 (BM8; BioLegend), 0.2 μg anti–Gr-1 BV510 (RB6-8C5; BD Biosciences), 0.3 μg anti-CD19 BUV737 (1D3; BD Biosciences), 0.4 μg anti-B220 BV786 or Pacific Blue (RA3-6B2; BioLegend), and 0.5 μl Fixable Viability Stain 620 (BD Biosciences) were added, and samples were incubated for 25 min on ice. After the incubation, ∼15 ml of FACS buffer was added, and the samples were centrifuged at 300 × g for 5 min at 4°C. The supernatant was discarded, and the pellet was resuspended before the addition of 8.33 μl of anti-allophycocyanin microbeads (Miltenyi Biotec). After a 30-min incubation on ice, tetramer-binding B cells were enriched as described previously (25). Flow cytometry was performed at the Fred Hutchinson Flow Cytometry Core Service on a five-laser (355, 405, 488, 561, and 640 nm) LSR II, LSRFortessa, or FACSymphony (BD Biosciences).

PBMCs were isolated via venipuncture and cryopreserved from adult volunteers before the COVID-19 pandemic as a part of the Fred Hutchinson Seattle Area Control study or from patients diagnosed with COVID-19 as part of NCT04338360 and NCT04344977 clinical trials. Institutional Review Board approval was obtained from Fred Hutchinson Cancer Center or the University of Washington, and all participants gave written informed consent. A total of 10⁷–10⁸ frozen PBMCs were thawed into DMEM containing 10% FCS, 100 U/ml penicillin, and 100 µg/ml streptomycin (all from ThermoFisher Scientific) and enriched for tetramer-binding B cells as previously described (26). Enriched cells were resuspended to 50 μl in FACS buffer containing 6.24 μg/ml anti-IgM FITC (G20–127; BD Biosciences), 2.5 μl anti-CD19 BUV395 (SJ25C1; BD Biosciences), 1 μl anti-CD3 BV711 (UCHT1; BD Biosciences), 1 μl anti-CD14 BV711 (M0P-9; BD Biosciences), 1 μl anti-CD16 BV711 (3G8; BD Biosciences), 2.5 μl anti-CD20 BUV737 (2H7; BD Biosciences), 2.5 μl anti-IgD BV605 (IA6–2; BD Biosciences), 3.125 μg/ml anti-CD27 PE-Cy7 (LG.7F9; eBioscience), and 0.5 μl Ghost Dye Violet 510 (Tonbo Biosciences) and incubated for 25 min on ice. After the incubation, ∼4 ml of FACS buffer was added, and the samples were centrifuged at 300 × g for 5 min at 4°C. Supernatants were discarded, cells were resuspended in FACS buffer, and tetramer-binding B cells were individually sorted using a five-laser (355, 405, 488, 561, and 640 nm) FACSAria (BD Biosciences) into empty 96-well low-profile PCR plates (Labcon), which were sealed with PCR microplate sealing foil (Eppendorf), and flash frozen before storage at −80°C.

A nested RT-PCR approach was used to sequence Ab H and L chain sequences from frozen PCR plates containing individually sorted B cells as described previously (27). Second-round PCR products were sequenced by Genewiz with the respective reverse primer used in the second-round PCR, and sequences were analyzed using IMGT/V-Quest to identify V, D, and J gene segments. Paired H chain VDJ and L chain VJ sequences were cloned into pTT3-derived expression vectors containing the human IgG1, Igκ, or Igλ constant regions using In-Fusion cloning (Clontech), and mAbs were produced and purified as previously described (23).

Biolayer interferometry (BLI) assays were performed on the Octet Red instrument (ForteBio) at room temperature with shaking at 500 rpm. For binding analyses, anti-human IgG Fc capture biosensors (ForteBio) were loaded with 10 or 40 μg/ml purified Ab in kinetics buffer (0.01% BSA, 0.02% Tween 20, and 0.005% NaN₃ in 1× DPBS [pH 7.4]) for 2.5 min. After loading, a baseline was recorded for 1 min in kinetics buffer. The sensors were then immersed in 62.5 or 500 nM S2P or RBD in kinetics buffer for 5 min to assess association, followed by immersion in kinetics buffer for 5 min to assess dissociation. A 1:1 binding model using ForteBio data analysis software was used for curve fitting.

MHC class I (MHCI) α chain (HLA-A2) and β₂-microglobulin (β2m) recombinant proteins were generated by overexpression in Escherichia coli and purification of inclusion bodies following a previously described protocol (28). To generate individual peptide:MHC monomer complexes, we set up a refolding reaction in an l-arginine–rich, glutathione redox refolding buffer with protease inhibitors in the presence of a high concentration of the HLA-A2–restricted UV labile peptide H-KLLT(1051)ILTI-OH (Mimotopes), where 1051 denotes a 3-amino-3 (2-nitro-phenyl)-propionic acid residue. In brief, peptide was dissolved in a spinning refolding buffer at 4°C before the slow injection of β2m followed by α chain protein dissolved in injection buffer (3 M guanidine-HCl, 10 mM Na acetate, and 10 mM EDTA [pH 4.2]). The solution was incubated at 4°C with slow stirring for 36 h with additional α chain added in a similar manner at 12 and 24 h. After incubation, precipitated debris was removed by centrifugation at 4000 × g for 5 min followed by filtration through 0.22-μm bottle-top filters (Corning). The monomer preparation was then concentrated, and buffer was exchanged into 10 mM Tris-HCl (pH 8.0) using 10 kDa MWCO Amicon Ultra-15 centrifugal filters (Millipore Sigma). Biotinylation of monomer was carried out using BirA biotin-protein ligase kit (Avidity) according to the manufacturer’s protocol. Biotinylated monomer was buffer exchanged and concentrated using 10 kDa MWCO Amicon Ultra-15 centrifugal filters before purification using the Superdex 200 HR 10/300 NGC Chromatography System (BioRad). The purified monomer was concentrated and stored at 0.5 mg/ml with 2 mg/ml leupeptin, 1 mg/ml pepstatin, 0.5 M EDTA, and 10% sodium azide at −80°C until use. Mouse MHC tetramers were made following the same protocol using mouse MHC H-2K^b H chain protein (provided by D. Masopust) and UV labile peptide H-SIINFE(1051)L-OH (Mimotopes).

To generate peptide-specific tetramers, we set up a UV exchange reaction as described by Fehlings et al. (29). Total protein concentration of tetramer was determined using Bradford assay kit (BioRad). Peptides used were EBV_GLC (GLCTLVAML; Biomatik), SARS-CoV-2_YLQ (YLQPRTFLL; Biomatik), and vaccinia virus B8R (TSYKFESV; Biomatik).

Human PBMCs were obtained from patients undergoing routine care following Dartmouth Institutional Review Board–approved protocols. Cryopreserved PBMCs were thawed and rested for 2 h in RPMI containing 10% FBS, 1% l-glutamine, 1% HEPES, 1% nonessential amino acids, 1% sodium pyruvate, 0.01% DNase, and 2-ME at 37°C. Cells were then washed with 1× PBS and stained with LIVE/DEAD Fixable Blue (ThermoFisher) for 30 min on ice. After splitting the sample into three, we centrifuged cells at 400 × g for 5 min at 4°C. The supernatant was discarded, and the cell pellet was stained with different batches of tetramer at 1:100 on ice for 30 min. Next, cells were centrifuged to remove supernatant and incubated for 30 min on ice in 50 μl FACS buffer (1× PBS with 0.2% BSA and 10% sodium azide) containing anti-CD4 PE-Cy5 (RPA-T4; BioLegend), anti-CD8 allophycocyanin-Fire750 (SK1; BioLegend), and anti-CD3 Alexa Fluor 700 (UCHT1; BioLegend). After incubation, cells were washed twice with 3 ml FACS buffer and fixed in 2% paraformaldehyde before flow cytometry.

For each Ab/bead conjugate, 50 μl of UltraComp eBeads Plus compensation beads (ThermoFisher Scientific) was added to 0.125 μg of control or Ag-specific Ab and incubated for 15 min on ice. A total of 2 ml of FACS buffer was added, and beads were centrifuged at 400 × g for 5 min at 4°C. Supernatant was aspirated and beads resuspended in 50 μl of FACS buffer and stored at 4°C. For use, 5 μl of beads was mixed with 45 μl of FACS buffer containing 0.25 pmol tetramer and incubated for 25 min on ice. After the incubation, ∼4 ml of FACS buffer was added, and the samples were centrifuged at 400 × g for 5 min at 4°C. The supernatant was discarded, and the beads were resuspended in 0.2 ml FACS buffer for flow cytometry.

For bead cocktails, for each Ab/bead conjugate, one drop of UltraComp eBeads compensation beads was added to 0.125 μg of control or Ag-specific Abs with 0.125 μg of fluorochrome-conjugated Abs. A total of 4 ml of FACS buffer was added, and beads were centrifuged at 400 × g for 5 min at 4°C. Supernatant was aspirated, and in some cases, beads were resuspended in 50 μl of FACS buffer before the addition of 100 μl of 2% paraformaldehyde and incubation for 18 min on ice. A total of 4 ml of FACS buffer was added, and beads were centrifuged at 400 × g for 5 min at 4°C. Supernatant was aspirated, and beads were resuspended in 50 μl of FACS buffer and stored at 4°C. For use, 3 μl of each bead type was mixed and brought up to 50 μl with FACS buffer containing 0.25 pmol Ag tetramer or 136 μg/ml peptide:MHC tetramer and incubated for 25 min on ice. After the incubation, ∼4 ml of FACS buffer was added, and the samples were centrifuged at 400 × g for 5 min at 4°C. The supernatant was discarded, and the beads were resuspended in 0.2 ml FACS buffer for flow cytometry.

Control Abs used in this study were anti-PE (PE001; BioLegend), anti-allophycocyanin (APC003 [BioLegend] or eBioAPC-6A2 [ThermoFisher Scientific]), anti-streptavidin (S10D4; ThermoFisher Scientific or Santa Cruz Biotechnology), anti-FITC (FIT-22; BioLegend), and anti-6×HIS tag DyLight680 (HIS.H8; ThermoFisher Scientific). A08 was identified via single-cell Ab sequencing of human PBMC B cells and did not bind RBD when expressed as a secreted IgG1. CV30 is an RBD-specific Ab characterized previously (22) and generously provided by L. Stamatatos. The Abs 204, 208, 211, and 215 are anti-RBD–specific Abs that were characterized previously (30) and generously provided by M. Pepper and J. Netland. MHC-specific Abs used in this study were anti-human β2m (A17082A; BioLegend), anti-human HLA-A2 (BB7.2; BioLegend), anti-human HLA-ABC (W6/32; BioLegend), and anti-mouse H-2K^b (Y3; BioXCell).

Fluorochrome-labeled Abs used for bead multiplexing included anti-human CD19 BUV395 (SJ25C1; BD Biosciences), anti-human CD19 BUV496 (SJ25C1; BD Biosciences), anti-human CD19 BUV563 (SJ25C1; BD Biosciences), anti-human CD79b BUV661 (HM79b; BD Biosciences), anti-human CD20 BUV737 (2H7; BD Biosciences), anti-human CD20 BUV805 (2H7; BD Biosciences), anti-human CD19 BV421 (HIB19; BD Biosciences), anti-human CD45 BV510 (HI30; BD Biosciences), anti-mouse CD43 BV605 (S7; BD Biosciences), anti-mouse CD93 BV650 (AA4.1; BD Biosciences), anti-human CD3 BV711 (UCHT1; BD Biosciences), anti-human CD19 BV786 (2H7; BD Biosciences), anti-human CD19 FITC (HIB19; BioLegend), anti-mouse B220/CD45R PerCP (RA3-6B2; BD Biosciences), anti-human CD10 PerCP-Cy5.5 (HI10a; BD Biosciences), anti-mouse CD4 BV711 (RM4-5; BioLegend), anti-mouse CD4 BV421 (GK1.5; BioLegend), anti-mouse CD103 allophycocyanin and BV510 (2E7; BioLegend), anti-mouse Thy1.2 PerCP (53-2.1; BioLegend), anti-mouse CD197 PE (4B12; BioLegend), anti-mouse CD8a BV605 (53-6.7; BioLegend), and anti-mouse PD-1 PE-Cy7 (29F.1; BioLegend).

The figures were generated using BioRender, FlowJo 10 (Becton Dickinson & Company), Prism 9 (Dotmatics), and Illustrator 2021 (Adobe).

As an initial assessment, we tested whether different batches of SARS-CoV-2 RBD tetramers would bind the same frequency of murine B cells. These different batches were made using the same protocol and either used the same day or stored for up to 42 wk before use. For this, spleen and lymph nodes were pooled from multiple animals and divided into three technical replicates for each tetramer batch to gauge technical variability (Fig. 1A). Cells were next incubated with SARS-CoV-2 RBD allophycocyanin tetramers and a control allophycocyanin 755 tetramer that allows for cells binding streptavidin-allophycocyanin to be excluded from the analysis. After incubation, tetramer-binding B cells were enriched from the sample using anti-allophycocyanin microbeads to allow for more robust visualization. Using this approach, ∼9% of B cells in the allophycocyanin-enriched fractions bound to SARS-CoV-2 RBD allophycocyanin tetramer batches made the day of the experiments (Fig. 1B, 1C), which amounted to ∼0.08% of B cells in samples without enrichment (Fig. 1D). Likewise, a similar frequency of cells can be found using some SARS-CoV-2 RBD allophycocyanin tetramers produced as many as 42 wk earlier when stored at −20°C in 50% glycerol (Fig. 1D, 1E). Tetramers stored at 4°C in 1× DPBS performed poorly after only a few weeks, as demonstrated by >30-fold fewer cells binding to tetramers from these batches. Importantly, however, storage at −20°C in 50% glycerol did not maintain consistent cell binding of tetramer for all batches (Fig. 1E). Analysis of batch 7G revealed that 0.06% of B cells bound this tetramer 6 wk after production, which was at the lower end of the range found with fresh tetramers (Fig. 1E). However, this frequency declined to 0.015% at 17 wk after production and 0.0031% of cells 29 wk after production (Fig. 1E). Batch 1G also exhibited a decline in cell binding over time, albeit at a rate slower than that of 7G (Fig. 1E). Together, these data indicate that although storage conditions can help to maintain tetramer performance over time, assessment of these tools is still important to ensure performance.

Given the laborious nature of using cell samples to validate tetramer performance, we aimed to develop an easy and robust assay that could be used before, or within, every experiment. For this, we expanded on previously developed bead-based assays (31, 32). For this, eComp Ultra Plus compensation beads were incubated with mAbs specific for SARS-CoV-2 RBD, positive control Abs specific for allophycocyanin, or negative control Abs specific for PE (Fig. 2A). After incubation with Abs, the beads were washed and incubated with tetramer before analysis. Using this approach, we found that RBD allophycocyanin tetramer batch 2G was bound strongly by both anti-allophycocyanin and the anti-SARS-CoV-2 RBD clone S309 (33), but not anti-PE (Fig. 2B). In contrast, tetramer batch 2P was not bound by S309 but maintained binding by anti-allophycocyanin (Fig. 2B). Together, these results suggested that a bead-based approach could be used to assess the performance of Ag tetramers.

We next reasoned that assessing tetramer binding by a single Ag-specific Ab clone was not enough to fully validate tetramer performance, and we expanded the analysis to include additional Ab clones with varied epitope specificities and binding characteristics. For this, we used a combination of RBD-specific Abs identified by others previously [S309, CV30 (22), 204, 208, 211, 215 (30)], as well as Abs we identified specific for a SARS-CoV-2 spike Ag. For the latter, allophycocyanin tetramers containing a stabilized spike trimer called S2P (34) were used to identify SARS-CoV-2 spike-specific B cells in allophycocyanin-enriched PBMCs from convalescent individuals (Fig. 3A). From this screen, six S2P-binding Abs, named E1, E7, E11, 2B5, 2B8, and 2B11, were identified and used in this study (Fig. 3B). Five of these Abs bound to RBD with notable differences in off-rate measured by BLI (Fig. 3B, 3C). The sixth Ab, 2B5, did not bind RBD (Fig. 3B) and was used as a negative control in subsequent RBD tetramer binding experiments. The mAbs 215 and 208 bound RBD weakly compared with the other Abs (Fig. 3C).

To assess the binding of RBD tetramers to multiple RBD-specific and control Abs, we coincubated aliquots of beads with an Ag-specific or control Ab, as well as an irrelevant fluorochrome-conjugated Ab (Fig. 4A). The fluorochrome-conjugated Ab was included to allow the different Ab-labeled bead populations to be pooled before tetramer labeling and distinguished during analysis (Fig. 4B). Using this approach, we included eight RBD specific Abs, anti-allophycocyanin, anti-PE, and an additional negative control Ab specific for SARS-CoV-2 spike outside of the RBD. After gating on the individual bead populations, we found that different RBD-specific Abs bound different amounts of RBD allophycocyanin tetramers as assessed by geometric mean fluorescence intensity (gMFI). In general, there was no correlation in the level of tetramer binding of the Abs bound to beads (Fig. 4C) and the strength of the binding by BLI (Fig. 3C). This disconnect may be the result of avidity gains because of tetramerization of RBD in the bead assay, but not BLI assessment.

When comparing binding in this assay with multiple batches of tetramers, four distinct profiles emerged. The first profile included most of the tetramer batches that were bound by ∼0.08% of murine B cells in Fig. 1. These tetramers, batches 1G–4G, were bound well by all RBD-specific Abs and anti-allophycocyanin (Fig. 4C). The second profile included batches 5P and 7P, which were the tetramer batches bound poorly by murine B cells in Fig. 1. Both 5P and 7P exhibited ∼100-fold lower binding by all RBD-specific Abs (Fig. 4C), consistent with poor performance in cell-binding assays in Fig. 1. The third profile was displayed by batch 7G, which exhibited ∼10-fold reduction binding to all RBD-specific Abs 19 wk after production (Fig. 4C), which is after this batch began performing poorly in enrichment experiments (Fig. 1E).

The final profile was shown by batches 9G and 5G, which exhibited poor binding by some of the RBD-specific mAbs but was bound well by others (Fig. 4C). Interestingly, no reduction in binding was detected for 215 or 208, and only slight reductions in binding for 2B8 and 204 (Fig. 4C). These Abs also exhibited the poorest binding to RBD when assessed by BLI (Fig. 3C), suggesting that lower-affinity Abs bound batches 9G and 5G well, whereas higher-affinity Abs could not. On assessment of the production of batch 9G, it was discovered that the concentration of Ag used to produce this tetramer was 10-fold higher than the 4:1 Ag/streptavidin ratio normally used to produce tetramers. Because the tetramers used in these experiments were not purified before use, this means that there were ∼36 excess unbound Ag molecules per tetramer molecule contaminating these preparations. Therefore, it appears likely that this unlabeled free Ag was bound by high-affinity Abs, interfering with tetramer binding and detection. Lower-affinity Abs would be expected to be less affected by the presence of free Ag because stable binding is not easily achieved without multimerization. Although we did not discover any production errors for batch 5G, we suspect that a similar issue occurred during preparation. Interestingly, batches 5G and 9G were assessed in cellular experiments and bound ∼0.08% of cells, similar to tetramer batches that were performing well (Fig. 1). To confirm that excess free Ag could block binding of high-affinity Abs, but not low-affinity Abs, we produced batches in which a 40:1 ratio of RBD-biotin to streptavidin-allophycocyanin was compared with a 4:1 ratio. In these experiments, beads loaded with the high-affinity Abs E1 or E7 exhibited reduced binding to the 40:1 batch compared with the 4:1 batch (Fig. 4D). In contrast, beads loaded with the low-affinity Abs 208 or 215 bound similarly to the 40:1 batch and 4:1 batch (Fig. 4D). Together, these results indicate that excess free Ag in tetramer preparations can block the binding of high-affinity Abs, but not binding of lower-affinity Abs expressed by the majority of cells in repertoires with no previous SARS-CoV-2 RBD exposure.

In these experiments, we also explored how probes containing fewer than four Ag molecules would perform compared with their fully loaded tetramer counterparts. For this, we compared probe batches made with RBD-biotin/streptavidin ratios of 0.1:1 and 1:1, compared with properly made 4:1 RBD tetramers. As expected, the beads loaded with the low-affinity Abs 208 and 215 exhibited ∼100-fold loss in binding to the 1:1 batch, which was further reduced at 0.1:1 (Fig. 4D). In contrast, binding by beads loaded with the high-affinity Abs E1 and E7 to the 1:1 batch was only ∼10-fold decreased (Fig. 4D). Together, these data highlight the utility of using a bead-based assay to reveal errors in Ag tetramer production resulting in underloaded probes.

Using a mixture of beads bound to different Abs, we conducted experiments assessing the performance of different batches of tetramers over time under different storage conditions. In total, four batches of RBD allophycocyanin or RBD PE tetramers were produced, split, and stored at −20°C in 50% glycerol, or stored at 4°C in 1× DPBS. Assessing the allophycocyanin tetramers revealed that all batches stored at −20°C in 50% glycerol largely maintained stable binding by RBD-specific Abs over time (Fig. 5). In contrast, RBD allophycocyanin tetramer batches stored at 4°C in 1× DPBS began to exhibit reduced binding by RBD-specific Abs as early as 1–3 wk after production when assessed with CV30, S309, and 2B11 (Fig. 5B). By 6 wk after production, most batches stored at 4°C had 5- to 10-fold reduced binding (Fig. 5B). Abs such as 2B8 and E11 also exhibited a decline in binding with slightly slower kinetics (Fig. 5B). These results highlight the importance of using multiple Abs to assess tetramer performance. Unexpectedly, the batch-to-batch decline in performance varied widely, with batch 2 exhibiting only a moderate loss in binding when stored at 4°C for the 12 wk assessed. Together, these results highlight the importance of using multiple Abs to assess tetramer performance of each tetramer batch.

We initially predicted that the decline in tetramer performance over time when stored at 4°C in 1× DPBS reflected the stability and half-life of the Ag under these conditions. However, this did not appear to be the case, because RBD PE tetramers were bound well by anti-RBD Abs for weeks longer than RBD allophycocyanin tetramers made at the same time (Fig. 5C, 5D). Overall, these data suggest that PE-conjugated tetramers are more stable at 4°C compared with allophycocyanin-conjugated tetramers.

Because allophycocyanin and PE are not the only fluorochromes used in experiments analyzing tetramer-binding cells, we adapted this assay to include additional positive controls. Because Abs specific for every possible fluorochrome are not available, we instead produced positive control beads bound to streptavidin-specific Abs. Compared with allophycocyanin-specific beads, streptavidin-specific beads bound lower levels of wild-type and omicron RBD allophycocyanin tetramers (Fig. 6A, 6B), but at a level sufficient to serve as a positive control for omicron RBD BV421 and BV650 tetramers (Fig. 6C, 6D). As an additional control in these experiments, we also included beads specific to the 6×-HIS tag included in the RBDs. HIS-specific beads exhibited reduced binding to both wild-type and omicron RBD allophycocyanin tetramers compared with streptavidin-specific beads, but binding was clearly detectable above the level of binding by the negative control PE-specific beads.

The generation of fluorescent multimerized peptide:MHC molecules (35) was a major advance in the detection of Ag-specific T cells. Many peptide:MHC tetramers are commercially available, but these tools are also commonly made by research laboratories. From both sources these tools can degrade over time to varying degrees. Testing tetramer integrity can be difficult, particularly in the case of human HLA tetramers, where clinical samples with a reliable target T cell population may be limiting, such as in the case of nonimmunodominant epitopes or cancer neoantigens. Given this, we next assessed whether our multiplexed bead-based assay could be used to test peptide:MHC tetramer stability. For peptide:MHCI monomers to retain stability, they must include three components: β2m, an α chain molecule, and an 8- to 10-aa peptide. A bead-based assay using MHC-specific Abs has been used to assess tetramer integrity; however, this relied on noncommercially available beads and has not been adapted to assess multiple components of the peptide:MHC complex in one reaction (31). To test whether our multiplex bead-based assay could be adapted for MHC tetramers, we used an HLA-A*02 PE tetramer loaded with the YLQPRTFLL peptide from SARS-CoV-2 spike (SARSCov2_YLQ) (36). SARSCov2_YLQ:HLA-A*02 monomers were made in-house by constructing HLA-A*02 monomer folded with a UV labile peptide as previously described (29, 37, 38). UV labile peptide was exchanged for the desired peptide under UV light; then the resulting monomer was tetramerized with streptavidin-PE. We incubated UltraComp beads with Abs specific for components of the peptide:MHC tetramer complex, including α chain (HLA-A2 or pan HLA-ABC), human β2m, anti-PE as a positive control, and anti-FITC as a negative control (Fig. 7A). As described for Ag tetramers, to enable multiplexing of these bead–Ab conjugates, we included irrelevant fluorochrome-conjugated Abs to allow the different Ab-labeled bead populations to be pooled before SARSCov2_YLQ PE tetramer labeling and distinguished during analysis (Fig. 7A). After gating on individual bead populations, we observed high PE gMFI on beads specific for PE, β2m, and HLA-A2, but not negative control beads specific for FITC (Fig. 7B, 7C). To confirm this assay can also be used for mouse MHCI tetramers, we incubated H-2K^b allophycocyanin tetramer loaded with the B8R peptide from vaccinia virus (H2K^b-B8R) (39) with bead–Ab conjugates loaded with Abs against murine H-2K^b, anti-streptavidin, or anti-allophycocyanin as positive controls and anti-FITC as a negative control. Consistent with results from the human SARSCov2_YLQ:HLA-A2 test, we observed strong binding of H2K^b-B8R allophycocyanin tetramer to H-2K^b-specific beads, streptavidin-specific beads, and allophycocyanin-specific beads, but not FITC-specific beads (Fig. 7D). Together, these results indicate that this assay can be adapted for the assessment of human and mouse MHC tetramers.

To assess batch-to-batch variability and stability over time, we used the multiplexed bead assay to compare different batches of HLA-A*02 PE tetramer loaded with the immunodominant peptide GLCTLVAML from EBV (EBV_GLC). These different tetramer batches were made using the same batch of UV labile peptide:MHC monomer, but UV peptide exchange and tetramerization were performed 82, 14, or 6 wk earlier. The results showed a clear difference in the binding ability of tetramer made 82 wk prior compared with more recently made tetramers. Specifically, we observed a reduction in PE gMFI with anti–HLA-A2, -β2m, and –HLA-ABC Abs, indicating a decline in tetramer integrity over time (Fig. 7E). Interestingly, we observed differences in the two recently made tetramers that did not track with time postproduction; Tetramer produced 14 wk prior showed a stronger binding profile than the 6-wk-old batch (Fig. 7E). This indicates that in addition to decay over time, there is likely batch-to-batch variability contributing to tetramer integrity.

To test whether the difference in binding abilities observed with beads correlates with CD8⁺ T cell binding ability, we stained a sample of PBMCs from an EBV⁺ donor with the different batches of EBV_GLC-PE tetramer. Similar to the bead-based assay, the tetramer batch made 82 wk prior exhibited poor performance with a very low frequency of EBV-specific T cells detected. In addition, whereas the 6- and 14-wk-old batches showed overall improved binding in cell-based assays, the 14-wk-old batch had a brighter, more distinct population of tetramer-specific T cells compared with the 6-wk-old batch, supporting the observed profile in the bead-based assay (Fig. 7F). This difference may indicate an error in tetramer production with the 6-wk-old batch resulting in poor binding ability. Together, these data demonstrate the feasibility of the multiplexed bead–based assay to test the performance of peptide:MHC tetramers.

In this study, we have developed an easy and robust assay for assessing the performance of Ag tetramers with the goal of minimizing technical variability and preventing failed experiments caused by unreliable and suboptimal tools. The assay uses commercially available beads typically used to generate single-color controls for compensation bound to mAbs specific for the Ag in the tetramer, or MHC complex components, and control Abs specific for the fluorochrome or streptavidin in the tetramer. By colabeling the beads with different fluorochrome-labeled Abs, different populations of beads can be multiplexed into a single tube and each population identified during analysis. This allows for multiple Ag-specific and control Abs to be used in this assay, allowing for coverage of a wide range of binding modalities, epitopes, and in the case of MHC tetramers, components of the peptide:MHC complex. Together, this assay allowed us to detect both improperly produced tetramers and tetramers that performed poorly over time.

Moving forward, an assay such as we have described could be standardized to help minimize laboratory-to-laboratory differences in reagent quality and ensure reproducibility in similar experiments. Large research consortiums would be well suited to standardize the Ab screening panels for commonly used Ags, as well as controls for normalization of data generated on different flow cytometers by different researchers. Reagent supply companies could also aid this effort by producing Ab-binding beads preloaded with fluorochromes and/or premade cocktails targeting commonly used Ags or peptide:MHC components as opposed to the versions we produced in-house. Together, these improvements would allow for standardized thresholds or guidelines to be used across the field.

One limitation of the described assay is that it relies on the availability of previously characterized Abs specific for the Ag of interest. For newly developed or understudied Ags, these may not be readily available. However, given that Ab cloning is often an early step for tetramer validation, we would encourage researchers to use these cloned Abs to assess future tetramer batches to ensure consistency with early studies. Alternatively, it may also be possible to assess tetramer performance using polyclonal Ag–specific Abs purified from serum.

For MHC tetramer validation, we expanded on a published bead-based assay previously used to assess tetramer performance, but which has not been widely adapted (31). Our use of commercially available beads and multiplexing provides a fast and cost-effective method that may facilitate more widespread adoption of such assays for MHC tetramer validation in the future. Although we focused here on MHCI tetramers, the same approaches could be adapted to test MHC class II tetramer integrity using commercially available Abs against human (HLA-DQ, -DR, and -DP) and mouse MHC class II (I-A and I-E). Although not tested in this study, Abs for specific peptide:MHC combinations, which can be generated by isolating B cells specific for peptide:MHC (40), may prove even more reliable in testing peptide:MHC tetramer integrity and would enable further multiplexing to test many different tetramers in one reaction.

The authors have no financial conflicts of interest.

We thank L. Stamatatos for CV30; M. Pepper and J. Netland for 204, 208, 211, and 215; M. J. McElrath for PBMCs from Seattle Area Control cohort; D. Koelle and A. Wald for COVID-19 PBMC samples from NCT04338360 and NCT04344977; B. Graham for SARS-CoV-2 S2P plasmid; Fred Hutchinson Flow Cytometry and Comparative Medicine Shared Resource staff for technical assistance; M. Lopez-Bernal, L. Yates, R. Putnam, and M. Gurtovnik for administrative assistance; S. Voght for manuscript editing; the bioMT Molecular Tools Core at Dartmouth for help with peptide:MHC monomer production; E. Ferris and M. Cole for use of their FPLC; the National Institutes of Health Tetramer Core for E. coli expressing HLA-A2 α chain; D. Masopust for E. coli expressing mouse H2K^b and human β2m; and E. Newell for helpful discussions regarding MHC tetramer production.