Contact with secondary lymphoid organs drives post-thymic T cell maturation

Abstract

T cell development, originally thought to be completed in the thymus, has recently been shown to continue for several weeks in the lymphoid periphery. The forces that drive this peripheral maturation are unclear. The use of mice transgenic for GFP driven by the RAG2 promoter has enabled the ready identification and analysis of recent thymic emigrants (RTEs). Here, we show that RTE maturation is a progressive process and is promoted by T cell exit from the thymus. Further, we show that this maturation occurs within secondary lymphoid organs and does not require extensive lymphocyte recirculation.

Introduction

T cell development has been thought to occur in its entirety in the thymus, giving rise to mature naïve T cells that circulate in the lymphoid periphery, patrolling for foreign antigenic insults. However, recent studies have suggested that to attain full immunocompetence, further development must occur in T cells that have recently entered the peripheral environment. Such recent thymic emigrants (RTEs) are a subset of naïve T cells that help maintain the diversity of the peripheral T cell repertoire, and are of particular importance to recovery from lymphopenia, when thymic output is needed to replenish the peripheral repertoire, and in infants, when RTEs first seed the periphery.

A challenge to studying RTEs is their lack of a unique cell surface marker. Recent work from our lab (1, 2) characterized a new model system that allows unambiguous identification of RTEs from unmanipulated mice, and enables the ready isolation and analysis of their function and phenotype. This system uses mice transgenic (Tg) for GFP under control of the RAG-2 promoter (3). Thymocytes from such RAG2p-GFP Tg mice express high levels of GFP, mirroring endogenous RAG expression. While RAG expression in RAG2p-GFP Tg mice is extinguished by the single positive (SP) stage, a residual, decaying GFP signal remains in cells following thymic egress. Thymectomy studies have indicated that the GFP signal declines with time in the periphery, such that GFP^hi and GFP^lo RTEs have been in the periphery for up to 1 and 2-3 weeks, respectively (1). GFP^- naïve T cells (non-RTEs) have exited the thymus at least 3 weeks previously.

Using this system, we demonstrated that RTEs exhibit a CD24^hiQa2^loCD45RB^loIL-7Rα^loTCR^hiCD3^hiCD28^lo phenotype relative to non-RTEs (1). We also showed that RTEs differ functionally from non-RTEs, exhibiting a dampened response to stimulation, with decreased IL-2 and IFN-γ production, proliferation, and high-affinity IL-2R upregulation (1). These striking differences between RTEs and non-RTEs hint that definable forces may drive cells from RTE status into the mature naïve T cell compartment. We now show that RTE maturation occurs progressively, requires egress from the thymus, and is driven to completion not in the blood, but within secondary lymphoid organs (SLOs).

Materials and Methods

Mice

C57BL/6 mice were from The Jackson Laboratory and RAG2p-GFP Tg mice (3), originally a gift from M. Nussenzweig (The Rockefeller University), were backcrossed in our lab at least 10 generations onto the C57BL/6 background. Mice were used between 6-12 weeks of age. All experiments were performed in compliance with the University of Washington Institutional Animal Care and Use Committee.

Mouse procedures

Mice were splenectomized (4) and thymectomized (1) as described previously. For blockade of thymic and lymph node (LN) egress, mice were i.p. injected daily for 3 or 6 d with 1 μg/g body weight AAL-R, a sphingosine 1-phosphate (S1P) mimetic, or AAL-S (gifts from Novartis), a biologically inactive enantiomer of AAL-R, made up at 1 mg/ml in water plus .25% DMSO. For blockade of LN entry, mice were given 200 μg lab-purified anti-CD62L (clone MEL-14) plus 100 μg anti-Very Late Antigen 4 (VLA-4, clone PS/2), or were given 200 μg IgG2a isotype control Ab (eBioscience) i.p. on d 0, 2, and 4.

Cell preparation, staining, enrichment and sorting

Single cell suspensions of thymus, LNs (brachial, axillary, inguinal, cervical, and mesenteric), and water-lysed blood and splenocytes were prepared and counted. Where noted, cells were labeled for 10 min at 37°C with 4 μm CFSE. For flow cytometric analysis, FcR were blocked with anti-CD16/32 (2.4G2, BD Biosciences), and cells were stained as described (1) with Abs against the following molecules: CD3 (145.2C11), CD4 (RM4-5), CD11c (N418), CD24 (M1/69), CD25 (PC61), CD44 (Pgp-1), CD45.1 (A20), CD45.2 (104), CD45RB (16A), CD62L (MEL-14), CD69 (H1.2F3), Qa2 (1-1-2), and CD90.2 (53-2.1), all from from eBioscience or BD Pharmingen. Biotinylated Abs were detected with allophycocyanin-conjugated streptavidin (eBioscience). Events were collected on a FACSCanto (BD Biosciences) and data were analyzed on FlowJo (TreeStar) after excluding doublets from live-gated samples. Fluorescence-Minus-One (5) samples were run where appropriate. For sorting, untouched T cells were enriched with an EasySep kit (StemCell Technologies), stained to eliminate non-T cell lineages with PE-conjugated anti-CD11b (M1/70), anti-NK1.1 (PK136), anti-B220 (RA3-6B2), and anti-Ter119 (Ly-76) (all from eBioscience or BD Biosciences). Staining with anti-CD62L was used as a positive marker for naïve cells. Cells were sorted on a FACSAria (BD Biosciences) as PE^-CD62L⁺, and either GFP⁺ or GFP^- to >97% purity (for blood sorts, purity was >95% for RTE and >80% for non-RTE).

Quantification of IL-2 secretion and cell proliferation

Per well, 25,000 sorted CD4 T cells were stimulated with 30 ng/ml anti-CD3 and 1 μg/ml anti-CD28 (BD Biosciences) in the presence of 175,000 irradiated splenocytes depleted of T cells by treatment with anti-CD4 (RC172.4R6), anti-CD8 (3.168.8), and anti-CD90.2 (13.4.6) plus rabbit complement (Cedarlane). Cells were cultured in 96-well plates (BD Biosciences) in complete RPMI at 37°C in 7% CO₂. IL-2 secretion was measured in 24-hour supernatants with the OptEIA IL-2 ELISA kit (BD Biosciences). At 48 hours, 1 μCi [³H] thymidine (PerkinElmer) was added per well, and [³H] incorporation was measured after overnight incubation.

Results and Discussion

Alterations in RTE phenotype result from progressive maturation

To determine whether the phenotypic changes that occur in RTEs are due to selective survival and outgrowth of a subpopulation of RTEs bearing a mature CD24^loQa2^hiCD45RB^hi phenotype or to maturation of RTEs on a per-cell basis, we adoptively transferred equal numbers of sorted CD4 RTEs and non-RTEs into lymphoreplete recipients. Comparable numbers of both cell types were recovered from recipient spleens short-term following transfer (Fig. 1A), and at 10% of input, were on a par with the generally accepted engraftment of transferred lymphocytes (6). Downregulation of CD24 and upregulation of Qa2 and CD45RB expression by RTEs was evident during the 7 d time course, even in undivided populations of RTEs (Fig. 1B). Thus, RTE maturation does not require cell division. The gradual decay of GFP in the transferred RTEs (Fig. 1C) is consistent with the estimated half life of GFP in T cells (7). At 7 d post-transfer, > 95% RTEs remained GFP⁺CD44^lo/midCD62L^hi and undivided (Fig. 1C and data not shown), arguing against significant homeostatic proliferation. The percent take and the absence of extensive proliferation together suggest that the phenotypic changes in RTEs result from progressive maturation of the bulk of the RTE population, rather than selective accumulation of a subset of RTEs already expressing a mature surface antigen phenotype.

Figure 1.

RTEs mature progressively in the lymphoid periphery. Sorted populations of CD4 RTEs (GFP⁺CD62L^hi) and naïve non-RTEs (GFP^-CD62L^hi) from RAG2p-GFP Tg mice were transferred to separate congenic lymphoreplete recipients on d 0 (2 × 10⁶ per mouse, >97% purity). At the indicated times thereafter, recipient splenocytes were stained for donor cell analysis. A, Mean donor cell numbers are expressed as a percentage of the number of cells transferred on d 0. Error bars represent SD. Differences were not statistically significant; p > .01, two-tailed Student’s t-test with equal variance. Representative CD24, Qa2, and CD45RB expression in B and GFP and CD44 expression in C by donor cells from T-enriched splenocytes at the indicated times post-transfer. Transferred cells were CFSE-labeled in B, and data collected from CFSE^hi, undivided cells. Data are from 3 recipients of each cell type per time point.

RTE maturation requires thymic egress

To explore whether the RTE maturation that occurs in the periphery is a cell-intrinsic program or one that is triggered by signals from the lymphoid periphery, we sequestered RTEs in the thymus by treating RAG2p-GFP Tg mice with AAL-R (8, 9). AAL-R is a synthetic mimetic of S1P, blocking S1P receptor 1 (S1P₁), the receptor that is required for T cell exit from both the thymus and LNs (9, 10). The resulting RTE “wannabes” were contained within the GFP⁺, developmentally mature (TCR^hiCD62L^hi SP) thymocyte compartments (Fig. 2A). RTE “wannabes” accumulated in the thymus of AAL-R-treated mice, as these most developmentally mature CD4 and CD8 SP compartments increased about 5-fold by percent (Fig. 2B) and about 3-fold by number (data not shown) relative to the thymus of untreated mice or mice treated with AAL-S, the biologically inactive enantiomer of AAL-R (8, 9).

Figure 2.

RTEs require contact with the lymphoid periphery to mature. A, Diagram of experimental conditions. RAG2p-GFP Tg mice were treated daily for 6 d with AAL-R or control AAL-S, and analyzed on d 7. B, Flow cytometric analysis of thymocytes from untreated, AAL-S treated, and AAL-R treated mice, gated on the indicated cell populations. C,D, CD24 expression by GFP⁺CD62L^hiTCR^hi CD4 and CD8 SP thymocytes from AAL-R treated mice (RTE “wannabes”) and GFP⁺CD44^lo/mid splenic CD4 and CD8 T cells from AAL-S treated mice (control RTEs). GFP normalization was performed on RTE “wannabes” and control RTEs are gated on GFP⁺ cells to match the age of these 2 comparison groups (GFP MFI = ∼1300 for CD4 and GFP MFI = ∼625 for CD8 RTE “wannabes” and control CD8 RTEs). Thymocytes and splenocytes from an untreated age-matched mouse were analyzed on the same day for CD24 expression by GFP⁺CD62L^hiTCR^hi SP thymocytes (Mature SP thymocytes) and GFP^-CD44^lo/mid splenic CD4 and CD8 T cells (naïve GFP^- peripheral T). Representative data are shown in C, and data in D are averaged MFIs from 4-5 mice per group from 2 independent experiments, with error bars representing SD. *, p < .01, two-tailed Student’s t-test with equal variance.

GFP level in RTEs is a faithful indicator of their age, as the decrease in GFP brightness correlates well with residence time in the periphery (7). To compensate for the unequal cellular age of RTE “wannabes” and control RTEs, we normalized GFP MFIs, excluding GFP^lo older RTEs from the control group and GFP^bright RTE precursors from the “wannabe” group (7). After GFP normalization, RTE “wannabes” were phenotypically immature relative to control RTEs by the marker CD24 (Fig. 2C,D). These results suggest that RTEs must exit the thymus to fully complete their phenotypic maturation within this time frame.

However, the phenotype of RTE “wannabes” is intermediate, as Qa2 and CD45RB were essentially at mature levels (data not shown). Thus, part of the maturation program may be on autopilot, or the factors driving RTE maturation in the periphery may also be present in the thymus, with lack of full maturation of RTE “wannabes” a result of quantitative rather than qualitative maturation signal differences. This idea is reinforced by the phenotypic maturity of S1P₁-deficient T cells that remain stuck in the thymus (11).

AAL-R (and other S1P mimetics such as FTY720) binds to the S1P₃, S1P₄, and S1P₅ receptors as well as to S1P₁, influencing factors such as heart rate (12). Our results using the S1P₁-specific agonist SEW2871 (13) were comparable to those obtained with AAL-R (data not shown), suggesting that the maturation defects are specific to blockade of thymic egress.

RTE maturation requires access to SLOs

SLOs are the sites where naïve T cells encounter many other cell types, such as DCs, and cytokines, such as IL-7 (14). To test whether RTE maturation takes place in SLOs, we blocked RTE access to SLOs through a combination of splenectomy and administration of anti-CD62L plus anti-VLA-4 to block LN entry, creating “homeless” RTEs (Fig. 3A). LN blockade was successful, as 6 d following initiation of Ab administration, there was a >50-fold reduction in naïve T cell numbers in LNs (data not shown). Phenotypic maturation of GFP normalized “homeless” RTEs relative to control RTEs was impaired for the markers Qa2 (Fig. 3B) and CD45RB (data not shown). When the Qa2 expression level on “homeless” RTEs was normalized to that of control RTEs, there was a statistically significant difference between “homeless” RTEs and control RTEs for both CD4 and CD8 T cells (p < .05, two-tailed Student’s t-test with equal variance). Access to either the spleen or LN compartment alone is sufficient for maturation, as the phenotypic profile of RTEs blocked from either compartment alone matched that of control RTEs (Fig. 3C). Surgical stress and homeostatic proliferation did not influence RTE maturation, as Qa2, CD45RB, and CD44 expression levels in RTEs from splenectomized mice matched those of untreated or Ab-treated mice (Fig. 3C and data not shown). The Abs coating RTEs to block LN entry did not impair maturation, as the phenotypic profile of Ab-coated RTEs in the spleen matches that of uncoated RTEs from the periphery of control mice (Fig. 3C).

Figure 3.

RTEs require SLO access to mature. A, Diagram of experimental conditions. RAG2p-GFP Tg mice were untreated or splenectomized and treated every other day with anti-CD62L plus anti-VLA-4 for 4 d and analyzed on d 6. Other controls included eusplenic mice treated with both Abs and splenectomized mice treated with isotype control Ab. B,C, Representative Qa2 expression by GFP⁺CD44^lo/mid CD4 and CD8 T cells from the blood of splenectomized, anti-CD62L plus anti-VLA-4 treated mice (“Homeless” RTEs), anti-CD62L plus anti-VLA-4 treated euthymic mice (“No LN” RTEs), splenectomized, isotype control Ab treated mice (“No Spl” RTEs), or untreated mice (control RTEs). GFP levels of “homeless” and control RTEs were normalized in B. Thymocytes and splenocytes from an untreated age-matched mouse were analyzed on the same day for Qa2 expression by GFP⁺CD62L^hiTCR^hi SP thymocytes (Mature SP thymocytes) and GFP^-CD44^lo/mid splenic CD4 and CD8 T cells (Naïve GFP^- peripheral T). CD24 expression levels could not be assessed on cells in the blood because of high background staining. Data are representative of at least 5 mice per condition in 2 independent experiments.

Deprivation of SLO access for 3 weeks causes death of much of the naïve T cell compartment due to lack of IL-7 and other cytokines (14). When RTEs were prevented from entering SLOs for 6 d, numbers of “homeless” blood naïve T cells were increased relative to control blood naïve T cells, with a specific increase in the RTE compartment (data not shown). Thus, 6 d deprivation of SLO access leaves the naïve T cell compartment largely intact, but does promote the survival of the younger RTE compartment relative to that of the older non-RTEs. We matched GFP levels to ensure that our phenotypic comparison of “homeless” and control RTEs was limited to cells that spent the same period of time in the periphery (Fig. 3B).

RTE maturation does not require continuous recirculation in the lymphoid periphery

Naïve T cells scan for antigen presented by DCs in SLOs, recirculating to another SLO if antigen is not found within 12-18 hr (15). To test whether RTE maturation requires continuous recirculation in the lymphoid periphery, we compared the phenotype of RTEs that were “stuck” in SLOs for 6 d with that of control RTEs continually recirculating for 6 d (Fig. 4A). Because AAL-R treatment blocks thymic egress, we thymectomized control mice at the onset of the experiment to age-match the RTEs in each group, and thereby matched GFP levels of “stuck” and control RTEs. There was no statistically significant difference in the CD24 or Qa2 MFIs between “stuck” and control RTEs for either CD4 or CD8 T cells (p = .56 to .98, two-tailed Student’s t-test with equal variance), indicating that continuous recirculation is not required for RTE maturation (Fig. 4B). To assess RTEs that did not recirculate extensively prior to AAL-R treatment, we analyzed GFP^hi RTEs from both groups, and again found comparable maturation (data not shown).

Figure 4.

RTE maturation does not require continuous recirculation. A, Diagram of experimental conditions. RAG2p-GFP Tg mice were treated daily for 6 d with AAL-R or AAL-S, and cells were analyzed on d 7. Control mice were thymectomized on d 0. B, CD24 and Qa2 expression by GFP⁺CD44^lo/mid splenic CD4 and CD8 T cells from AAL-R treated mice (“Stuck” RTEs) and AAL-S treated thymectomized mice (Control RTEs). Thymocytes and splenocytes from an untreated age-matched mouse were analyzed on the same day for CD24 and Qa2 expression by GFP⁺CD62L^hiTCR^hi SP thymocytes (Mature SP thymocytes) and GFP^-CD44^lo/mid splenic CD4 and CD8 T cells (Naïve GFP^- peripheral T). Data are representative of at least 4 mice per group from 2 independent experiments.

Maturation is unimpaired in RTEs blocked from accessing either the spleen or LNs alone (Fig. 3C), and in RTEs that access a given LN but are blocked from further recirculation (Fig. 4B), suggesting that the factors driving RTE maturation are shared between all SLOs. CD8 RTEs have been shown to circulate to the gut within the first 24 h of thymic egress (16), but RTEs do not preferentially accumulate in the gut at 1-2 weeks post-egress (data not shown). Given that S1P mimetic treatment blocks gut homing of conventional T cells (17), our data also show that access to the unique environment of the gut is not required for RTE maturation.

Functional RTE maturation requires access to SLOs

In addition to phenotypic defects, stimulated “homeless” CD4 RTEs secrete less IL-2 (Fig. 5A) and proliferate less relative to control RTEs (Fig. 5B). A similar comparison of “homeless” CD8 RTEs and control non-RTEs was precluded due to small blood volumes and the low proportion of CD8 T cells among RTEs (1). “Homeless” and control CD4 non-RTEs secreted more IL-2 than did control RTEs (Fig. 5A and data not shown), demonstrating that functional defects are limited to RTEs without access to SLOs. Thus, access to SLOs is important for RTEs to complete both phenotypic and functional maturation.

Figure 5.

RTEs require SLO access for functional maturation. A,B, “Homeless” or control CD4 RTEs (GFP⁺CD62L^hi) and CD4 non-RTEs (GFP^-CD62L^hi) were generated as in Fig. 3, sorted from the blood, and stimulated with anti-CD3 plus anti-CD28 in the presence of irradiated splenocytes. Sorted “homeless” and control RTEs had similar GFP MFIs. Following stimulation, IL-2 secretion at 24 hr in A and cell proliferation at 48 hr in B were quantified. Data are from blood pooled from 9-14 mice per condition, with 2-3 wells per stimulation condition. Error bars represent SD. *, p < .02 as compared to control RTEs, two-tailed Student’s t-test with equal variance.

What is the relationship between phenotypic and functional maturation? The increase in CD28 and IL-7Rα and the decrease in CD3 expression levels upon RTE maturation could modulate immunocompetence. While CD24 has roles in both naïve T cell homeostasis (18) and tolerance (19), less is known about the function in T cells of Qa2, a non-classical class Ib MHC molecule, or CD45RB, an isoform of the CD45 glycoprotein phosphatase (20, 21). Our data do suggest that phenotypic and functional maturation go hand in hand, making phenotypic marker expression a reliable indicator of the overall maturation state of RTEs.

Peripheral maturation was not seen in a study of RTEs labeled 10 d previously by intrathymic CFSE injection (22). At best, this 10 d interim would reveal only subtle differences between these “aged” RTEs and non-RTEs (data not shown). Furthermore, while our analyses of RTEs marked by intrathymic FITC injection recapitulate our findings from the RAG2p-GFP system (data not shown), CFSE injection in our hands often labels some extrathymic T cells. RTEs can also be exclusively identified as CD24^hiQa2^lo peripheral T cells, although ∼85% of RTEs are excluded by this gating system (data not shown).

In conclusion, we show here that not only does T cell development continue after thymic egress, but that this process is dynamically regulated. While RTEs are adjusting to the lymphoid periphery, their immune competence is dampened for a period of 2-3 weeks, and they rely on signals received in SLOs to drive them to the full competence of the non-RTE subset, ready to defend against invading pathogens.

Non-standard abbreviations

LN

lymph node

MFI

mean fluorescence intensity

RTE

recent thymic emigrant

SLO

secondary lymphoid organ

S1P

sphingosine 1-phosphate

S1P₁

sphingosine 1-phosphate receptor 1

SP

single positive thymocyte

Tg

transgenic