Abstract
To explore the TCR sensitivity of recent thymic emigrants (RTEs), we triggered T cells with altered peptide ligands (APLs). Upon peptide stimulation in vitro, RTEs exhibited increased TCR signal transduction, and following infection in vivo with APL-expressing bacteria, CD8 RTEs expanded to a greater extent in response to low-affinity Ags than did their mature T cell counterparts. RTEs skewed to short-lived effector cells in response to all APLs but also were characterized by diminished cytokine production. RTEs responding to infection expressed increased levels of VLA-4, with consequent improved entry into inflamed tissue and pathogen clearance. These positive outcomes were offset by the capacity of RTEs to elicit autoimmunity. Overall, salient features of CD8 RTE biology should inform strategies to improve neonatal vaccination and therapies for cancer and HIV, because RTEs make up a large proportion of the T cells in lymphodepleted environments.
Introduction
Development of functional and self-tolerant T cells requires selection in the thymus and further maturation in the periphery. Recent thymic emigrants (RTEs) make up the subset of peripheral T cells that have most recently undergone thymic maturation and egress. RTEs, although an understudied population, contribute significantly to the T cell pool in neonates and during immune reconstitution following HIV infection or chemotherapy, as well as to T cell diversity throughout life.
A mouse model in which RTEs can be readily identified and isolated from their mature naive (MN) T cell counterparts has expanded our understanding of RTE biology (reviewed in Ref. 1). In RAG2p-GFP–transgenic (Tg) mice, GFP is expressed under the control of the RAG2 promoter; although expression is extinguished in mature thymocytes, a residual GFP signal can be detected in a population of peripheral T cells defined as RTEs (2, 3). We and other investigators used RAG2p-GFP–Tg mice to show that RTEs are phenotypically and functionally distinct from mature T cells (2–10), dissimilarities that are mirrored by human RTEs (11, 12). The full extent of these differences remains an open question.
An effective CD8 T cell response to infection requires robust expansion of effector T cells, access to peripheral tissues, activation of appropriate effector functions, such as the production of IFN-γ and TNF-α, and generation of durable memory to protect the host from reinfection (13, 14). The pattern of killer cell lectin-like receptor subfamily G member 1 (KLRG1) and IL-7Rα expression is commonly used (14) to define CD8 T cells as short-lived effector cells (SLECs) or memory precursor effector cells (MPECs). SLECs are terminally differentiated T cells that rapidly contract after the peak of the immune response and contribute minimally to the memory T cell pool. When CD8 T cell responses go awry, autoimmunity, rather than host protection, can result. The ability of RTEs to maintain the proper balance between protection and self-destruction is largely unstudied.
Initiating T cell activation requires TCR recognition of peptide presented by MHC. Membrane proximal events, such as phosphorylation of ZAP70, propagate this signal via an intracellular cascade that results in distal activation of kinases, such as ERK, which help to regulate gene expression to drive the appropriate effector T cell response (15, 16). Recent evidence suggests that the strength of the TCR signal can impact the relative SLEC/MPEC skewing of the CD8 T cell response (17). Altered peptide ligands (APLs), peptide variants that differ in their TCR-binding affinity, can probe monoclonal T cell responses to high- and low-affinity Ags (15). In the OVA-specific OT-I TCR–Tg system, APLs of OVA have been well studied in thymic selection (18) and in the mature T cell response to Listeria monocytogenes expressing APLs of OVA (Lm.APLs) (17, 19).
This study aimed to determine whether the maturation state of peripheral T cells impacts the response to bacterial infection due to intrinsic differences in TCR signaling. We show that RTEs have increased TCR signal transduction and enhanced skewing toward the SLEC compartment during the acute effector response to a pathogen. Independent of T cell fate, RTEs are less able to produce cytokines compared with mature cells, but this dampened function is balanced by the enhanced sensitivity of RTEs to low-affinity ligands, reflected in upregulated CD44 and VLA-4 expression, facilitating RTE migration into inflamed tissues and pathogen clearance. These distinct traits suggest that RTEs are able to clear a broad spectrum of pathogens without causing undue cytokine-mediated destruction, although these advantages can be undermined by an enhanced tendency to autoimmune destruction through efficient tissue localization.
Materials and Methods
Mice
RAG2p-GFP–Tg OT-I TCR–Tg mice were backcrossed in our laboratory for 12 generations onto CD45.1+ or CD45.2+ C57BL/6J (B6) backgrounds. These mice, as well as B6xB6.SJL-PtprcaPepcb/BoyJ mice and B6-Tg(Ins2-TFRC/OVA)296Wehi/WehiJ Tg mice expressing membrane-bound OVA under the control of the rat insulin promoter (RIPmOVA) in pancreatic β cells, were housed under specific pathogen–free conditions and used in accordance with the University of Washington Institutional Animal Care and Use Committee guidelines.
Reagents
19). Vesicular stomatitis virus (VSV) Indiana strain and VSV engineered to express OVA (VSV.OVA; 20) were obtained from M. Bevan (University of Washington).
Cell preparation, flow cytometry, and sorting
Water-lysed single-cell suspensions from spleens and lymph nodes were negatively enriched using CD8 T Cell Isolation Kits (STEMCELL Technologies), and FcRs were blocked (anti-CD16/32; 2.4G2) before sorting to >98% purity as NK1.1−CD4−Ter119−CD11b−B220− (dump gate) CD44loCD62Lhi cells that were either GFP+ (RTEs) or GFP− (MN T cells). On day 7 postinfection, splenocytes were stained for surface markers or intracellular cytokines, as previously described (4). All samples were acquired on an LSR II or FACSCanto (BD Biosciences) and analyzed using FlowJo software (TreeStar).
In vitro peptide stimulation and staining for phosphorylated signaling molecules
Splenocytes from RAG2p-GFP–Tg OT-I TCR–Tg mice were resuspended (3 × 106/50 μl) in 37°C serum-free HBSS, and prewarmed peptide was added at the times indicated in Fig. 1 to a final concentration of 0.1 μM. Cells were fixed in Fix Buffer I (BD Biosciences), permeabilized with Perm Buffer III (BD Biosciences), and stained for both surface and phosphorylated signaling molecules for 30 min at room temperature.
Adoptive transfers, infections, tissue inflammation, and diabetes induction
A total of 104 sorted RTEs and MN T cells was transferred i.v. either separately or as a 1:1 mix into congenic hosts. Lm.APL strains were grown until mid-log phase, and 2000 CFU were injected i.v. into mice 1 d after cell transfer. For inducing ear inflammation or VSV.OVA challenge, mice were sedated with ketamine/xylazine 5.5 d postinfection, and one ear was injected intradermally with 10 μl 1:1 CFA/PBS (Sigma-Aldrich) or 104 PFU VSV.OVA. Equivalent application of PBS or VSV served as respective controls. Where indicated, 100–150 μg anti–VLA-4–blocking Ab (PS/2; University of California, San Francisco mAb Core) was administered i.p. 12 h before and 12 h after CFA treatment. To isolate skin-infiltrating T cells, ears were harvested, minced, and digested three times using 0.14 U/ml Liberase (Roche) in medium for 40–45 min at 37°C. For diabetes induction, 104 sorted RTEs or MN T cells were transferred i.v. into RIPmOVA mice that were infected the following day with Lm.APL. Mice were monitored for blood glucose daily beginning 4–5 d postinfection and were considered diabetic with a reading >350 mg/dl.
Statistics
Unpaired or paired two-tailed Student t tests were used for comparisons, and the log-rank test was used for survival, as indicated.
Results and Discussion
RTEs exhibit increased TCR signal transduction
RTEs and MN T cells from RAG2p-GFP–Tg OT-I TCR–Tg mice were stimulated in vitro with 0.1 μM N4, Q4, or T4 peptide, and phosphorylation of downstream mediators of TCR signaling was measured using flow cytometry (Fig. 1A, 1B). ZAP70 phosphorylation was detected in RTEs stimulated for as little as 5 min with N4, the high-affinity ligand (Fig. 1B, left panels). ZAP70 phosphorylation peaked in N4-stimulated RTEs at ∼75%, whereas mature T cells lagged behind at 26%. Q4 stimulation induced ∼40% of RTEs and 10% of mature T cells to phosphorylate ZAP70. T4, the lowest-affinity ligand tested, generated very little p-ZAP70, but RTEs displayed increased levels over the barely detectable mature T cell response (Fig. 1B). The response kinetics to low- and high-affinity peptides were similar in both populations.
RTEs have heightened TCR signal transduction following in vitro stimulation with high- and low-affinity peptide Ags. RAG2p-GFP–Tg OT-I TCR–Tg splenocytes were stimulated with 0.1 μM peptide. (A) Representative graphs of p-ZAP70 (upper panels) and p-ERK (lower panels) staining following activation of RTEs (solid black line) and MN T cells (dashed line) for 1 h with N4 (left panels), Q4 (middle panels), or T4 (right panels). The “no peptide” control for MN T cells is shown in gray. (B) A full time course for RTE and MN T cell activation with the indicated peptide, as measured by p-ZAP70 (left panels) and p-ERK (right panels) expression. Data are representative of two independent experiments.
ERK phosphorylation was examined to determine whether the enhanced proximal signals in RTEs are propagated downstream. ERK activation after N4 stimulation was detected within 5–10 min (Fig. 1B, right panels). This response peaked in RTEs at 90 min, with 93% of cells phosphorylating ERK compared with 80% of mature T cells (peaking at 120 min). Stimulation with Q4 and T4 resulted in a detectable p-ERK signal at 30 min that was enhanced in RTEs relative to mature T cells at all time points thereafter. Thus, RTEs are better able than mature T cells to transduce both proximal and distal TCR-signaling events upon stimulation with low- and high-affinity ligands. The surprising increase in TCR signal transduction by RTEs suggests that they may be able to respond to a broader range of peptides, as well as to receive increased homeostatic survival signals (21). Although the elongated CDR3 regions of TCRs expressed by RTEs (22) and their elevated TCR/CD3 expression (2) may help to explain this enhanced signal transduction (15), further characterization of molecules that modulate TCR signaling is called for.
RTEs expand more and are skewed toward a SLEC phenotype
We turned to an in vivo model of Lm.APL infection to explore the impact of this increased TCR signal transduction in RTEs. In response to Lm.APL infection, RTEs expanded to a greater extent than did mature T cells in the same inflammatory environment, as measured both by frequency (Fig. 2A, left panels) and absolute number (Fig. 2A, right panels) of transferred cells in the spleen. As expected, mature T cell expansion in response to low-affinity ligands was diminished (19), although, even in this case, RTEs predominated (Fig. 2A). This increased expansion correlated with the skewing of RTEs toward a SLEC phenotype, as determined by KLRG1 and IL-7Rα expression (Fig. 2B, 2C). Lower-affinity ligands drive the generation of a diminished proportion of SLECs and an elevated proportion of MPECs among mature CD8 T cell responders (17); RTEs follow this trend, generating a lower proportion of SLECs in response to infection with L. monocytogenes expressing OVA containing T4 (Lm.T4) compared with L. monocytogenes expressing OVA containing N4 (Lm.N4). However, RTEs still skew to a SLEC phenotype relative to mature T cells responding to low-affinity stimulation (Fig. 2C). Taken together, these data suggest that the enhanced TCR signal transduction in RTEs drives increased cell expansion and skews cell fate toward the SLEC compartment in response to bacterial infection. This SLEC-centric response would be advantageous in a neonatal setting, in which mounting a strong effector response to ligands of a broad range of affinities could help to ensure survival of the infected lymphopenic individual.
RTEs display greater expansion and are skewed to a SLEC phenotype in response to infection with Listeria expressing high- and low-affinity ligands. Sort-purified OT-I TCR–Tg RTEs and MN T cells were cotransferred into congenically marked hosts that were infected the following day with the indicated Lm.APL. Splenocytes were analyzed on day 7 postinfection. (A) Percentage (left panel) and absolute number (right panel) of RTEs and mature T cells. (B) Representative flow cytometric plots of the relative SLEC (KLRG1hiIL7Rαlo) and MPEC (KLRG1loIL7Rαhi) phenotype. Numbers represent the percentage of transferred cells within that quadrant. (C) Proportional SLEC (left) or MPEC (right) phenotype. Data are mean ± SEM compiled from eight independent experiments, n = 19 (N4) or n = 22 (T4). *p < 0.05, ***p < 0.001, paired Student t test.
A lower proportion of CD8 RTEs makes effector cytokines
A significantly lower proportion of RTEs than mature T cells produced the effector cytokines IL-2, TNF-α, and IFN-γ upon Lm.N4 infection (Fig. 3A, 3B). RTEs also exhibited reduced production of IL-2 and TNF-α following Lm.T4 infection, although no notable differences in IFN-γ production between RTE- and MN-derived T cells were apparent. A higher proportion of MPECs produced cytokine, most noticeably so for IL-2 (Fig. 3C, 3D). A lower proportion of RTEs produced cytokines within both the SLEC and MPEC compartments, indicating that, although some differences in cytokine production by RTEs can be attributed to their increased SLEC skewing, cell fate phenotype cannot fully account for these cytokine defects, which are, therefore, intrinsic to RTEs. The lower cytokine production by RTEs may prove beneficial in a neonatal host in which RTEs predominate and unrestrained release of inflammatory cytokines in response to commensal Ags would be problematic (23), although these cells did not produce more of the anti-inflammatory cytokine IL-10 than their mature counterparts (data not shown). It remains to be determined whether the dampened cytokine production by CD8 RTEs, also seen directly ex vivo (9), is a result of increased methylation at cytokine promoter regions, as was shown for CD4 RTEs (7).
Regardless of cell fate, a lower proportion of RTEs produces cytokines in response to bacterial infection. Sort-purified OT-I TCR–Tg RTEs and MN T cells were cotransferred into congenically marked hosts that were infected the following day with the indicated Lm.APL. Splenocytes were analyzed on day 7 postinfection after a 5-h restimulation in vitro with SIINFEKL peptide. (A) Representative flow cytometric plots of cytokine staining; gates were set using cells from an infected recipient that were not stimulated in vitro. Numbers represent the percentage of transferred cells within that quadrant. (B) RTE- or MN-derived cells producing IL-2 (left panel), TNF-α (middle panel), and IFN-γ (right panel). IL-2, TNF-α, and IFN-γ production by RTE- or MN-derived SLECs (C) or MPECs (D). Data are mean ± SEM compiled from eight independent experiments (n = 19 [N4] or 22 [T4]) (B) and two independent experiments (n = 6 [N4 and T4]) (C and D). *p < 0.05, **p < 0.01, ***p < 0.001, paired Student t test.
CD8 RTEs upregulate VLA-4 expression and are better able to enter inflamed tissue
RTE- and MN-derived T cells encountering Lm.N4 highly upregulate VLA-4 expression (measured by the expression of CD49d/α4 and CD29/β1 integrin subunits), with RTEs exhibiting slightly increased expression relative to their mature counterparts (Fig. 4A). This difference is magnified in response to the low-affinity ligand, during which only 75% of mature cells upregulate VLA-4 expression compared with 90% of RTEs (Fig. 4A). The heightened sensitivity of RTEs to low-affinity ligands is also reflected in their enhanced upregulation of CD44 (Fig. 4A).
RTEs demonstrate both increased VLA-4–dependent homing to inflamed tissues and viral clearance in the ear following challenge. Sort-purified OT-I TCR–Tg RTEs and MN T cells were transferred together (A–C) or separately (D) into congenically marked hosts that were infected the following day with the indicated Lm.APL. (A) VLA-4 (α4β1 integrin, left panel) and CD44 (right panel) expression on splenocytes 7 d postinfection. Data are mean ± SEM compiled from eight independent experiments (n = 19–22). (B) Number of cells in the ear on day 7 post-Lm.T4 infection following CFA treatment on day 5.5 and/or anti-VLA-4 (PS/2) administration. Data are mean ± SEM and are representative of two independent experiments. (C) Proportion of cells in the CFA-treated (inflamed) ear corrected for their proportion in the untreated ear of Lm.T4-infected mice. Data are compiled from five independent experiments, with three pooled ears/data point. (D) VSV PFU/ear 24 h following injection of 104 PFU VSV.OVA intradermally in the ear on day 5.5 post-Lm.APL infection in recipients of individually transferred RTEs or MN T cells. Data are mean ± SEM compiled from two or three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, paired or unpaired (D) Student t test.
Knowing that the integrin VLA-4 is important in promoting access to peripheral tissues (24), we induced ear inflammation by CFA administration to evaluate the impact of this disparate VLA-4 expression on effector T cells generated upon Lm.T4 infection. Transferred effector cells were quantified in separately digested inflamed and control ears. The number of transferred cells in the CFA-treated ear increased by >30-fold over that in the control ear (Fig. 4B). T cell homing to the inflamed tissue was largely VLA-4 dependent, being significantly reduced by anti–VLA-4 treatment (Fig. 4B). Most interestingly, RTEs were better able to enter the inflamed ear compared with mature T cells (Fig. 4C). To determine whether this increased invasiveness is protective, RTEs and MN T cells were transferred separately into recipient mice that were then infected with Lm.APL, challenged 5.5 d later in the ear with VSV or VSV.OVA, and assessed for viral PFU in the ear the next day. VSV-challenged recipients of RTEs and mature T cells had comparable virus levels (data not shown). In contrast, mice receiving RTEs had significantly reduced viral titers in the ear following VSV.OVA challenge (Fig. 4D). Therefore, despite the impaired capacity of RTEs to produce effector cytokines, their increased invasiveness allows them to better access sites of inflammation and control viral challenge in an Ag-dependent manner. Thus, efficient tissue localization compensates for the dampened cytokine production by RTEs. It remains to be determined whether the enhanced tissue invasiveness that RTEs demonstrate impacts their relative contribution to the pool of skin-resident memory T cells.
Increased invasiveness of RTEs drives disease when cells recognize self-Ags
To determine whether the enhanced invasiveness of RTEs could contribute to increased autoimmunity, we measured diabetes induction in mice expressing a model pancreatic Ag, in which the diabetogenic capacity of mature CD8 T cells correlates with VLA-4 expression (17). Sort-purified OT-I TCR–Tg RTEs or MN T cells were transferred separately into RIPmOVA hosts that were then infected with Lm.N4 or Lm.T4. All Lm.N4-infected recipients also receiving RTEs rapidly developed diabetes, whereas fewer of the mice receiving MN T cells became diabetic and did so with slower kinetics (Fig. 5A). Infection with the low-affinity Lm.T4 strain drove less diabetes overall, but RTEs still appeared more diabetogenic (Fig. 5B), consistent with their increased VLA-4 expression following stimulation with low-affinity ligand. Thus, the increased invasiveness of RTEs, while contributing to host protection, also can be detrimental when self-tolerance is lost. It is clearly crucial that postthymic maturation be coupled with the induction of tolerance to (often low-affinity) self-Ags whose expression is restricted to the lymphoid periphery. This facet of RTE biology is a focus of current research.
RTEs drive increased diabetes in RIPmOVA recipients following Lm.APL infection. A total of 104 sort-purified OT-I TCR–Tg RTEs or MN T cells was transferred separately into RIPmOVA mice that were infected the following day with Lm.N4 (A) or Lm.T4 (B). Shown is disease incidence in infected RIPmOVA recipients, as measured by a blood glucose reading >350 mg/dl. Numbers in parentheses indicate diabetic/total mice. Data are combined from three independent experiments. *p < 0.05, log-rank test.
To our knowledge, we show for the first time that RTEs demonstrate increased TCR signal transduction, notably even in response to low-affinity ligands, and that this response can result in host protection through better access to peripheral tissues via enhanced VLA-4 expression. This positive outcome is balanced by the risk for increased autoimmunity when the response is directed to self-Ags. These newly described properties that distinguish RTE responses from those of mature cells have important implications for neonatal vaccine design and anticancer therapy, because RTEs make up the majority of the responding T cells in these lymphopenic environments.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank members of the Fink Laboratory and Drs. M. Bevan, M. Pepper, and D. Zehn for discussion, as well as A. Woodward-Davis for advice on the plaque assay.
Footnotes
This work was supported by National Institutes of Heath Grant R01 AI064318 (to P.J.F.) and Predoctoral Training Grants T32 GM07270 and AI06677 (to A.M.B.).
Abbreviations used in this article:
- APL
- altered peptide ligand
- B6
- C57BL/6
- KLRG1
- killer cell lectin-like receptor subfamily G member 1
- Lm.APL
- Listeria monocytogenes expressing APLs of OVA
- Lm.N4
- L. monocytogenes expressing OVA containing N4
- Lm.T4
- L. monocytogenes expressing OVA containing T4
- MN
- mature naive
- MPEC
- memory precursor effector cell
- N4
- SIINFEKL peptide
- Q4
- SIIQFEKL peptide
- RIPmOVA
- Tg mice expressing membrane-bound OVA under the control of the rat insulin promoter
- RTE
- recent thymic emigrant
- SLEC
- short-lived effector cell
- Tg
- transgenic
- T4
- SIITFEKL peptide
- VSV
- vesicular stomatitis virus
- VSV.OVA
- VSV engineered to express OVA.
- Received July 22, 2014.
- Accepted August 4, 2014.
- Copyright © 2014 by The American Association of Immunologists, Inc.
References
In this issue
