A T-bet-independent role for IFN-α/β in regulating IL-2 secretion in human CD4+ central memory T cells

J Immunol. Author manuscript; available in PMC 2009 December 15.

Published in final edited form as:

J Immunol. 2008 December 15; 181(12): 8204–8208.

PMCID: PMC2596627

NIHMSID: NIHMS77895

A T-bet-independent role for IFN-α/β in regulating IL-2 secretion in human CD4⁺ central memory T cells¹

Ann M. Davis,^* Hilario J. Ramos,^* Laurie S. Davis,^†^‡ and J. David Farrar^*^§²

^*Department of Immunology, The University of Texas Southwestern Medical Center, Dallas, Texas 75390

^†Department of Rheumatology, The University of Texas Southwestern Medical Center, Dallas, Texas 75390

^‡Department of Internal Medicine, The University of Texas Southwestern Medical Center, Dallas, Texas 75390

^§Department of Molecular Biology, The University of Texas Southwestern Medical Center, Dallas, Texas 75390

²To whom correspondence should be addressed: J. David Farrar, Department of Immunology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390−9093. Ph: (214) 648−7315, Fax: (214) 648−7331. E-mail: David.Farrar/at/UTSouthwestern.edu.

The publisher's final edited version of this article is available free at J Immunol

See other articles in PMC that cite the published article.

Abstract

IL-2 is a hallmark cytokine secreted by central memory CD4⁺ T cells (T_CM). Although naïve cells rapidly secrete IL-2 in response to antigen stimulation, IL-12 inhibits IL-2 secretion in daughter cells as they differentiate into Th1 cells. In this study, we uncover a unique role for IFN-α in regulating IL-2 secretion by human T_CM cells. IFN-α synergized with IL-12 to enhance a subset of cells which secreted high and sustained levels of IL-2. These IL-2-secreting cells displayed phenotypic and functional characteristics of T_CM and were capable of generating IFN-γ-secreting effectors upon secondary activation. T-bet has been implicated in negatively regulating IL-2 secretion in murine T cells; however, T-bet expression did not inhibit IFN-α-dependent IL-2 secretion in human T_CM cells. Thus, our results highlight a unique role for IFN-α in regulating the development of IL-2-secreting human T_CM cells.

Introduction

IL-12 promotes type I responses by inducing IFN-γ and TNF-α secretion from CD4⁺ T helper (Th) cells via STAT4 and T-bet (1). In addition to IFN-γ and TNF-α, early studies suggested that IL-2 was also selectively expressed in Th1 cells (2). However, as naïve precursors differentiate, they lose the ability to secrete high levels of IL-2 (3, 4). Following the clearance of pathogen, effector Th1 cells give way to the emergence of both effector (T_EM) and central (T_CM) memory subsets (5). T_EM retain the ability to secrete IFN-γ and TNF-α upon secondary antigen encounter and circulate through peripheral tissues via expression of proinflammatory chemokine receptors such as CXCR3 (6) and CCR5 (7). In contrast, T_CM circulate through secondary lymphoid tissues via selective expression of CCR7 (8). T_CM do not secrete proinflammatory cytokines, but rather secrete high levels of IL-2, allowing for expansion of secondary CD4⁺ and CD8⁺ effectors during recall responses. As various models of memory cell development propose that T_CM transit through the effector stage, it is unclear how T_CM emerge with the renewed ability to secrete IL-2.

The role of IFN-α in promoting adaptive responses has been controversial. Initially, IFN-α was proposed to mediate human Th1 responses (9). However, our recent studies demonstrated that, while IFN-α promoted STAT4 phosphorylation, this signaling pathway was not sustained and did not lead to stable T-bet expression or Th1 commitment (1, 10). Thus, IL-12 is unique in driving Th1 effector development, while the effects of IFN-α on Th responses remain unclear. In this study, we have uncovered a unique pathway whereby IFN-α preserves the ability of T_CM cells to secrete IL-2. IL-12 and IFN-α acted in synergy to promote the outgrowth of a subpopulation of IL-2-secreting T_CM cells capable of generating secondary effector Th1 cells. Further, the maintenance of IL-2 secretion in T_CM was independent of T-bet. These findings demonstrate for the first time an important role for IFN-α in shaping human T_CM responses.

Materials and Methods

Human subjects

120 ml peripheral blood was drawn from healthy adult donors. Informed consent was obtained from all donors, and this study was approved by the Internal Review Board (UT Southwestern).

Cytokines, antibodies, and reagents

Human IL-4, IL-12, CCL19, CXCL10, anti-IL-4, and anti-CCR7-APC antibodies were from R&D Systems (Minneapolis, MN). IFN-αA and anti-IFN-α/β receptor (hIFNAR2) antibody were from PBL (Piscataway, NJ). Anti-CD3, anti-CD28, and anti-IL-2-AlexaFluor700 antibodies were from BioLegend (San Diego, CA). Anti-CD4-PE and anti-CD45RA-Pacific Blue antibodies were from Caltag (Burlingame, CA). Anti-CD45RA-FITC, anti-CXCR3-PE, and anti-IFN-γ-PE-Cy7 antibodies were from BD Bioscience (San Diego, CA). Anti-T-bet antibody was from Santa Cruz (Santa Cruz, CA). Biotin-conjugated goat anti-rabbit Fab was from Jackson (West Grove, PA). Streptavidin-Qdot655 was from Invitrogen (Carlsbad, CA).

Human CD4⁺ T cell cultures

PBMCs were isolated from whole blood of healthy adult donors as described (10). Cells were stained with anti-CD45RA-FITC and anti-CD4-PE, and CD45RA⁺CD4⁺ cells were isolated using a MoFlo sorter (Cytomation, Fort Collins, CO) at > 90% purity. Alternately, naïve cells were isolated using the Human Naïve CD4 T Cell Enrichment Set (BD Bioscience). Cells were activated with plate-bound anti-CD3/anti-CD28 for 7 days in complete medium containing 10% FBS (cIMDM) in the presence of cytokines and/or neutralizing antibodies as described (10). In some experiments, cells were restimulated on day 7 for an additional 7 days.

Quantification of IL-2 production

Naïve cells were differentiated for two consecutive weeks as described above. On day 14, cells were stimulated for 24 hrs with plate-bound anti-CD3 (5 μg/ml). IL-2 concentration was assessed in supernatants by Cytometric Bead Array (BD Bioscience).

Flow cytometry

Polarized cells were restimulated for 4 hours with 0.8 μg/ml PMA (A.G. Scientific, Inc., San Diego, CA) + 1 μM ionomycin (Sigma, St. Louis, MO) in the presence of 1 μg/ml Brefeldin A (Epicentre, Madison, WI). Intracellular staining was performed as described (1). For analysis of apoptosis, cells were stained with 7-AAD and Annexin V-FITC (BD Bioscience), and analyzed on an LSR II cytometer (BD Bioscience), and the data were processed using FlowJo software (TreeStar, Inc., Ashland, OR).

Cell migration assays

Polarized cells were added to the upper chambers of 24-well transwell plates with a 5 μm pore membrane (Corning, Lowell, MA), in which the lower chambers contained 10 ng/ml CCL19 or 10 ng/ml CXCL10 in cIMDM. Media was used as a control. Cells were incubated for 2 hours at 37^oC, 5% CO₂ and then on ice for 10 minutes. Cells which had migrated to the lower chamber were restimulated for 4 hours with PMA + ionomycin in the presence of Brefeldin A, and intracellular staining for IL-2 and IFN-γ was performed.

Sorting of live IL-2- and IFN-γ-secreting cells

Day 7 polarized cells were washed and rested overnight in cIMDM without IL-2. Cells were stimulated for 2 hours with PMA + ionomycin, and labeling was performed using MACS Cytokine Secretion Assay for IL-2 and IFN-γ (Miltenyi Biotech, Auburn, CA). Cells were isolated on a FACSAria sorter (BD Bioscience) at > 90% purity.

Retroviral transduction

The GFPRV and T-bet-GFPRV constructs have been previously described (1). Naïve cells were activated with plate-bound anti-CD3/anti-CD28 in cIMDM in the presence of anti-IL-12, anti-IL-4, anti-IFN-γ, anti-IFNAR2 and 600 U/ml IL-2. Cells were incubated for 16 hours followed by retroviral transduction. Transduction was repeated for three consecutive days. On days 2 and 3, IL-12 and IFN-αA were added at concentrations described (10). On day 4, cells were split 1:10 into fresh cIMDM containing 50 U/ml IL-2 and rested to day 7. On day 7, cells were washed, rested overnight, and restimulated with PMA + ionomycin, and intracellular staining was performed.

Statistical analysis

Statistical analysis was performed by one-way and two-way ANOVA using Prism software (GraphPad, San Diego, CA). Values of p < 0.05 were considered significant.

Results and Discussion

Type I IFN and IL-12 promote IL-2 secretion in T_CM cells

Previous work from our lab has indicated that type I IFN does not promote Th1 development as assessed by IFN-γ and TNF-α secretion (1, 10). However, the regulation of IL-2 secretion in human Th1 effectors has not been examined. While various in vitro models of activation are available to promote Th development, we chose to utilize anti-CD3/anti-CD28 stimulation in order to strictly control the cytokine environment. As expected, activation of cells with either IL-12 or IFN-α alone induced slight induction of IL-2 secretion. Surprisingly, IL-12 and IFN-α acted synergistically to promote high levels of IL-2 secretion, and this effect was maintained for two consecutive weeks of activation (Fig. 1A). The synergy observed with IL-12 and IFN-α was consistent among 5 donors (Fig. 1B, p < 0.05 vs. neutral). However, not all cells were capable of IL-2 secretion, as intracellular stain analysis demonstrated that IL-12+IFN-α enhanced the generation of distinct IL-2⁺IFN-γ⁻ and IL-2⁺IFN-γ⁺ populations (Fig. 1C). Thus, IFN-α synergized with IL-12 to promote elevated IL-2 secretion from a subpopulation of human Th cells.

FIGURE 1

IFN-α/β and IL-12 promote IL-2 secretion in human CD4⁺ T cells. Naïve human Th cells were activated for one or two consecutive weeks with plate-bound anti-CD3 and anti-CD28 in the presence of IL-2 and the indicated cytokines as (more ...)

Because IL-2 secretion is a hallmark of T_CM (5), we wished to investigate whether the IL-2-producing cells demonstrated other characteristics of the T_CM phenotype. We assessed expression of various markers of human T_EM and T_CM phenotypes. As expected, naïve progenitors gave rise to subpopulations of cells with both T_CM (CD45RA^loCCR7^hi) and T_EM (CD45RA^loCCR7^lo) phenotypes, even in the absence of innate cytokines (Fig. S1). In addition, we found that the proinflammatory chemokine receptor CXCR3 was expressed on cells that differentially expressed CCR7 (Fig. S1). Further, cells that expressed high levels of CXCR3 with low expression of CCR7 were found to reside within the T_EM population, whereas the majority of CCR7^hiCXCR3^lo cells mapped to the T_CM subset (data not shown).

The overall percentages of naïve, T_CM, or T_EM were not dramatically altered in response to IL-12+IFN-α activation. Thus, it was unclear whether the IL-2-secreting cells that developed in response to IL-12+IFN-α were T_CM. We focused our analysis on defining the nature of these IL-2-producing cells. Naïve human CD4⁺ T cells were polarized with IL-12+IFN-α to day 7 followed by analysis of surface markers and expression of IL-2 and IFN-γ. First, cells were gated as follows: naïve (CD45RA^hiCCR7^hi), T_CM (CD45RA^loCCR7^hi), and T_EM (CD45RA^loCCR7^lo). Secretion of IFN-γ and IL-2 was assessed from cells within each gate (Fig. 2A). In agreement with observations in freshly isolated cells, T_CM cells secreted high levels of IL-2, whereas T_EM cells produced substantially more IFN-γ and less IL-2 (Fig. 2A). In parallel analyses, cells that secreted only IL-2 (IL-2⁺IFN-γ⁻) were contained predominantly within the CD45RA^loCCR7^hi population, and approximately 40% of these cells were CCR7^hiCXCR3^lo (Fig. 2B, upper and lower left panels). In contrast, IFN-γ single positive cells (IL-2⁻IFN-γ⁺) mapped to the CD45RA^loCCR7^lo quadrant and were uniformly CXCR3^hi (Fig. 2B, upper and lower middle panels). Interesting, while a significant proportion of the IL-2⁺IFN-γ⁺ cells mapped to the CD45RA^loCCR7^hi quadrant, these cells were also predominantly CXCR3^hi, and suggest that these cells may represent a transitional population. Nonetheless, the IL-2⁺IFN-γ⁻ population enhanced by IL-12+IFN-α predominantly displayed T_CM characteristics.

FIGURE 2

IL-2-secreting human CD4⁺ T cells display T_CM phenotypes. Naïve human Th cells were activated for 7 days with plate-bound anti-CD3 and anti-CD28 in the presence of IL-2, IL-12, IFN-αA, anti-IFN-γ, and anti-IL-4. On day 7, cells (more ...)

IL-2 was once considered a Th1-associated cytokine based on selective coexpression with IFN-γ (11). However, as Th1 effectors differentiate in response to IL-12, they lose the ability to secrete high levels of IL-2 (4). This may be due to IL-12-mediated induction of T-bet, which can negatively regulate IL-2 (3). Furthermore, T_EM cells that migrate to peripheral sites express IFN-γ, but not appreciable levels of IL-2 (12), while IL-2 is predominantly expressed by T_CM cells. IL-2 secretion is an innate property of naïve Th cells, and retention of IL-2 expression is an important characteristic of T_CM. Here, we demonstrate for the first time a synergistic role for IL-12 and IFN-α in inducing high levels of IL-2 secretion in human T_CM cells.

IFN-α-regulated IL-2-secreting cells exhibit functional properties of T_CM

In addition to phenotypic characteristics, we proposed that the IFN-α-driven IL-2-secreting cells would display functional properties of T_CM, including CCL19-dependent trafficking, enhanced survival, and the ability to reconstitute IFN-γ-secreting effectors upon secondary activation. To examine this, a transwell migration assay was performed using either CCL19, a CCR7 ligand, or CXCL10, a CXCR3 ligand. Cells that migrated across the membrane were analyzed for IFN-γ and IL-2 expression. IL-2-secreting cells showed enhanced migration in response to CCL19 but not CXCL10 (Fig. 3A, p < 0.05 vs. media alone). In contrast, cells which secreted IFN-γ but not IL-2 showed elevated migration in response to CXCL10 but not CCL19 (Fig. 3A, p < 0.05 vs media alone). IFN-γ⁺IL-2⁺ cells migrated equivalently in response to both chemokines (Fig. 3A, p < 0.05 vs media alone), as expected based on their coexpression of CCR7 and CXCR3. Therefore, the expression patterns of CCR7 and CXCR3 on subsets of IL-2- and IFN-γ-secreting human Th cells correspond to functional specificity in migration.

FIGURE 3

IL-2- and IFN-γ-secreting cells demonstrate functional characteristics of T_CM and T_EM cells. Naïve human Th cells were activated as described in Fig. 2. (A) Migration assays were performed for 2 hours in the presence of CXCL10 or CCL19. (more ...)

We next isolated the IL-2- and IFN-γ-secreting populations in order to more closely examine their individual functions. We performed live cell sorting of IL-2- and IFN-γ-producing populations after 7 days of activation in the presence of IL-12+IFN-α (Fig. S2). These cells were stimulated with plate-bound anti-CD3, and functional properties were assessed. The plasticity of IL-2- and IFN-γ-producing populations was examined by reactivation for a further 7 days in the presence of IL-12+IFN-α. Cells secreting only IL-2 or both IL-2 and IFN-γ were able to give rise to multiple populations of cytokine-expressing cells indicating significant plasticity in their ability to repopulate effector cells (Fig. 3B). Further, IL-12+IFN-α was required for development of IFN-γ-secreting cells from the IL-2⁺IFN-g⁻ population upon secondary activation (data not shown). However, regardless of the presence of IL-12+IFN-α, cells which secreted only IFN-γ were unable to give rise to cells which only secreted IL-2, suggesting a more terminally differentiated phenotype.

T_CM cells have also been shown to be more resistant to apoptosis than their T_EM counterparts (13). In order to probe the survival of IL-2- and IFN-γ-producing populations, we reactivated sorted cells for 3 days with plate-bound anti-CD3 and assessed apoptosis with Annexin V and 7-AAD staining. Cells secreting only IL-2 showed the lowest degree of apoptosis, while cells expressing IFN-γ demonstrated a greatly enhanced tendency to undergo apoptosis (Fig. 3C). Further, the percentage of cells that labeled with Annexin V and 7-AAD corresponded with the total percentage of live cells in the IL-2- and IFN-γ-expressing populations (Fig. S3).

Our data indicate that IL-2 secretion is linked with the ability to survive and regenerate new effectors. This segregation of IL-2 and IFN-γ production to the T_CM and T_EM compartments, respectively, was proposed at the first description of these subsets (5). Likewise, our studies suggest that cells that secrete only IFN-γ, but not IL-2, would have a severely impaired ability to reconstitute additional effectors upon rechallenge. In agreement with our findings, Harrington et. al. demonstrated the generation of T_EM from primary effectors that also secreted IL-2 (14).

Development of IL-2-secreting T_CM by IFN-α is independent of T-bet

Graded expression of T-bet has been proposed to play a role in the generation of CD8⁺ T_EM vs. T_CM cells (15). T-bet regulates IFN-γ secretion from Th cells, and mice deficient in T-bet show elevated IL-2 expression and generate a large population of T_CM cells (4). T-bet is responsive to both IL-12 and IFN-α, but unlike IL-12, IFN-α does not maintain T-bet expression (1). Thus, we reasoned that the IL-2-secreting population enhanced by IFN-α might have lower expression of T-bet as a result of local exposure to IL-12 vs. IFN-α. Surprisingly, we found that T_EM cells showed slightly lower T-bet expression compared with naïve or T_CM cells (Fig. 4A). Furthermore, both IL-2- and IFN-γ-expressing cells demonstrated similar T-bet content, whereas only a fraction of the cells that did not produce either cytokine were T-bet^lo.

FIGURE 4

T-bet expression controls effector but not memory CD4⁺ T cell phenotypes. (A) Naïve human Th cells were activated for 7 days with plate-bound anti-CD3 and anti-CD28 in the presence of IL-2, IL-12, IFN-αA, anti-IFN-γ, and anti-IL-4. (more ...)

In order to more directly examine the role of T-bet in the generation of T_EM and T_CM phenotypes, we expressed T-bet in developing human Th cells by retroviral transduction. Increased expression of T-bet was confirmed in transduced cells by intracellular staining (Fig. 4B). In the presence of IL-12+IFN-α, ectopic expression of T-bet did not significantly alter the proportions of T_CM and T_EM cells based upon CD45RA and CCR7 expression (Fig. 4B and C). T-bet has been shown to directly regulate the CXCR3 promoter (16), and T-bet markedly enhanced expression of CXCR3 (Fig. 4C). In addition to CXCR3, T-bet markedly increased IFN-γ secretion, in agreement with our previous findings (Fig. 4D). Surprisingly, T-bet expression failed to alter IL-2 expression (Fig 4D). This result suggests that the induction of IL-2 secretion by IFN-α within T_CM cells is not blocked by IL-12-induced T-bet during Th1 development. In contrast, our data indicate that T-bet expression in Th1 cells controls generation of effector phenotypes without impacting IL-2 secretion or the regulation of T_CM development.

IL-2 secretion is a critical component of T_CM function. IL-2 provides proliferative signals to T and B cells during anamnestic responses and has been implicated as a crucial signal for CD8⁺ T cell memory development. Bevan and colleagues recently demonstrated a requirement for IL-2 signaling for productive secondary CD8⁺ T cell responses and suggested that Th cells might be a primary physiological source of IL-2 (17). Given that both IL-12 and IFN-α/β are secreted in response to many intracellular infections, our study highlights the combined role of these two cytokines in driving the development of IL-2-secreting T_CM.

Supplementary Material

Click here to view.^{(2.5M, doc)}

Acknowledgments

We acknowledge Angela Mobley and the UT Southwestern Flow Cytometry Core Facility for help with cell sorting and flow cytometric analysis. We thank Drs. M. Siegelman and C. Pasare for helpful discussions and critical reading of the manuscript.

ABBREVIATIONS


APC	allophycocyanin

CBA	cytometric bead array

cIMDM	complete IMDM

T_CM	central memory T cell

T_EM	effector memory T cell

Footnotes

¹This work was supported by the NIH/NIAID (AI056222 awarded to J.D.F.). A.M.D. was supported by a training grant from the NIH/NIGMS (GM00820317), and H.J.R. was supported by a pre-doctoral fellowship from the NIH/NIAID (AI068622).

References

1. Ramos HJ, Davis AM, George TC, Farrar JD. IFN-α is not sufficient to drive Th1 development due to lack of stable T-bet expression. J. Immunol. 2007;179:3792–3803. [PMC free article] [PubMed]

2. Mosmann TR, Cherwinski H, Bond MW, Giedlin MA, Coffman RL. Two types of murine helper T cell clone. I. Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol. 1986;136:2348–2357. [PubMed]

3. Szabo SJ, Kim ST, Costa GL, Zhang X, Fathman CG, Glimcher LH. A novel transcription factor, T-bet, directs Th1 lineage committment. Cell. 2000;100:655–669. [PubMed]

4. Hwang ES, Hong JH, Glimcher LH. IL-2 production in developing Th1 cells is regulated by heterodimerization of RelA and T-bet and requires T-bet serine residue 508. J Exp Med. 2005;202:1289–1300. [PMC free article] [PubMed]

5. Sallusto F, Lenig D, Forster R, Lipp M, Lanzavecchia A. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature. 1999;401:708–712. [PubMed]

6. Kim CH, Rott L, Kunkel EJ, Genovese MC, Andrew DP, Wu L, Butcher EC. Rules of chemokine receptor association with T cell polarization in vivo. J. Clin. Invest. 2001;108:1331–1339. [PMC free article] [PubMed]

7. Kohlmeier JE, Miller SC, Smith J, Lu B, Gerard C, Cookenham T, Roberts AD, Woodland DL. The chemokine receptor CCR5 plays a key role in the early memory CD8+ T cell response to respiratory virus infections. Immunity. 2008;29:101–113. [PMC free article] [PubMed]

8. Debes GF, Arnold CN, Young AJ, Krautwald S, Lipp M, Hay JB, Butcher EC. Chemokine receptor CCR7 required for T lymphocyte exit from peripheral tissues. Nat Immunol. 2005;6:889–894. [PMC free article] [PubMed]

9. Rogge L, D'Ambrosio D, Biffi M, Penna G, Minetti LJ, Presky DH, Adorini L, Sinigaglia F. The role of Stat4 in species-specific regulation of Th cell development by type I IFNs. J Immunol. 1998;161:6567–6574. [PubMed]

10. Davis AM, Hagan KA, Matthews LA, Bajwa G, Gill MA, Gale MJ, Farrar JD. Blockade of virus infection by human CD4⁺ T cells via a cytokine relay network. J. Immunol. 2008;180:6923–6932. [PMC free article] [PubMed]

11. Seder RA, Paul WE. Acquisition of lymphokine-producing phenotype by CD4+ T cells. Annual review of immunology. 1994;12:635–673. [PubMed]

12. Román E, Miller E, Harmsen A, Wiley J, von Andrian UH, Huston G, Swain SL. CD4 effector T cell subsets in the response to influenza: heterogeneity, migration, and function. J. Exp. Med. 2002;196:957–968. [PMC free article] [PubMed]

13. Macallan DC, Wallace D, Zhang Y, de Lara C, Worth AT, Ghattas H, Griffin GE, Beverley PCL, Tough DF. Rapid turnover of effector-memory CD4⁺ T cells in healthy humans. J. Exp. Med. 2004;200:255–260. [PMC free article] [PubMed]

14. Harrington LE, Janowski KM, Oliver JR, Zajac AJ, Weaver CT. Memory CD4 T cells emerge from effector T-cell progenitors. Nature. 2008;452:356–360. [PubMed]

15. Intlekofer AM, Takemoto N, Kao C, Banerjee A, Schambach F, Northrop JK, Shen H, Wherry EJ, Reiner SL. Requirement for T-bet in the aberrant differentiation of unhelped memory CD8+ T cells. J Exp Med. 2007;204:2015–2021. [PMC free article] [PubMed]

16. Beima KM, Miazgowicz MM, Lewis MD, Yan PS, Huang TH, Weinmann AS. T-bet binding to newly identified target gene promoters is cell type-independent but results in variable context-dependent functional effects. J Biol Chem. 2006;281:11992–12000. [PubMed]

17. Williams MA, Tyznik AJ, Bevan MJ. Interleukin-2 signals during priming are required for secondary expansion of CD8+ memory T cells. Nature. 2006;441:890–893. [PMC free article] [PubMed]