Abstract
TCR affinity for peptide MHC dictates the functional efficiency of T cells and their propensity to differentiate into effectors and form memory. However, in the context of chronic infections, it is unclear what the overall profile of TCR affinity for Ag is and if it differs from acute infections. Using the comprehensive affinity analysis provided by the two-dimensional micropipette adhesion frequency assay and the common indirect affinity evaluation methods of MHC class II tetramer and functional avidity, we tracked IAb GP61–80–specific cells in the mouse model of acute (Armstrong) and chronic (clone 13) lymphocytic choriomeningitis virus infection. In each response, we show CD4 T cell population affinity peaks at the effector phase and declines with memory. Of interest, the range and average relative two-dimensional affinity was equivalent between acute and chronic infection, indicating chronic Ag exposure did not skew TCR affinity. In contrast, functional and tetramer avidity measurements revealed divergent results and lacked a consistent correlation with TCR affinity. Our findings highlight that the immune system maintains a diverse range in TCR affinity even under the pressures of chronic Ag stimulation.
This article is featured in In This Issue, p.3
Introduction
One key parameter regulating T cell activation and functional differentiation in the CD4 T cell response is TCR affinity for Ag. Affinity between receptor and ligand has often been determined using surface plasmon resonance (SPR) measurements (1, 2). However, the need for soluble forms of TCR and peptide MHC (pMHC) has made the use of this assay impractical for tracking the affinity of large numbers of Ag-specific polyclonal T cells in an ongoing immune response. As a result, insights into TCR–pMHC affinity and how it instructs the T cell response during an infection have relied on the use of monoclonal TCRs and altered peptide ligands with SPR-defined affinities (3–5). Polyclonal TCR affinity analysis, in contrast, has depended on ex vivo functional avidity and pMHC tetramer staining assays for indirect estimation of TCR affinity based on the positive correlation between these methods and SPR affinity measurements in monoclonal TCR systems (6–10). Hence, in functional avidity assays, T cells able to mount functional responses by cytokine production, proliferation, or cytotoxic activity in response to low-dose Ag have generally possessed TCRs with high affinity for Ag (1, 3, 11). Similarly, the increased staining or avidity of T cell clones for tetramerized pMHC has been correlated to the inherently high-affinity interaction between monomeric TCR and pMHC (12).
During an immune response, acute Ag exposure models that have examined both monoclonal and polyclonal populations have demonstrated that T cell clones with increased avidity for tetramer and a high functional avidity are preferentially expanded in primary and secondary responses and are selected to become memory cells (6, 7, 9, 13–15). The observed narrowing of the affinity diversity of Ag-specific cells to preferentially enrich for high-affinity T cell clones has been equated to a form of T cell affinity/avidity maturation (16). In contrast, experiments in chronic infections have demonstrated the loss of high-avidity clones and later enrichment of lower-avidity cells, suggesting a decrease in TCR affinity under continuous Ag experience and selection for a T cell affinity profile distinct from the one generated during an acute response (17, 18). As these observations stem from unrelated models that have yet to probe the same Ag specificity and affinity evolution under acute and chronic Ag exposure, it is unclear if affinity skewing actually differs under these conditions. Furthermore, tetramer and functional avidity assays have shown bias toward sampling the highest-affinity fraction of the Ag-specific repertoire, potentially missing the full breadth and diversity in a polyclonal response to infection (19, 20).
Re-evaluation of T cell affinity profiles using the more comprehensive analysis of affinity provided by the two-dimensional micropipette adhesion frequency assay (2D-MP) has now shifted our understanding of TCR affinity breadth and the prevalence and contribution of high- and low-affinity T cells during a polyclonal immune response (19, 21–23). This microscopy-based assay measures monomeric TCR–pMHC affinity at the single-cell level, with the TCR anchored in its natural T cell membrane context and pMHC coated on a surrogate APC, hence providing a two-dimensional (2D) affinity analysis highly predictive of T cell function (24–27). Mounting data highlight that low-affinity and tetramer-negative CD4 clones participate in the immune response, form functional memory, and can compose a larger portion of the Ag-specific compartment for a given epitope (3, 19, 23, 28, 29). Affinities ranging from 100- to 1000-fold have been shown for various Ags in differing immune responses, indicating the immune system maintains a wide breadth of affinities, with all possessing the capacity to undergo clonal expansion and form memory (19, 21–23, 27). Although skewing toward high or low affinities has been noted in some models (19, 23), 2D affinity characterization of a polyclonal CD4 T cell response to chronic infection is lacking. In comparison with an acute infection, understanding how the chronic infection environment shifts and shapes the host’s Ag-specific CD4 T cell populations’ TCR affinity diversity can prove beneficial in CD4 T cell therapies aimed at rescuing immune responses in chronic infections (30, 31).
To understand the evolution of T cell affinity and functional responses during infection, we directly compared polyclonal CD4 TCR affinity to the same MHC class II (IAb) restricted GP61–80 epitope in the well-studied lymphocytic choriomeningitis virus (LCMV) model of acute Armstrong (ARM) and chronic clone 13 (CL13) infections. Although they lead to different infection outcomes, the two viruses share CD4 and CD8 T cell epitopes, allowing for a direct comparison of T cell responses across the two models (32). As the majority of the CD4 T cells target the GP61–80 peptide, with a lower frequency of cells being specific to other minor epitopes, we focused our studies on this immunodominant response (33). 2D-MP, TCR tetramer avidity, tetramer half-life, and functional avidity measurements were used to compare the biophysical TCR–pMHC binding and functional characteristics of CD4 T cells as they transitioned from peak effectors to memory cells. Our overall findings demonstrate CD4 TCR affinity diversity is equally maintained under acute and chronic infection with early effectors and late memory CD4 T cells in acute infection having 2D affinities identical to their chronic infection counterparts. Despite the dominance of high-affinity cells at the peak of the response, in both acute and chronic infection, overall affinity decreased at memory paralleling Ag clearance. Functional and peptide MHC class II (pMHC II) tetramer avidity and half-life measurements lacked a consistent correlation to 2D affinity measurements, confirming TCR affinity contributes to but is not the sole parameter readout by tetramer and functional avidity assays. As affinity skewing of the CD4 T cell response is similar between acute and chronic infection, our data indicate that other regulators modulate T cell function in chronic infection to limit immune pathology without altering affinity diversity in the CD4 T cell response. Hence, therapeutic interventions may be able to recover CD4 T cell responses without the need to increase the breadth of TCR affinity.
Materials and Methods
Mice and virus infection
C57BL/6 (B6) mice were purchased from the National Cancer Institute and Charles River Laboratories. B cell–deficient (Ighm−/−) (34) mice on the B6 background were purchased from The Jackson Laboratory. For acute and chronic LCMV infection, 6–10-wk-old female mice were i.p. injected with 2 × 105 PFU of ARM or i.v. infected with 2 × 106 PFU of CL13, respectively (32, 35–37). Virus stocks were kindly provided by Dr. R. Ahmed’s laboratory at Emory University, Atlanta, Georgia. All animals were housed at the Emory University Department of Animal Resources facility, and all experiments were performed in accordance with the guidelines for the Care and Use of Laboratory Animals under Emory University Institutional Animal Care and Use Committee–approved protocols.
Intracellular cytokine staining: functional avidity
Splenocytes isolated from infected mice were plated at 2 × 106 cells per tested GP61–80 peptide concentration ([GLKGPDIYKGVYQFKSVEFD] synthesized on a Prelude peptide synthesizer [Protein Technologies]). Tested concentrations ranged from 100 μM to 1 nM at 10-fold dilutions. Cells were incubated for 6 h at 37°C in T cell culture media and 5% CO2 in the presence of 10 μg/ml Brefeldin A (MP Biomedicals). T cell media contained RPMI 1640 (Mediatech), 10% heat-inactivated FBS (HyClone), 10 mM HEPES buffer (Mediatech), 2 mM l50, the frequency of cytokine-positive cells (with frequency of unstimulated control subtracted) at the 100 μM concentration was used as the maximal cytokine producer frequency and equated to a 100% with the remaining frequencies at lower peptide doses normalized to this maximal response, as previously described (14, 38). Normalized values were graphed against log-transformed peptide concentrations and the data fitted to a nonlinear regression (log [agonist] versus normalized response) using GraphPad Prism 7 analysis software.
TCR staining
hi (Ag-experienced) and CD4+CD44loCD62+ (naive) cells were quantified and compared between the two infections per experiment. QuantiBRITE PE quantification beads (BD Biosciences) were used per manufacturer instructions to determine the number of TCRs per cell. For further comparison between infections and across different time points, MFI of Ag-experienced cells was normalized to the naive population within the same sample and quantified as the percentage of naive ([Ag-experienced MFI/naive MFI] × 100) (39).
Tetramer avidity
Tetramer staining was performed with splenocytes at a density of 2 × 106 cells (100 μl volume) per tested concentration (10–0.01 μg/ml at 10-fold dilutions, 5–0.05 μg/ml at 10-fold dilutions, 2.5, 0.25 μg/ml) of the PE-conjugated IAb GP66–77 tetramer acquired from the National Institutes of Health Tetramer Core Facility at Emory University, Atlanta, Georgia. Labeling was done in T cell media at room temperature for 1 h. As control, IAb hCLIP103–11750, the frequency of tetramer-stained cells at 10 μg/ml concentration was used as the maximal response and equated to 100% with the remaining frequencies at lower tetramer doses normalized to this maximal frequency, as previously described (14, 18, 38). Normalized values were graphed against log-transformed tetramer concentrations, and the data were fitted to a nonlinear regression (log [agonist] versus normalized response) using GraphPad Prism 7 analysis software.
Tetramer decay
Splenocytes were stained with tetramer at 10 μg/ml for 1 h at room temperature at a cell density of 20 × 106 cells in 1 ml of FACS staining buffer (0.05% sodium azide). Cells were washed with ice-cold buffer to remove excess tetramer and kept on ice until an aliquot (2 × 106b hCLIP103–117 staining used to draw positive gate) was normalized to tetramer staining intensity at time zero ([MFI at time (y)/MFI at zero] × 100), and the normalized MFI was graphed against time. The data were fitted to a one-phase exponential decay curve using Prism 7 analysis software (GraphPad), and the half-life was determined accordingly (40, 41).
2D-MP
The relative 2D affinity of polyclonal IAb GP66–77–specific cells was measured using the previously characterized 2D-MP (19, 26). In this 2D assessment, the frequency of adhesion between ligand- (pMHC on human RBC [hRBC]) and receptor- (TCR on T cell) carrying cells held on opposing micropipettes was observed using an inverted Zeiss microscope. The presence of adhesion was denoted by the interaction-induced stretching of the highly flexible RBC membrane as the two cells were separated after an equilibrium contact time of 2 s. To serve as a surrogate APC, the RBC was first biotinylated (biotin-X-NHS; Calbiochem) then incubated with streptavidin (Thermo Fisher) followed by the addition of biotinylated pMHC monomers (IAb GP66–77 or a control monomer [IAb hCLIP103–117, MTB IAb Ag85B280–294 and IAb ESAT-61–19 and Influenza IAb NP311–325]). T cell samples were prepared from splenocytes by CD4+CD44hi cell enrichment through CD4 T cell purification using EasySep mouse CD4 T cell negative selection kit (Stemcell Technologies) and the simultaneous depletion of CD62L+ cells through the addition of biotinylated anti-CD62L Ab (MEL-14; eBioscience at 10 μg/ml per 1 × 108 cells) to the manufacturer-recommended volume of the isolation mixture. For 2D affinity measurements of tetramer-positive cells, CD4+CD62−-enriched samples were stained with tetramer (described in tetramer avidity methods) and sorted on a FACSAriaII (BD Biosciences Biosciences). To determine relative 2D affinity, 30 independent 2 s contacts were tested per T cell to generate an adhesion frequency value (Pa[2s]). Cells showing an adhesion frequency above 10% at the highest pMHC coating densities (>1000/μm2) were considered Ag reactive [previously identified cutoff (19, 21)]. Cells exhibiting 100% adhesion were further resolved using lower pMHC densities until frequency values below 90% were obtained (27). The adhesion frequency was used to derive the relative 2D affinity of the cell with the following equation: AcKa = −ln (1−Pa[2s])/mrml where mr and ml represent receptor (TCR) and ligand (pMHC) density per area (square micrometers), Pa(2s) is the adhesion frequency at the 2 s equilibrium contact time, Ac is the contact area (kept constant), and AcKa is the 2D affinity (in μm4) (24). TCR and pMHC density per cell was determined using QuantiBRITE PE quantification beads (BD Biosciences Biosciences) per manufacturer instructions and staining of TCR with anti-mouse TCRβ PE Ab (H57-597; BD Biosciences Biosciences) and MHC staining with anti-IA/IE Ab (M5/114/15/2; eBioscience), both at saturating concentrations. Calculations of molecules per area were done by dividing the number of TCR and pMHC per cell by the respective surface areas (hRBC 140 μm2, T cell during assay measured diameter of an individual T cell and the surface area equation of a sphere) (26). A total of 70–400 cells were tested per sample with 20–70 binders/Ag-specific cell used to derive the geometric mean affinity for a given population.
Statistical analysis
Statistical significance of measured values was determined with ordinary one-way ANOVA, Tukey multiple comparison, Sidak multiple comparison, and two-tailed parametric Student t tests with robust regression and outlier removal test (coefficient Q = 1%) used to eliminate outliers, all using the Prism 7 Software (GraphPad). Statistical significance indicated as no significance, *p > 0.05, **p > 0.01, ***p > 0.001, and ****p > 0.0001.
Results
Chronic Ag stimulation leads to T cell dysfunction but maintains tetramer-positive cells at a number comparable to the acute response
In the LCMV B6 model, the ARM virus causes an acute infection that lasts 8–10 d, whereas CL13 persists 40–80 d in serum and spleen but remains in tissues like the brain and kidney (42–44). Chronic Ag exposure leads to CD4 T cell dysfunction early in CL13 infection and results in a faulty memory pool that is unable to mount a secondary response (36, 45). To compare and contrast how the duration of Ag exposure alters the frequency and number of IAb-restricted, GP61–80–specific CD4 T cells, we assessed the TH1 cytokine response and the overall prevalence of pMHC II tetramer-positive cells (i.e., higher-affinity cells) at time points that corresponded to prior- and postviral Ag clearance. CL13 infection led to the increased and sustained expression of the inhibitory and recent Ag experience marker PD-1 on GP61–80–specific IFN-γ–producing TH1 (Fig. 1A) and pMHC II tetramer–positive cells early in the infection (Fig. 1B). PD-1 expression was upregulated at the early effector (day 8 postinfection [d8]) and chronic stage [d35] of the CL13 response, confirming continuous Ag exposure as compared with the acute infection. A marked decrease in PD-1 expression was noted at d120, indicating viral clearance from serum and spleen at this late time point (Fig. 1A, 1B). As is the hallmark of the chronic infection, the dysfunction in the TH1 response was observed with a significant reduction in the frequency and number of IFN-γ–positive (Fig. 1C) and polyfunctional IFN-γ–, TNF-α–, and IL-2–coproducing CD4 T cells (Supplemental Fig. 1), as compared with the more robust acute response (32, 36). However, the two responses had an equal prevalence of epitope-specific tetramer-positive CD4 T cells at the later time points (d35 and d120) and only differed in the number of expanded effectors at d8 (Fig. 1D) (36, 45). Although this suggested that the inhibitory environment in CL13 infection limited early CD4 T cell expansion, the decay in this population occurred relatively slower, with a similar number of tetramer-positive T cells maintained between d8 and d35 and a significant contraction occurring at d120 with Ag clearance (Fig. 1D). In contrast, a continuous decline in tetramer-positive cells was observed as the acute response progressed from the expansion of peak d8 effectors to the formation and maintenance of early (d35) and late (d120) memory cells. Hence, more contraction and culling of Ag-specific cells occurred throughout memory long after clearance of viral Ag. The decay pattern in the TH1 population (Fig. 1C) mirrored the observed decrease in tetramer-positive cell numbers (Fig. 1D) in both acute and chronic infection and confirmed previous observations of time-dependent memory CD4 T cell decline in acute infections (14, 38, 46). As pMHC II tetramers are partial to high-affinity cell detection, these data suggested high-affinity CD4 T cells are retained in the presence of Ag with chronic infection (18) and that their decline in the absence of Ag was not unique to chronic Ag stimulation but occurred in response to acute infection as well.
Chronic Ag stimulation leads to T cell dysfunction but maintains tetramer-positive cells at a number comparable to the acute response. (A) A representative flow plot with the frequency of IFN-γ– and PD-1–expressing cells in ARM and CL13 at the indicated days postinfection gated on CD4+CD44hi splenocytes. (B) A representative flow plot with the frequency of IAb GP66–77 tetramer+– and PD-1–expressing cells under the conditions mentioned in (A). (C) Frequency (left) and log-transformed absolute numbers (right) of IFN-γ–producing cells at the time points and infections represented in the flow plot in (A). (D) Frequency (left) and log-transformed absolute numbers (right) of IAb GP66–77 tetramer+ T cells represented in (B). (C and D) Cumulative data with three to five independent experiments and a total n = 7–19 mice per group at n = 2–5 mice per experiment per group. Bar graphs with mean ± SEM. Statistical significance, ns, no significance, *p > 0.05, **p > 0.01, ***p > 0.001, ****p > 0.0001, Student t test (ARM versus CL13), ordinary one-way ANOVA Tukey multiple comparison test (between days postinfection for individual infections).
CD4 T cell affinity peaks at effector phase and declines equally with memory in acute and chronic infection
As low-affinity cells are expected to participate in the immune response and potentially dominate (18, 23), our observation of a decline in the high-affinity (tetramer-positive) Ag-specific CD4 T cell population in both acute and chronic infection alluded to a possible decrease in the overall affinity of the response in the progression toward viral clearance. To evaluate if affinity declines in the total Ag-specific population, inclusive of tetramer-positive and -negative CD4 T cells, we used the 2D-MP to measure single-cell CD4 TCR affinity. In the 2D-MP assay, the adhesion between TCR and pMHC on micropipette-anchored CD4 T cells and pMHC-coated hRBCs was visually assessed using an inverted microscope (19, 24, 26, 27). Adhesion or binding was evidenced as the elongation of the flexible RBC membrane as the two cells were brought in contact and separated sequentially (Supplemental Fig. 2A). These adhesion or binding events along with surface density of TCR and pMHC were used to generate a relative 2D affinity for individual T cells, and the geometric mean affinity was used to compare different populations. To perform these measurements, we enriched for CD4+CD62L− T cells to maximize the frequency of Ag-experienced CD4+CD44hi cells within the sample (Supplemental Fig. 2B). Determining the adhesion frequency to the IAb GP66–77 monomer as compared with a control non-LCMV monomer tested the LCMV specificity of individual T cells within each sample (Supplemental Fig. 2C). Using this method, high- and low- affinity cells were detected in both the acute and chronic infection at all the time points tested with a >1000-fold affinities represented in each response. However, the peak TCR affinity of the polyclonal response occurred at d8, with both infections showing increased prevalence of higher-affinity CD4 T cells before a decline occurred in the transition to memory time points (Fig. 2A, 2B). Of interest, at both the d8 peak effector and d120 memory time points, the acute (ARM Fig. 2A) and chronic (CL13 Fig. 2B) infections generated an Ag-specific population with equivalent average affinities (ARM; CL13 d8, 1 × 10−4 μm4; 2 × 10−4 μm4 and d120, 8 × 10−6 μm4; 3 × 10−6 μm4). Although the ARM population affinity steadily declined to the d120 time point, the CL13 responders maintained the same high-affinity effector levels between d8 and d35 and only declined in the transition between d35 and d120 (Fig. 2C). Therefore, in the acute and chronic infections, the CD4 T cell population affinity was highest during the effector response with Ag presence, maintaining a larger frequency of higher-affinity T cells, as observed with tetramer staining frequency and numbers. Viral clearance in both infections resulted in a lower-affinity memory CD4 T cell population concurrent with a decline in tetramer-positive T cells.
CD4 T cell affinity peaks at effector phase and declines equally with memory in acute and chronic infection. (A and B) 2D affinity of IAb GP66–77–specific cells in CD4+CD62L−–enriched samples from ARM- (A) and CL13- (B) infected splenocytes, and a comparison of the two infections (C) at the designated days. (D and E) A comparison of tetramer and 2D detected frequency of IAb GP66–77–specific cells in above-mentioned samples from (D) ARM and (E) CL13. (F) 2D affinity of sorted tetramer+ and total (tet+ and tet−) CD4+CD62L− cells from d7 ARM-infected splenocytes. (G) 2D affinity of IAb GP66–77–specific CD4+CD62L− T cells from ARM-infected B cell–deficient (Ighm−/−) mice at d8 and d85 postinfection with p value (0.0522). (H) Comparison of tetramer and 2D frequency in samples from (G). All data representative of two to three independent experiments with splenocytes from two to three mice pooled pre-CD4+CD62L− enrichment per time point and per infection. Affinity data log transformed with (+) sign depicting mean affinity in box and whisker graphs with minimum to maximum range of measured single-cell affinities. Tetramer + high-affinity cell cutoff as a dotted line at 1 × 10−4. Mean + SEM in bar graphs. Statistical significance, ns, no significance, *p > 0.05, **p > 0.01, ****p > 0.0001, (A and B) ordinary one-way ANOVA Tukey multiple comparison test, (C) Sidak multiple comparison test, (D, E, G, and H) Student t test. TET−, tetramer negative; TET+, tetramer positive.
Identification of virus-specific CD4 T cells by 2D-MP revealed a greater frequency of responding T cells than determined by using pMHC II tetramer, as previously noted (19, 23). Employing an in vivo limiting dilution assay along with the Nur77GFP reporter mice with peptide Ag delivered in CFA, we had reported the number and affinity of naive Ag-specific precursors inclusive of lower-affinity T cells for several Ags (23). Lower-affinity, tetramer-negative precursors were found to dominate the naive population for all the tested epitopes, including the LCMV epitope studied in this work. In the naive repertoire, tetramer-negative IAb GP66–77–specific lower-affinity T cells outnumbered tetramer-positive high-affinity precursors by >3-fold and similarly dominated during ARM and CL13 infections, reaching a 7-fold difference at d120 (Fig. 2D, 2E). To validate that tetramer preferentially stains higher-affinity CD4 T cells, we sorted pMHC II tetramer–positive cells from d7 ARM-infected mice to >99% enrichment (Supplemental Fig. 2D) using FACS and measured their 2D affinity. Compared to the total Ag-specific CD4 T cell population, the sorted tetramer-positive T cells had a 10-fold higher affinity and demonstrated a narrowed affinity range (Fig. 2F). Based on this and previous observations from LCMV and myelin oligodendrocyte glycoprotein–specific tetramer-positive cells (19), 1 × 10−4 μm4 was used as the threshold 2D affinity cutoff for tetramer binding, with higher- and lower-affinity TCRs falling above and below this line, respectively. As the population affinity decreased below this threshold during memory (Fig. 2A–C), the increasing disparity between the Ag-specific cell frequencies detected by 2D-MP and tetramer largely occurred from pMHC II tetramer missing lower-affinity TCRs.
In a Friend virus protracted infection model, Ag presentation by activated B cells was necessary for the expansion and overrepresentation of lower-affinity T cell clonotypes and the decay in high-affinity T cells late in the infection (18). To determine if the affinity decline in the LCMV model can also be attributed to a B cell role and thus be abolished in an environment devoid of B cells, we measured 2D affinity of IAb GP66–77–specific CD4 T cells in ARM-infected B cell–deficient mice (Ighm−/−). Similar to the infection of wild type (WT) B6 mice, Ighm−/− mice demonstrated a decrease in average TCR affinity between effector (d8) and memory (d85) T cells (Fig. 2G, trending significance p = 0.0522). 2D-MP identified more Ag-specific cells than pMHC II tetramer during both the effector and memory stages of the response, with the transition from the d8 5-fold frequency difference to the 24-fold change by d85, confirming the increased prevalence of low-affinity cells at the later time point (Fig. 2H). The increased frequency of low-affinity cells at memory, coupled with the drop in 2D affinity, demonstrated the skewing to lower-affinity cells can occur in the absence of B cells. Although this does not rule out a possible role for B cells in the B6 model, it suggested that other mechanisms contribute to the observed decay in affinity.
Tetramer avidity changes in chronic infection in the absence of TCR affinity differences
Overall, the skewing to lower-affinity T cells in the acute response was in opposition to other findings of affinity maturation or selective enrichment of higher-affinity cells at memory (6–9). As these observations were made using pMHC tetramers as surrogate readouts for TCR affinity, we assessed tetramer avidity in the two infections for a comparison with these previous studies and our 2D affinity data. As measures of TCR affinity, 2D-MP and pMHC II tetramer avidity analysis require an accounting of TCR density and total number of receptors per cell (12, 26, 47). Accordingly, we monitored TCR expression levels in acute and chronic infection with the TCR-specific anti-TCRβ (clone H57-597) mAb. As performed with our 2D-MP protocols, we used quantification beads to determine the number of TCRs on CD44hi cells. CD4 T cells in chronic infection had a higher TCR expression at d35, whereas d8 and d120 numbers remained equivalent to ARM responders (Fig. 3A). A lack of TCR downregulation between d8 and d35 in CL13 infection contributed to this observed difference. Of note, TCR numbers decreased past Ag clearance time points within each infection, indicating TCR downregulation can occur independent of Ag presence. This finding was analogous to a study that revealed TCR downregulation as a programmed event set early during Ag encounter but manifesting later in the response despite the absence of Ag (39). Although the contribution of the CD4 coreceptor was found to be minimal in binding to pMHC II (12, 20, 48, 49), we next assessed CD4 expression levels on CD44hi cells relative to the naive population (Fig. 3B). The difference was limited to d8 effectors, with identical expression patterns observed between acute and chronic infections at d35 and d120 (Fig. 3C). Flow cytometry analysis of FSC-A was also used to assess T cell size and identify TCR and CD4 density differences. In both infections, activated T cells were larger at the peak of the response and reduced in size towards memory (Fig. 3D). CL13-specific cells at d8 were larger than their ARM counterparts, whereas d35 and d120 cells were of equivalent size. Although in CL13, infection-exhausted CD8 T cells were previously found to be smaller in size compared with late memory cells (50), CD4 T cell size at d120 was not significantly different between exhausted and memory cells, despite showing a similar trend. The cell size data together with the TCR expression changes suggested potential TCR density differences between early and late time points in both the acute and chronic infection. Although normalized in our 2D-MP measurements, these differences can affect tetramer avidity (12, 39).
TCR expression higher in CL13 infection. (A) TCRβ per cell numbers for CD44hi CD4 T cells from ARM- and CL13-infected splenocytes at the designated days using PE quantification beads. (B) A representative flow plot showing the gating strategy for naive (CD44loCD62L+) and Ag-experienced (CD44hi) cells from total CD4 T cells. (C) Relative CD4 and (D) FSC-A MFI of CD44hi cells normalized to naive MFIs. (A, C, and D) Points representing individual mice (n = 7–10 mice with n = 2–5 mice per experiment per group). MFI of CD44hi cells divided by MFI of naive cells and multiplied by 100 to get the percentage of relative MFI. Mean ± SEM. Statistical significance, ns, no significance, *p > 0.05, ***p > 0.001, ****p > 0.0001, ordinary one-way ANOVA Tukey multiple comparison test (between days postinfection for individual infections), Sidak multiple comparison test (ARM versus CL13).
To measure pMHC II tetramer avidity, we stained splenocytes with decreasing concentrations of IAb GP66–77 tetramer and used IAb CLIP103–117 control tetramer to determine the frequency of cells with specific staining. The frequency at the highest tetramer concentration was defined as the maximal frequency (100%), and the data normalized accordingly and fitted to a dose–response curve (Fig. 4A, 4C) for quantification of tetramer EC50 concentrations (Fig. 4B, 4D). Tetramer avidity remained unchanged between effector and memory cells in the ARM response (Fig. 4A, 4B), whereas in contrast, avidity changes were noted in the chronic infection, with d35 CD4 T cells showing increased tetramer avidity (Fig. 4C, 4D). A comparison of acute and chronic responders showed CD4 T cells in the CL13 response had significantly higher tetramer avidity at all the time points tested (Fig. 4E). Given that the TCR expression difference across the two infections was only limited to d35, the d8 and d120 increased avidity in CL13 suggested a CD4 and TCR number-independent effect; thus, we next assessed 2D affinity. For a fair comparison between 2D affinity and tetramer avidity, we excluded tetramer-negative T cells from our 2D-MP analysis using the tetramer-binding threshold affinity (1 × 10−4 μm4) to group single-cell measurements into high-affinity tetramer binders and tetramer-negative cells. In both acute and chronic infection, the average 2D affinity of Ag-specific cells falling above this threshold remained unchanged throughout each infection and across the two responses (Fig. 4F, 4G). In the ARM infection, this relative 2D affinity mimicked the observed static tetramer avidity, whereas in the CL13 response, tetramer avidity changes occurred despite the equivalent 2D affinities measured at each time point. The lack of TCR number and 2D affinity differences at d8 and d120 suggested the two responses generate a CD4 T cell population with comparable affinities, whereas other affinity- and TCR number–independent factors influenced tetramer avidity (EC50). The CD4 coreceptor’s role in binding pMHC II has previously been shown to be negligible (12, 20, 48, 49) and does not explain differences in avidity and affinity measurements identified in these assays.
Tetramer avidity changes in chronic infection in the absence of TCR affinity differences. Tetramer avidity curves showing the percentage of maximum tetramer+ cells at the different tetramer concentrations used to stain splenocytes from (A) ARM- and (C) CL13-infected mice at the different days postinfection. The frequency of tetramer+ cells at the different doses was divided by the frequency at the highest concentration (10 μg/ml) to calculate the percentage of maximum tetramer binders. Curves fitted to a nonlinear regression (normalized frequency × log concentration − dose–response curve with a variable slope). EC50 tetramer concentrations obtained from the dose–response curves for individual mice from (B) ARM and (D) CL13 infections. (E) Comparison of EC50 values between ARM and CL13 at the different days postinfection. 2D affinity of Ag-specific cells that fall above the tetramer staining affinity cutoff for (F) ARM and (G) CL13 infection and (H) a comparison of the two responses. (A and C) Curves represent the mean ± SEM of n = 12–16 mice. (B, D, and E) Mean ± SEM in bar graphs with n = 8–12 mice with each symbol representing individual mice. (F–H) Pooled sample from two to three mice, affinity data log transformed with (+) sign depicting mean affinity in box and whisker graphs with minimum to maximum range of measured single-cell affinities and tetramer + high-affinity cell cutoff as a dotted line at 1 × 10−4. All data representative of three to five independent experiments at n = 2–5 mice per experiment per group. Statistical significance, ns, no significance, *p > 0.05, **p > 0.01, ***p > 0.001, (B, D, F, and G), ordinary one-way ANOVA Tukey multiple comparison test (between days postinfection for individual infections), (E) Student t test, (H) Sidak multiple comparison test (ARM versus CL13).
Tetramer half-life similar between memory and exhausted CD4 T cells
A direct correlation between TCR affinity and TCR–pMHC interaction half-life has previously been observed, in which high-affinity T cells had a longer interaction duration with pMHC displaying APCs (4, 51). To determine if pMHC II tetramer interaction half-life correlated to avidity or 2D affinity measurements, we next performed tetramer decay assays. After pMHC II staining of samples from acutely or chronically infected mice, tetramer-labeled cells were incubated with an anti-MHCII Ab that binds dissociated tetramer and prevents rebinding to TCR (6, 39, 52). The decay in tetramer staining intensity was measured over time and normalized to the signal detected at time 0. The data were fitted to a one-phase exponential decay nonlinear regression and half-life was determined accordingly (Fig. 5A, 5B) (40, 41). To avoid the inadvertent early skewing to a negative signal (MFI) that occurs with grouping tetramer-positive and -negative polyclonal populations together, decay in tetramer staining intensity was measured for T cells falling in the tetramer-positive gate established using the control tetramer. To validate our analysis, we compared the frequency of tetramer-positive cells detected at every decay time point. The frequency of tetramer-positive cells detected at time 0 was similar to the frequency at all other time points during the assay, indicating the decay in tetramer MFI does not lead to a loss in detection of Ag-specific T cells (Supplemental Fig. 3A–C, data not shown). Using this analysis method, tetramer decay measurements in both the ARM and CL13 response showed a similar interaction half-life across all infection time points (Fig. 5A, 5B, Supplemental Fig. 3D, 3E). Equivalent pMHC II tetramer half-lives were also noted between acute and chronic infection at all points of comparison (Fig. 5C), aligning half-life measurements with 2D affinity but not pMHC II tetramer avidity.
Tetramer half-life similar between memory and exhausted CD4 T cells. Tetramer decay curves (left) and bar graphs of half-lives (right) for CD4 T cells from (A) ARM- and (B) CL13-infected splenocytes and (C) a comparison of the two shown for the different days postinfection. (A and B) Left, Tetramer MFI at decay time points was normalized to time 0 MFI, and the percentage of MFI was fitted to a one-phase exponential decay curve. (A and B) Right, Half-lives derived from curves for individual mice in bar graphs with symbols representing each mouse. Mean ± SEM representative of n = 13–14 mice, three to five independent experiments at n = 3–5 mice per experiment per group. Statistical significance, ns, no significance. (A and B) Ordinary one-way ANOVA Tukey multiple comparison test (between days postinfection for individual infections), (C) Student t test (ARM versus CL13).
Exhausted CD4 T cells have a lower functional avidity compared with memory cells
Functional avidity is often used to infer the affinity of a polyclonal T cell response with a high functional avidity generally predicting the presence of TCRs with a greater affinity for pMHC (3). Furthermore, this ex vivo testing of a virus-specific CD4 T cell populations’ ability to respond to decreasing doses of cognate Ag allows for a comparison of their sensitivity and potential to generate a functional response in vivo (53). In this dose–response assay, the frequency of cytokine-producing CD4 T cells at each dose of cognate Ag was normalized to the frequency observed at the highest peptide concentration (100 μM), and the data were fitted to a nonlinear curve for deriving half-maximal effective concentration (EC50) values. With acute infection (Fig. 6A), IFN-γ–producing GP61–80-specific CD4 T cells showed increased functional avidity in the transition from effectors to early memory but no further increase at late memory, contrary to previous findings (14). A similar functional avidity increase was also detected in CD4 T cells responding to CL13 infection (Fig. 6B). Despite this similar trend, the CD4 T cells in the acute infection had a significantly higher functional avidity compared with the chronic responders as seen with EC50 measurements of IFN-γ (Fig. 6C). Within each infection, the Ag sensitivity seen with IFN-γ producers was recapitulated with IL-2 (Fig. 6D) and TNF-α (data not shown) producers. However, between acute and chronic responders, unlike the IFN-γ avidity difference seen at all time points, IL-2 (Fig. 6D) and TNF-α (data not shown) avidity equalized at the d120 time point. Collectively, the data demonstrated virus-specific CD4 T cells increase sensitivity to Ag in response to acute and chronic infection, but initial activation conditions set degree of sensitivity. Furthermore, this increased sensitivity can occur without an accompanying increase in 2D affinity or pMHC II tetramer avidity.
Exhausted CD4 T cells have a lower functional avidity compared with memory cells. Functional avidity dose–response curves showing the percentage of maximal IFN-γ producers at the different doses of GP61–80 peptide used for ex vivo stimulation of splenocytes from [(A), left] ARM- and [(B), left] CL13-infected mice at the different days postinfection. The frequency of IFN-γ producers at the different doses was divided by the frequency of producers at the highest peptide dose (100 μM) to calculate the percentage of maximal producers. Curves were fitted to a nonlinear regression (normalized frequency × log concentration − dose–response curve with a variable slope). EC50 peptide concentrations derived from the dose–response curves for individual mice are represented in bar graphs for [(A), right] ARM and [(B), right] CL13. (C) Comparison of IFN-γ and (D) IL-2 EC50 values between ARM and CL13 at the different days postinfection. Mean ± SEM (A–C) representative of n = 9–18 mice and (D) n = 7–10 mice. Three to five independent experiments at n = 2–5 mice per experiment per group. Statistical significance, ns, no significance, *p > 0.05, **p > 0.01, ***p > 0.001, ****p > 0.0001 [(A and B) right], ordinary one-way ANOVA Tukey multiple comparison test (between days postinfection for individual infections), (C and D) Student t test (ARM versus CL13).
Discussion
In an infection setting, a T cell’s potential for activation, expansion, cytokine production, and survival as a memory cell is dependent on TCR affinity as well as other inflammatory environment-incited selective pressures (14, 52, 54, 55). In this study, using the 2D-MP assay, we sought to further our understanding of how the Ag-specific CD4 T cell population’s average TCR affinity changes under the pressures of acute and chronic infection in the well-studied LCMV model. In LCMV and other infections, CD4 T cells adept at controlling acute infections demonstrate robust cytokine production and form long-lived memory that protects the host against subsequent infections. However, in chronic infections, as seen with the LCMV mouse model as well as the human pathogens HIV and hepatitis C virus, continuous exposure to high Ag levels render CD4 T cells functionally exhausted and with altered T helper and memory differentiation outcomes (32, 35, 45, 56–59). Although the efficacy of the CD4 response in acute infections has previously been correlated to the enrichment of high-affinity/avidity T cells (6, 9, 14, 38), recent observations in retroviral infection models have demonstrated the chronic response to be enriched in lower-affinity/avidity clones late in the infection (17, 18).
Our data highlight that under acute and chronic infection, CD4 T cell affinity diversity is equally maintained. Both responses expanded peak effector CD4 T cells of equivalent 2D affinities, and the average affinity of the Ag-specific population similarly declined with Ag clearance. Although both high- and low-affinity T cells expanded to infection and later contracted, the relative abundance of each population changed depending on the stage of the immune response. At the peak of the effector response and in the presence of Ag, high-affinity cells were more prevalent (18) in both acute and chronic infection. However, with Ag clearance, increased contraction of high-affinity T cells (tetramer-positive) (60) and a potential outgrowth of lower-affinity (18, 61) clones gave way to a lower-affinity T cell–dominated memory population in both ARM and CL13 infection. As previously noted (19, 23), the 2D-MP detected more Ag-specific cells (2–7-fold higher) in both acute and chronic infection compared with pMHC II tetramer, which selectively identified higher-affinity cells. In evidence of this affinity bias, we found pMHC II tetramer–positive cells had a 10-fold higher average 2D affinity and spanned a narrower affinity range compared with the tetramer-positive and -negative inclusive sample. As the CD4 T cell population affinity declined with memory, more Ag-specific cells were identified using 2D-MP than pMHC II tetramer. At the naive precursor level, a similar skewing to lower-affinity cells was previously observed for GP61–80–specific CD4 T cells, which showed a 3:1 bias (23) before settling in to the observed 7:1 ratio at memory. This would suggest the immune system maintains and resets the ratio of low- to higher-affinity T cells back to this similar preinfection hierarchy that existed prior to infection.
Active mechanisms that promote affinity diversity and maintain lower-affinity cells in the immune system have previously been proposed (8, 18, 62). In the CD4 T cell response to Friend virus protracted infection model, a similar decrease in affinity and enrichment for lower-affinity cells was observed late in the infection (18). Although showing this similar skewing to lower-affinity cells occurred in the chronic LCMV CL13 model, we also observed this phenomenon was not unique to chronic infection, as the acute response shared this progression from a peak affinity at effector time points before a decrease in overall affinity at memory. Of interest, the decline in overall T cell affinity was dependent on B cell Ag presentation in the Friend virus infection model. In our 2D affinity measurements of ARM-infected B cell–deficient Ighm−/− mice, a similar affinity decline between effector and memory cells suggested other T cell intrinsic and extrinsic factors regulated CD4 T cell population affinity. Of note, in a previous characterization of ARM infection in Ighm−/− mice, CD4 T cell memory numbers and cytokine production were found significantly reduced when compared with the WT response (63). Although the poor secondary lymphoid architecture in Ighm−/− mice also contributed toward the deficit in CD4 memory, transient depletion of B cells in WT mice identified a clear B cell role in sustaining CD4 T cell memory numbers during the T cell contraction phase (64). These findings together with our observed affinity decline both in the presence and absence of B cells suggest that these cells maintain memory CD4 T cell numbers as a whole and likely without partiality toward higher- or lower-affinity clones in the LCMV model. Redirection of CD4 T helper differentiation and prevention of immune pathology in Friend virus infection was also dependent on B cells, unlike the LCMV chronic response (58, 59), hence confirming the potential role of other mechanisms in the enrichment of lower-affinity cells in the LCMV model. However, further investigation is needed to identify said mechanisms and the contribution of B cells toward this event without the confounding factors present in Ighm−/− mice. Of interest, the lack of affinity differences between acute and chronic LCMV infection suggested that the shift toward a TFH response in CL13 also occurred independent of a differential TCR affinity skewing. In agreement with our observations, the potential for differentiation into TFH cells has been noted for high– to low–TCR affinity interactions with Ag dose playing a significant role in driving monoclonal T cells down a TFH or alternative differentiation path (55, 58, 65).
The CD27/CD70 costimulatory pathway has previously been reported as a mechanism that ensures memory T cell survival and TCR affinity diversity through maintenance of lower-affinity T cells in the Ag-specific repertoire (8, 62, 66). Characterization of costimulatory molecules on pMHC II tetramer–positive memory and exhausted T cells in the same LCMV models of acute and chronic infection have shown increased CD27 expression on exhausted CD4 T cells (36). Although we also noted CL13-specific early effectors had increased CD27 expression compared with ARM responders, tetramer-positive and total CD44hi cells did not sustain high CD27 expression levels in the progression to lower-affinity memory time points (data not shown). Despite the early difference in CD27 expression, acute and chronic responders demonstrated identical 2D affinities in the high-affinity T cell population (tetramer-positive) at all time points tested. The total population also equally shifted to a lower-affinity population at late memory, making CD27 expression an unlikely mechanism leading to low-affinity T cell enrichment. However, direct approaches that abrogate CD27/CD70 interaction may further clarify if this pathway plays a role in the observed decline in CD4 T cell affinity.
Contrary to our 2D affinity observations, selection into the memory pool has previously been correlated with increased population affinity, as measured by avidity for pMHC II tetramer, longer tetramer binding half-lives, and a higher functional avidity (6, 9, 14, 38). Given the correlation between these measurements and monomeric TCR–pMHC affinity, memory is thought to enrich for higher-affinity/avidity T cells, resulting in T cell affinity maturation (7, 13, 16, 67). Our tetramer avidity measurements in acute and chronic infection revealed neither population enriched for higher-avidity cells in the progression to memory time points. In fact, pMHC II tetramer avidity remained identical in the transition from peak effectors to memory cells in the ARM infection, whereas a higher avidity was measured at d35 in the CL13 response. However, TCR expression difference can be a confounding factor when comparing avidity within each infection. Between ARM and CL13 responders, the similar TCR expression in d8 and d120 samples allowed for a direct comparison of avidity with the data, demonstrating increased avidity in the CL13 response. Our reanalysis of average 2D affinities for T cell populations falling above the 1 × 10−4 2D affinity cutoff for tetramer binding showed affinity to be static in each high-affinity (tetramer-positive) response and equivalent between acute and chronic infection. This suggested TCR affinity–independent mechanisms might be playing a role in the observed tetramer avidity differences. CD4 has no contribution toward binding pMHC II tetramer (12, 48, 49), and expression differences were limited to d8 samples; thus, coreceptor did not explain the increased avidity. Tetramer staining relies on the binding of multiple TCRs to the same pMHC II tetramer complex with one interaction increasing the likelihood of a second TCR–pMHC binding, hence the avidity (10, 40). T cell activation-induced changes in TCR clustering, membrane lipid raft organization, and decreased membrane stiffness or enhanced fluidity can lead to altered tetramer binding capabilities at the T cell surface and could be different in the inflammatory environment with continuous Ag stimulation (68–70). Memory T cell population skewing based on increased pMHC II tetramer interaction half-life and in the absence of a tetramer avidity-based enrichment has previously been noted (52). In our comparisons, tetramer half-life measurements remained equivalent over the course of each response and between acute and chronic responders, showing a correlation to 2D affinity measurements of the high-affinity CD4 T cell population.
As Ag-specific peak TH1 effectors progressed down the differentiation path to early and late memory cells, a stepwise increase in functional avidity was noted in ARM infection (14, 38). Although functional avidity remained similar between early and late memory cells, our data also demonstrated a shift toward increased Ag sensitivity in the transition from peak effectors to early memory cells. The increased Ag sensitivity occurred in the absence of comparable 2D affinity, tetramer avidity, and half-life changes. Other studies of monoclonal and polyclonal responses have also reported a similar disassociation between TCR affinity and functional avidity (14, 37, 67). Under continuous Ag exposure, exhausted TH1 cells also increased functional avidity as the immune response progressed toward Ag clearance, but the degree of sensitivity was significantly reduced as compared with the robust IFN-γ response in ARM infection. Ag dose, the inflammatory environment, and the upregulation of inhibitory receptors in the CL13 response can dampen T cell activation and functional responses (53, 54).
Overall, our 2D affinity findings highlight that CD4 TCR affinity diversity in the Ag-specific polyclonal population is equally maintained under the pressures of acute and chronic infection with both systems expanding CD4 T cell populations of identical 2D affinities. High-affinity T cells dominated peak effector populations, whereas increased prevalence of lower-affinity cells coincided with Ag clearance. A correlation between functional and tetramer avidity measurements and 2D micropipette-based affinity analysis was not consistent, confirming the influence of TCR affinity–independent mechanisms on these assays. Identification of parameters that predict and correlate with the efficacy of a CD4 T cell’s response are critical in tailoring therapies and vaccines toward effectively combating acute and chronic infections. Although our 2D TCR affinity, pMHC II tetramer avidity, and half-life analysis did not differentiate CD4 memory cells from their exhausted counterparts, our data confirmed that increased functional avidity better correlated to the T cell response difference between acute and chronic infection and the generation of functional memory. We now show that in chronic infection, the immune response can maintain a population with an intact TCR affinity distribution that can be targeted by therapies that restore Ag sensitivity and boost CD4 T cell functionality. Future studies are required to identify immune mechanisms that are in place for maintaining this affinity diversity and to further elucidate the role of high- and low-affinity cells within the immune response. Comparisons of as yet unmeasured TCR–pMHC kinetic and biophysical parameters can explain the response difference in the two infections and provide better correlates to protection and future targets for immunotherapies.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Matthew A. Williams and J. Scott Hale for helpful comments on this manuscript. We thank all members of the Evavold laboratory for useful scientific discussions and L.A. Lawrence for maintaining mouse colonies. We thank the Ahmed laboratory for providing virus stocks and the National Institutes of Health Tetramer Core Facility for providing pMHC reagents. We also would like to acknowledge the Emory Children’s Pediatric Research Center flow cytometry core facility for performing cell sorting.
Footnotes
This work was supported by National Institutes of Health Grants R01 NS071518 and R01 AI096879 and National Institutes of Health Training Grant T32 AI007610 (to B.D.E.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- ARM
- Armstrong
- B6
- C57BL/6
- CL13
- clone 13
- 2D
- two-dimensional
- 2D-MP
- two-dimensional micropipette adhesion frequency assay
- FSC-A
- forward scatter-area
- hRBC
- human RBC
- LCMV
- lymphocytic choriomeningitis virus
- MFI
- mean fluorescence intensity
- pMHC
- peptide MHC
- pMHC II
- peptide MHC class II
- SPR
- surface plasmon resonance
- WT
- wild type.
- Received February 27, 2018.
- Accepted April 30, 2018.
- Copyright © 2018 by The American Association of Immunologists, Inc.
References
In this issue
