Early Effector cells Survive the Contraction Phase in Malaria Infection and Generate both Central and Effector Memory T cells

Abstract

CD4 T cells orchestrate immunity against blood-stage malaria. However, a major challenge in designing vaccines to the disease is poor understanding of the requirements for the generation of protective memory T cells (Tmem) from responding effector T cells (Teff) in chronic parasite infection. Here, we use a transgenic mouse model with T cells specific for the Merozoite Surface Protein (MSP)-1 of Plasmodium chabaudi to show that activated T cells generate three distinct Teff subsets with progressive activation phenotypes. The earliest observed Teff subset (CD127⁻CD62L^hiCD27⁺) are less divided than CD62L^lo Teff and express memory genes. Intermediate (CD62L^loCD27⁺) effector subsets include the most multi-cytokine producing T cells, while fully activated (CD62L^loCD27⁻) Late effector cells have a terminal effector T cell phenotype (PD-1⁺, Fas^hi, AnnexinV⁺). We show that while IL-2 promotes expansion, it actually slows terminal effector differentiation. Using adoptive transfer, we show that only Early Teff survive the contraction phase and generate the terminal late effector T cell subsets, while in uninfected recipients, they become both central and effector Tmem. Furthermore, we show that progression towards full Teff activation is promoted by increased duration of infection, which in the long-term promotes Tem differentiation. Therefore, we have defined markers of progressive activation of CD4 effector T cells at the peak of malaria infection, including a subset that survives the contraction phase to make Tmem, and show that antigen and cytokine levels during CD4 T cell expansion influence the proportion of activated cells that can survive contraction and generate memory in malaria infection.

INTRODUCTION

Malaria infection induces severe immunopathology with significant long-term global health and economic consequences (1). People in endemic areas quickly become resistant to severe disease, but develop immunity only after repeated and chronic infections over many years, with prevalence of parasitemia falling off by adulthood by roughly 70% (2-4). During the symptomatic erythrocytic stage of Plasmodium infection, both CD4 T and B cells are essential for production of antibody for clearance of parasites and memory cells are essential throughout life to recognize cumulative parasite diversity (5, 6). Th1 cells are also critical to activate phagocytes both enhancing parasite killing (7), and regulating immunopathology in both mice and humans by production of IL-10, IL-27 and TGF-β (8, 9). Therefore, CD4 T cells are essential for protection from lethal disease (10), and study of the CD4 T cell response to the pathological blood-stage of parasitemia will lead to development of protective vaccine strategies.

We have studied the immune response to Plasmodium chabaudi, a rodent model of chronic malaria infection (11), with the goal of understanding the activation and development of CD4 T cells in malaria parasite infection. As in other chronic parasitic infections (12-14), chronic P. chabaudi infection maintains protection to re-infection (15). Effector memory T cells (Tem) (12, 16-19) and effector T cells (13, 14) have been implicated in this phenomenon, called premunition, but the cellular mechanisms of maintaining protection, by continuous effector generation or maintenance of Tem, and avoiding full exhaustion, remain unclear. While PD1 expression has been identified during infection (17, 20, 21), it is clear that functional effector cells are generated and that some functional memory cells emerge (16, 17). Therefore, we investigated CD4 T cell memory in P. chabaudi, and found a reduction of classical Tmem (CD44^hiCD127^hi) in chronic infection, suggesting that traditional central memory cells did not provide the improved protection. Importantly, the majority of CD44^hi memory T cells in the chronic infection did not survive by homeostatic proliferation, did not proliferate quickly to secondary infection, and had the phenotype of Tem (22). We also studied B cell memory, and found only a small effect of chronic infection (23), suggesting that T helper cells may induce this effect. We showed the important role of CD4 memory cells in protection by a transfer of chronically-stimulated, CD44^hiCD25⁻ CD4⁺ T cells, which protect better than resting memory T cells (17). Subsequently, looking at MSP-1 specific B5 TCR Tg T cells (24), we identified an increase in CD44^hiCD62L^lo IFNγ⁺TNF⁺IL-2⁻ T cells at memory timepoints in chronic infection compared to animals with a treated infection (17). However, there was no decrease in the number of effector memory T cells (CD44^hi CD127^hi CD62L^lo) in the absence of infection for one month, suggesting that Tem cytokine potential, or their ability to generate Teff, and not necessarily their survival, depends on persistent infection. Furthermore, we showed that protective Tem were derived from central memory T cells in chronic infection (17), suggesting that they are not simply derived from long-lived Teff. This data led us to hypothesize that Tem cells survive after persistent infection clears and provide sentinel activity for some time, even after clearance of infection. This idea has been well supported by research on resident effector memory T cells (25). Therefore, the question becomes how to drive the generation of cytokine-producing Tem for protective vaccination protocols for chronic infection.

Recent evidence suggests that CD4 memory T cell differentiation is determined in the earliest stages of activation, and that Pre-Tcm and Tem can be detected concurrently with effector differentiation (26-28). This important work suggests that in order to understand Tem differentiation, and the potential to make protective Tem against malaria by vaccination, we must follow effector activation from the beginning to find the precursors of long-lived Tem (reviewed in (29)). We hypothesized that Tem were either generated from early (27, 30) or late effector cells (26, 31). However, it was not possible to identify both of these Teff subsets in our BALB/c MSP1-specific B5 TCR Tg adoptive transfer system, since the only marker of CD4 terminal Teff described to date, Ly6C (28), does not stain CD4 cells in BALB/c. Therefore, we identified new markers of progressive activation of CD4 effector cells for the purpose of determining the derivation of CD4 memory T cells in P. chabaudi malaria.

The markers CD127, CD62L and CD27 are well-described molecules downregulated upon activation (17, 32, 33), each with different kinetics. Therefore, we tested if these facile surface markers, that we previously used to define progressively differentiated CD4 memory T cell subsets (17), provide a good marker panel to distinguish Teff phases along the spectrum of activation. In these studies, we show progressive activation of CD4 Teff from Early (Teff^Early, CD127⁻CD62L^hi), which express high levels of anti-apoptotic genes, to Intermediate effector cells (Teff^Int, CD127-CD62L^loCD27⁺), that make more cytokines to fully activated Late Teff (Teff^Late, CD127-CD62L^loCD27⁻), which are terminally differentiated. By adoptive transfer at the peak of infection, we define a linear progression of activation phenotypes and confirm their ability to make terminally differentiated Teff and survive the contraction phase, to generate memory T cells, as suggested by previous studies on early effector cells (26-28). Furthermore, we show that only the early effector cell subset generates both central and effector memory T cells in the long-term, and that the ratio of these is regulated by the duration of antigen exposure. This data will inform future studies on the differentiation pathway of CD4 effector memory T cells in malaria infection, which are necessary to understand how to develop vaccine strategies that will induce protective immunity.

MATERIALS AND METHODS

Mice, Parasites and in vivo experiments

Thy1.1 BALB/cByJ were backcrossed to BALB/cJ (N4, Jackson Labs, Bar Harbor, ME) and maintained in our SPF animal facility with ad libitum access to food and water. B5 TCR Transgenic mice, a generous gift from Jean Langhorne (MRC-NIMR Mill Hill, UK), were generated as described (10) and backcrossed to BALB/cJ (N4-9). The B5 TCR recognizes MSP1 (1157–1171, ISVLKSRLLKRKKYI/I-E^d); B5 TCR Tg mice were typed using primers Valpha2, gaacgttccagattccatgg and atggacaagatcctgacagcatcg; and Vbeta8.1, cagagaccctcaggcggctgctcagg and atgggctccaggctgttctttgtggttttgattc. Ifng/Thy1.1 Knock-In mice on a C57BL/6 background (26) were a kind gift of Casey Weaver (University of Alabama, Birmingham, AL). Mice 6–12 weeks old were infected with 10⁵ Plasmodium chabaudi chabaudi (AS) infected erythrocytes i.p.. Parasites are counted in thin blood smears stained with Giemsa (Sigma, MO) by light microscopy. In some experiments, mice were treated with 4mg/kg mefloquine hydrochloride (MQ) anti-malarial drug by oral gavage for three days starting day 3 or 5 post-infection. Anti-IL-2 (mixture of clones S4B6 and JES6-1) or Rat IgG2a control (clone 2A3 BioXcell): 0.25 mg of each Ab/mouse) were administered in saline i.p. every other day starting on day 1 post-infection.

All experiments were carried out in accordance with the protocol approved by the University of Texas Medical Branch Institutional Animal Care and Use Committee.

Flow cytometry

Single-cell suspensions from spleens were made in HEPES buffered Hank’s Balanced Salt Solution (Gibco, Lifetechnologies), incubated in red blood cell lysis buffer (eBioscience, San Diego, CA), and stained in PBS 2% FBS (Sigma, MO) and 0.1% sodium azide with anti-CD16/32 (2.4G2) supernatant (Bio Xcell, West Lebanon, NH) followed by combinations of FITC–, PE–, peridinin chlorophyll protein (PerCP)-Cy5.5, PE/Cyanine 7 (Cy7), PE/Cy5, (APC)–, or APC/eflour780-conjugated antibodies (all from eBioscience) and CD62L PE-TexasRed (Invitrogen), CD44-Brilliant Violet 785 (Biolegend, San Diego, CA) and CD5-PE (eBiosciences). PSGL-1 PE (BD Bioscience), Ly-6C FITC, CD25 PE-Cy7 (eBioscience), and CXCR5-Biotin followed by StreptavidineFluor 450. Cells were collected on a LSRII Fortessa using FACSDiva software (BDbiosciences, San Jose, CA) and analyzed in FlowJo version 9.7 (TreeStar, Ashland, OR). Surface stained cells were washed with PBS and resuspended in 100 μM AnnexinV antibody in proprietary buffer (Invitrogen, Grand Island, NY) for 15 min in the dark at room temperature. Cells were analyzed within 1 hour of staining. For Bcl-2, cells were fixed in 2% paraformaldehyde (Sigma) and permeabilized using BD Perm/Wash buffer, cells were stained with Bcl2-PE (BD biosciences) for 40 min. at 4^oC and washed. Compensation was performed in FlowJo using single stained splenocytes (using CD4 in all colors). For presentation, data from 3–4 mice is concatenated to achieve sufficient cell numbers for presentation and Boolean gating analysis, after each mouse was analyzed and averages and standard error of the mean calculated. For intracellular cytokine and CFSE staining, cells were treated as previously described (17). Gating strategies are shown in Supplemental Figure 1.

Cell Sorting

Splenic CD4⁺ T cells from day 8 infected B5 TCR Tg donors were purified by EasySep biotin selection kit (STEMCELL, Vancouver, Canada) using biotinylated anti-CD8α (55-6.7), B220 (RA3-6B2), CD11b (MI/70), CD11c (N418), F4/80 (BM8) and Ter119 (eBioscience). Enriched T cells were then stained with anti-CD4-FITC, CD44-APC-Cy7, CD127-PE, CD62L-Texas Red and CD27-APC for Teff subset sorts. Cells were sorted on a FACSAriaI with FACDiva software (BD biosciences). Naïve (CD44^loCD25⁻), or effector (CD44^int-hiCD127^neg) CD4⁺ T cells were sorted from 5-12 week old B5 TCR Tg mice and transferred i.p.. Transfer of one million cells models the physiological precursor number of MSP1-specific CD4 T cells seen in the memory phase of infection (0.01% or 1/8800), as previously shown by limiting dilution analysis of splenocytes for responsiveness to this fragment of MSP-1 (34). We have also replicated results in this system with as few as 5,000 cells (17). Gating strategy and purity are shown in Supplemental Figure 1.

Real Time PCR

RNA from purified B5 TCR Tg Teff cell subsets (spun into RLT buffer (Qiagen Valencia, CA), every hour) was analyzed by real time PCR using a custom PrimePCR assay (BioRad, Hercules CA) with SYBR green iQ Supermix, and iScript (BioRad) for cDNA synthesis. Normalized Relative Expression was calculated using GAPDH primers run on each plate as an internal control [2^(CT of Gene of Interest-CT GAPDH)], where gene of interest CT was compared in two subsets (Teff^Int/Teff^Early (Fig. 3B) or relative to Naïve (Fig. 3C) using Excel or Spotfire software for heatmap, using the z-score to display relative expression of each gene (TIBCO software).

Statistics

Where indicated, experiments were analyzed by one way ANOVA followed by students t-test or Tukey’s or Wilcoxon Rank-Sum test for non-parametric data in Prism (GraphPad, La Jolla, CA), *p≤0.05, **p≤0.01, ***p≤0.001, NS=not significant.

RESULTS

Effector T cell Subsets mark progressive activation

Although significant work has been done to understand the differentiation of CD4 memory T cells in vivo, the pathway for the development of memory cells from responding effector T cells remains unclear, especially in chronic infection. In the current study, we sought to demonstrate the cellular pathway of activation of effector CD4 T cells in malaria infection, and to use the MSP1-specifc TCR Transgenic model in BALB/c to identify effector T cells with the potential to survive through the contraction phase and form protective effector memory T cells. In order to do this, we found it essential to identify subsets of responding Th1 cells that represent progressively activated effector T cells. To observe the kinetics of Ifng expression in responding CD4 T cells in P.chabaudi infection, and the phenotypes of Ifng+ effector cells, we took advantage of an IFN-γ reporter mouse on a polyclonal C57BL/6 background where cells express Thy1.1 transiently on the surface when the Ifng gene is transcribed (Ifng/Thy1.1 Knock-In). These mice were infected with P. chabaudi, and the phenotype of the Teff during expansion was determined on days 5, 7 and 9 post-infection (p.i.), and during the contraction phase as measured on day 12 p.i.. Gating on CD127⁻ Teff, as shown in supplemental Fig. 1, we observed three effector T cell subsets, (CD62L^hiCD27⁺, CD62L^loCD27⁺ and CD62L^loCD27⁻), which appeared to change in proportion over time in response to P. chabaudi infection (Fig. 1A). The CD62L^hiCD27⁺ Teff population was enriched on day 5 p.i, and decreased by day 7, while the CD62L^loCD27⁺ and CD62L^loCD27⁻ Teff were enriched by day 9-12, correlating with the peak of infection (days 8-10) in this model (Fig. 1A, graph). Within Ifng/Thy1.1⁺CD4⁺ T cells, we observed that an average of 95% of the responding cells were CD127⁻ on day 9 post-infection (Fig. 1B), indicating that they were indeed activated effector cells (17, 33), and contained all three of the observed subsets. The CD62L^loCD27⁺ effector population was the highest in IFN-γ production, and numbers of all Teff sub-populations increased to day 9 post-infection, and decreased by day 12, as parasite was cleared. The highest levels of T-bet were observed on the CD62L^loCD27⁻ subset on day 5 and 9 post-infection (Fig. 1C), suggesting full effector activation. As we previously showed a linear path of differentiation for memory cells from CD62L^hiCD27⁺ Tcm to CD62L^loCD27⁺ Early Tem to CD62L^loCD27⁻ Late Tem (17), we hypothesized a similar effector progression from CD62L^loCD27⁺ to CD62L^loCD27⁻ Teff before deletion in the contraction phase. We have therefore named these CD4⁺CD127⁻ effector populations Early (Teff^Early, CD62L^hi CD27⁺), Intermediate (Teff^Int CD62L^lo CD27⁺), and Late (Teff^Late, CD62L^lo CD27⁻) effector T cells and tested this progression in this report.

Figure 1. IFN-γ⁺ cells contain three subsets in the effector population.

Ifng/Thy1.1 reporter mice were infected with P. chabaudi and (A) Proportions of the Teff subsets identified by CD62L and CD27 were determined within CD4⁺CD127⁻ on day 5, 7, 9 through 12 and quantification shown in graph. (B) Teff (Thy1.1⁺CD127⁻CD44^hi) on day 9 p.i. and proportions of effector subsets (right plot) with numbers (graph) of cytokine producers (CD4⁺Ifng/Thy1.1⁺CD127⁻) in each Teff subset over the course of infection with parasitemia in shaded gray area. (C) MFI of T bet⁺ cells in CD4⁺Ifng/Thy1⁺ cells on days 5, 7 and 9. Data is representative of two experiments with 3 mice per group. Error bars represent SEM, Student t-test was used to compare Teff^Early vs. Teff^Int on the same day. *p<0.05, **p<0.01, ***p<0.001. (Teff^Early, CD62L⁺CD27⁺), intermediate (Teff^Int, CD62L⁻CD27⁺), and late (Teff ^Late, CD62L⁻CD27⁻).

We sought to test the hypothesis of progressive Teff activation in this C57BL/6 model by using recently reported markers of early Teff/Pre-Tcm and late Teff/Th1 cells (27, 28). We therefore, validated the overlap of Pre-Tcm in effector subsets defined by CD62L and CD27 (Supplemental Fig. 2A) and measured enrichment for Pre-Tcm and terminal effectors defined as PSGL-1⁺Ly6C⁻ (Supplemental Fig. 2B) and CXCR5⁺CD25⁻ (Supplemental Fig. 2C). Interestingly, the Pre-Tcm population was enriched in the CD62L^hi Teff^Early population described here, as measured by CXCR5⁺PSGL1⁺, and Th1 and Tfh terminal effector cells are enriched in CD62L^lo Late and Intermediate Teff (Supplemental Fig. 2D).

Malaria-specific CD4 effector T cells are progressively activated synchronously

In order to show that malaria-specific T cells progress through these clearly defined Teff phases, we used a transgenic mouse that expresses a T cell receptor (TCR) specific for the P. chabaudi Merozoite Surface Protein-1 (B5 Tg). Naïve (CD4⁺ CD44^lo CD25⁻) B5 TCR Tg T cells were adoptively transferred (10⁶) into congenic Thy1.1 mice, which were subsequently infected with P. chabaudi. The cells were recovered on d9 post infection, and gated on divided cells (CFSE⁻) to identify responding Teff, which are predominantly CD127⁻ (Fig. 2A). As seen in the polyclonal system, we observed that within the CD127⁻ Teff population there are three populations as defined by CD27 and CD62L. We found that CFSE⁻CD127⁻CD62L^hiCD27⁺ Early effector cells constitute an average of 6.5% of malaria-specific effector T cells on day 9 post-infection. When day 9 B5 TCR Tg cells were gated on CD127⁻ Teff, we observed that the CD62L^hiCD27⁺ population contained all of the CFSE^hi cells (Fig. 2B), suggesting that they are the first to divide. Following the kinetics of development, we observed that on day 5 post-infection, B5 Tg Teff cells were primarily CD27⁺CD62L^hi Early effector cells (71.9% average, Fig. 2C). The CD62L^lo subsets began to increase on day 7, with the CD27⁺ (average 54.95%) preceding the CD27⁻ Teff on day 9 p.i. (average 67.3%), suggesting that the B5 T cells progress through activation in a synchronized manner. The sequential proportional increase of these three Teff subsets during infection further confirmed our hypothesis that they represent a linear progressive activation of T cells in response to malaria infection.

Figure 2. The Teff subsets define progressive activation.

Naïve B5 TCR Tg T cells (CD44^loCD25⁻) were transferred (1-2×10⁶) into Thy1.1 congenic mice that were subsequently infected with P. chabaudi. Splenocytes were analyzed by flow cytometry. (A) Dividing, malaria-specific Teff (CFSE⁻Thy1.2⁺CD127⁻) d9 p.i. include early (red, CD629L^hiCD27⁺), intermediate (blue, CD62L^loCD27⁺), and late (green, CD62L^loCD27⁻) effector subsets. (B) B5 TCR Tg donor cells (CD4⁺Thy1.2⁺CD127⁻) showing proliferation of Early, Intermediates and Late as measured by CFSE-violet dilution at day 9 p.i. (C) B5 TCR Tg donor cells (CD4⁺Thy1.2⁺) showing changing phenotype of Teff (CD127⁻) from day 5-9 p.i.. Data is concatenated from three recipient hosts and quantified (right). Parasitemia is shown in shaded gray area. Gates and quadrants are set on all CD4⁺ cells. Error bars represent SEM, t-test was used to compare the subsets on the same day. **p<0.01, ***p<0.001

To test this hypothesis at the molecular level, we investigated the expression of pro- and anti-survival factors in these activation subsets. We measured expression of the anti-apoptotic molecule Bcl-2, and markers of terminal differentiation (PD-1, Fas and Annexin V binding) in adoptively transferred B5 TCR Tg cells on day 7 p.i. (Fig. 3A). Only CD127⁻CD62L^hi Early B5 Teff cells expressed high levels of Bcl-2, while the CD62L^lo subsets were Bcl-2⁻, and expressed the inhibitory receptor PD-1 and death-domain-containing receptor Fas, suggesting terminal differentiation of the CD62L^lo subsets. Supporting this conclusion, CD27⁻ Teff^Late had the highest expression of Annexin V⁺, suggesting that these cells had entered early apoptosis. As an indication of the maturation of functional avidity of the developing effector T cells, we measured expression of CD5, an inhibitor of TCR signaling (35). We observed progressive decline in expression of CD5 from Teff^Early to Teff^Int and Teff^Late cells (Fig 3A), suggesting that as cells become more activated, they decrease regulation of the TCR, and increase in signal strength, thereby correlating with a gain in functional avidity with increasing activation (36). We also measured mRNA levels of other pro-and anti-apoptotic molecules by real time PCR of B5 TCR Tg T cells (Fig. 3B). In the transition from Early (CD62L^hiCD27⁺) to Intermediate (CD62L^loCD27⁺) effectors, we detected significant up-regulation of proapoptotic (fas, bcl2l15 (Bfk), bbc3 (Puma), but not bak1 or bax, and down-regulation of anti-apoptotic (bcl2, mcl1, pim2, pim3) genes at the transcriptional level (Fig. 3B). Because the Teff^Early express high levels of pro-survival molecules suggesting longevity, we also measured expression of memory and effector differentiation genes (Fig. 3C). Interestingly, Teff^Early expressed higher levels of memory genes including (foxo1, id2, klf2 and S1pr1) as well as Teff genes (id3, prdm1, tbx21,) than the other two subsets. A notable exception is Bcl6, which is reported to be a memory promoting transcription factor (37), but is not increased in Early CD4 effector cells in this infection. Taken together these data suggest that early-activated CD4 Teff contain a longer-lived Pre-Tmem population, while the highly activated CD62L^lo cells appear terminally differentiated.

Figure 3. Mature Effector T cells are terminally differentiated.

Naïve B5 TCR Tg CD4 T cells (1×10⁶) were transferred into Thy1.1 mice subsequently infected with P. chabaudi. Splenocytes were stained for Teff subsets (as defined in Figure. 2) and (A) Bcl-2, PD-1, CD95, Annexin V and CD5 were determined. Plots show representative histograms while and graphs (below plots) show percent quantification for Bcl-2 and PD1 on d7 p.i or MFI for CD95, Annexin V and CD5 on d16 p.i of positive cells in the histograms for all the subsets. Data is concatenated for B5 TCR Tg cells from three mice and is representative of two independent experiments. Gates set on isotype controls, but isotype and and positive controls not shown for clarity. Error bars represent SEM comparing different subsets on the same day. (B) Expression of selected pro- and anti-apoptotic genes showing fold change from early to intermediate on d8 p.i. of sorted Teff subsets. (C) Effector (labeled in Red) and memory (labeled in Blue) gene expression for the three Teff subsets. PCR data was normalized to GAPDH and z-score of relative expression of each gene compared to naïve across the subsets is shown. Error bars represent SEM, t-test was used to compare the subsets. *p<0.05, **p<0.01, ***p<0.001.

In order to study the progression of effector cytokine production during activation, we determined the cytokine profile of the malaria-specific effector T cell subsets by intracellular cytokine staining in adoptively transferred B5 TCR Tg cells on day 7 post-infection. The CD62L^loCD27⁻ Intermediate Teff subset contained the highest percentage of cytokine producers for IFN-γ, TNFα and IL-2 (Fig. 4A), showing that this is the most responsive subset. Interestingly, some Teff^Early also make cytokines; however, 40% of Teff^Early cytokine producers are TNF⁺IL-2⁺, which is the cytokine profile of CD4 Tcm (17, 38, 39). Furthermore, only the CD62L^lo Teff subsets (Teff^Int and Teff^Late) contain triple cytokine producers (TNF⁺IFN-γ⁺IL-2⁺, Fig. 4B), indicating that the multi-cytokine producing phenotype is associated with highly activated effectors at this stage of infection.

Figure 4. Mature Effector T cells are multi-cytokine producers.

Naïve B5 TCR Tg CD4 T cells (1×10⁶) were transferred into Thy1.1 mice subsequently infected with P. chabaudi. Splenocytes were stained d7 p.i. for Teff subsets (as defined in Figure 2) and cytokine secretion of the subsets was determined using intracellular staining. (A) Histograms of cytokine profile of each Teff (CD4⁺CD127⁻) subset d7 p.i. after adoptive transfer, including early (Red, CD62L^hiCD27⁺), intermediate (Blue, CD62L^loCD27⁺), and late (Green, CD62L^loCD27⁻) effector cells are shown. Isotype controls were used to define positive gates (not shown for clarity). The plots represent percent quantification of all the mice in the experiment. (B) Pie charts show percentage of B5 Tg cytokine-producing cells in each subset expressing 1-3 cytokines as calculated by Boolean analysis. Data represent concatenated B5 TCR Tg cells from three mice and is representative of two independent experiments. Error bars represent SEM, comparing different groups. *p<0.05.

CD62L^lo classical effector T cell subsets are short-lived effector cells

Based on the expression of memory and anti-apoptotic genes and the timing of their appearance, we hypothesized that early effector cells may survive the contraction phase and contribute to memory differentiation, while CD62L^lo subsets would lose this potential. Therefore, we determined the fate of each effector subset over the peak of infection and observed survival in the contraction phase. Teff subsets were sort-purified from infected B5 TCR Tg mice (day 8 p.i.), based on the gating strategy and purity shown in Supplemental Figs 1C, D and E. Each subset was transferred into infection-matched Thy1.1⁺ congenic recipients days 8 through day 11 of infection (Fig. 5). Despite historic reluctance to infect intact TCR transgenic animals in favor of adoptive transfer, we have previously verified similar degrees of T cell activation in B5 TCR Tg and wildtype animals, and found phenotypes and protection comparable (17). Adoptive transfer of the three purified effector T cell populations at the peak of infection into infection-matched recipients showed that only the Teff^Early population was capable of surviving three days of infection during the contraction phase (Fig. 5A) demonstrating that Early CD62L^hi effector cells could survive the T cell contraction during parasite clearance. The CD62L^lo Teff were terminally differentiated, as there were minimal T cell numbers recovered from the recipients of these subsets. Furthermore, the recovered Teff^Early cells generated both Teff^Int and Teff^Late populations after three days (Teff^Early OUT) in this period of the T cell response when the parasitemia is decreasing and the T cells are contracting (as seen in Fig. 1B).

Figure 5. Early T effector cell subset generates all effector and memory subsets.

Malaria-specific B5 CD4⁺ effector T cell subsets were sorted on d8 p.i. and transferred (7×10⁵) into (A) infection-matched or (B) naive Thy1.1 recipients. (A) Numbers (left) and phenotype of sorted (IN) and of recovered (OUT) donor cells (CD4⁺Thy1.2⁺) were determined on d3 post-transfer (d11 p.i.) in infection-matched recipients. Purity of sorted early Teff (IN), and CD4⁺Thy1.2⁺CD127⁻ cells recovered from early Teff recipients (OUT) are shown. (B) Numbers (left) and phenotype of recovered (OUT) Early donor cells (CD4⁺Thy1.2⁺) were determined on d14 post-transfer in naïve recipients. All recovered cells are represented as contour plots with outliers. Pie charts show percentage of each Teff (CD127⁻) subset (left pie) or Tmem (CD127^hi) subset, right pie) within recovered Early subset. Error bars represent SEM, data was analyzed by one-way ANOVA followed by Tukey’s test comparing all the groups. *p<0.05, **p<0.01, ***p<0.001.

In order to test the survival potential of the three subsets of effector cells in the absence of antigen, we investigated their phenotype on recovery from an uninfected recipient. We transferred each purified Teff subset into uninfected congenic Thy1.1 hosts, and analyzed the recovery and phenotype of the cells 14 days after transfer. Again, we recovered significantly higher numbers of B5 TCR Tg T cells from recipients of Teff^Early than Teff^Int or Teff^Late (Fig. 5B). Interestingly, in the absence of antigen, an average of 65% of the recovered CD127⁻ Teff retained their original CD62L^hiCD27⁺ phenotype, and did not progress as in the infection-matched recipients (Fig. 5B, Teff^Early OUT, Teff pie), indicating that infection promotes the progressive activation through these phenotypes from CD62L^hi to CD62L^loCD27⁻ (Fig. 5A, B). Furthermore, on transfer, they had been completely CD127 negative, while after two weeks without parasite exposure, the majority re-upregulated CD127 to become CD127 intermediate (average 61.5%). In order to identify true memory phenotype cells, we gated on the CD127^hi and identified both central memory (Tcm) and effector memory cells (Tem) derived from the Teff^Early (Fig. 5B, Teff^Early OUT, Tmem pie). These data suggest that the Early Teff subset can make all the later Teff subsets in the presence of antigen and a fraction can also survive the contraction and potentially make memory T cells upon antigen elimination.

Antigen and IL-2 are important in progressive activation of Effector T cells

Since effector T cells expand in response to IL-2, we predicted that blocking IL-2 would impede progressive differentiation of the Teff subsets. We therefore administered anti-IL-2 over the first 9 days of infection to Thy1.1 congenic recipients of B5 TCR Tg CD4 T cells (Fig. 6). Unexpectedly, there was a significant decrease in the proportion of the Teff^Early cells, as opposed to an increase, and an increase in the terminal MSP1-specific Teff subsets in the anti-IL-2 treated animals compared to isotype controls on day 5p.i., which persisted through day 9 (Figs. 6A, B and C), although the differences seen on day 9 were not statistically significant. As expected, animals treated with anti-IL-2 did not have significant T cell accumulation by day 9 p.i. (Fig. 6D), when T cell numbers are at their peak in this infection. These data suggest that progressive differentiation of the Teff from Early to Late activation is slowed by IL-2, and that licensing of full effector activation and T cell expansion are separable.

Figure 6. IL-2 deprivation promotes Effector T cell Progression.

Naïve B5 TCR Tg CD4 T cells (1×10⁶) were transferred into Thy1.1 mice subsequently infected with P. chabaudi. Mice were treated with anti-IL-2 on alternate days starting d1 p.i. Splenocytes were stained for B5 Teff subsets as in Figure 3. (A) Plots of Teff subsets comparing isotype to anti-IL-2 treatment on days 5 p.i.. (B) Percent of each Teff subset out of CFSE^loThy1.2⁺CD127⁻ Teff on day 5 p.i. (C) Plots of Teff subsets on d9 comparing isotype to anti-IL-2. (D) Absolute cell numbers of proliferating effector (CFSE^loCD127⁻) B5 subsets on days 5 and 9 p.i.. Plots are representative of one mouse from a group of 4-5 mice. Data was analyzed using Student t-test and shows mean and SEM. *p<0.05.

As the duration and strength of antigen stimulation has been proposed to determine the extent of CD4 T cell activation and differentiation (30), and this has been persuasively shown for CD8 T cells (40), we tested whether shortening the period of infection would affect the full activation of effector T cells and the ratio of memory subsets generated. Adoptively transferred recipient mice were infected but treated with mefloquine hydrochloride (MQ), an anti-malarial drug, on days 3-5 (MQ d3) or 5-7 (MQ d5) post-infection to reduce the load of parasitemia, or left untreated, and effector T cell phenotype was determined in all the mice on day 9 post infection (Fig. 7A). There were strikingly higher proportions of the CD127⁻CD62L^hiCD27⁺ Early effector T cell subset on day 9, when parasite grew unimpeded for only 3 days, with intermediate levels of Teff^Early population in MQ d5 treated animals. This decreased to 4.4% Teff^Early (average) when mice were not treated, suggesting that a longer duration of parasitemia indeed induced a stronger total signal and pushed the effector cells through the phases of Teff activation from Teff^Early to Teff^Late. Early antigen elimination was also accompanied by a significant increase in CD44^hiCD127^hi cells (data not shown), indicating the possibility of early memory generation.

Figure 7. Antigen drives the progressive activation of the Teff subsets.

Naïve B5 TCR Tg CD4 T cells (1×10⁶) were transferred into Thy1.1 mice subsequently infected with P. chabaudi. (A) Mice were treated with mefloquine hydrochloride (MQ) on d3-5 or d5-7 and Teff subset phenotypes were determined on d9 p.i. Plots show phenotype of CD4⁺CFSE⁻CD127⁻ T cell subsets from d3, d5 and no treatment groups. Data represent one mouse of 3 mice per group. Graphs show quantification of the CD127-CD62L^hiCD27⁺ early Teff subset or parasitemia in the three groups. Student t-test was used to compare the mefloquine (MQ) treated vs. not treated (NT) and error bars represent SEM. (B) Mice were treated with MQ (d3-7) or CQ (d30-34) and phenotypes of memory subsets was determined on d53 p.i.. Plots show phenotypes of memory subsets (CD4⁺CD127⁺) while pie charts show quantification of Tmem in the two groups. Numbers in the pies represent average and SEM of all mice in each group. Graph shows total numbers of recovered cells. Student t-test was used to compare MQ vs CQ treated groups and error bar represent SEM. *p<0.05, ***p<0.001.

Our previous work shortening the infection with chloroquine on day 30, did not change the Tcm to Tem ratio (17). To determine if earlier elimination of antigen might influence the transition of Teff into memory, we treated the recipient mice with MQ on days 3-7 and analyzed memory cell formation on day 53 p.i.. As a control, we treated another group with chloroquine (CQ) d30-34 to eliminate parasite exposure for the three weeks before analysis for memory T cells. We observed a significantly higher proportion of CD127^hiCD62L^hiCD27⁺ central memory T cells in the mice treated with MQ on day 3 compared to those exposed to a full infection (Fig. 7B), and a correspondingly higher proportion of CD127^hiCD62L^hiCD27⁺ early effector memory T cells in mice exposed to infection for a month, suggesting that exposure to chronic infection induces Tem, while shorter exposure to the same parasite induces Tcm. Unsurprisingly, there were also significantly more total cells recovered from the animals with the full month of infection. While this agrees with previous studies suggesting that chronicity generates Tem, it is striking how early the infection had to be treated to reduce Tem generation.

DISCUSSION

In the current study, we have identified a pathway of progressive activation of effector CD4 T cells in malaria infection to generate memory T cells. We have described three subsets of effector cells, identified by their transient downregulation of CD127, that define the gamut of activation states leading to terminal differentiation versus memory survival. The earliest of these activation states is a population quickly re-expressing CD62L (32), that potentially contains both short-lived effector cells and memory precursor effector cells. We show that effector cells go through a linear pathway of activation starting from an early CD127⁻CD62L^hi Teff precursor, which are the least divided, and are not terminally differentiated likely due to higher expression of pro-survival genes. The later, CD62L^lo, stages include CD27⁻ late effector cells that express PD-1, Fas and AnnexinV⁺, indicating early stages of cell death; and CD27⁺ Intermediate Teff, with the maximum effector function. We show that highly activated CD4 Teff lose the ability to survive after the contraction phase, while the CD62L^hi early Teff are recovered both after contraction and after being “rested” in naïve hosts and have the phenotypes of both central and effector memory T cells.

This work is the first step to defining the fate of various CD4 effector cells generated in parasitic infections, and may eventually contribute to resolving the controversial derivation of effector memory T cells, as they can derive from both IFN-γ-producing (26, 41, 42), and IFN-γ⁻ Teff (43). Residual Teff have also been identified several months post-infection (14, 17, 28, 44-46), suggesting that some cells with Teff phenotypes survive beyond contraction. On the other hand, it has also been suggested in the progressive differentiation model for CD4 T cell differentiation, that Tem are generated from fully activated effector cells (26, 28), although this is hard to reconcile with the above-mentioned data. The progressive activation states and terminal differentiation of Teff cells demonstrated here, support the current paradigm from studies in CD8 T cells that as Teff become progressively activated, they lose the potential to generate memory T cells (40). Our data suggest that like effector CD8 T cells (KLRG1⁺, Ly6C^hi), CD4 T cells lose their potential to generate memory with increasing activation, and in the case of CD8 cells, with exposure to IL-12 and increased T-bet expression (47, 48), suggesting that highly activated T cells are terminal and not likely to survive past the contraction. Supporting this interpretation, effector and Tmem differentiation appear to be mutually exclusive as they are regulated by pairs of transcription factors including T-bet and eomes, Id3 and Id2, and Blimp and Bcl6, which each promote terminal differentiation over memory T cell development respectively (40, 49, 50).

Several studies indicate that memory cell precursors can be detected during the early stages of activation in other models and infections (27, 51). In this study, we used CD127 downregulation to mark Teff cells because effector cells from the peak of infection are CD127 negative (17). Our CD62L^hi Teff^Early effector population overlaps significantly with early effector CXCR5⁺CD25⁻ cells proposed to be Pre-Tcm by Pepper et al., while the CD62L^lo populations contain terminal effector cells (27, 28). We show that the low proportional enrichment of Ly6C⁺ in later populations is due to a large expansion of the CXCR5⁺ population, which has been shown to contain a mixture of Tfh and Pre-Tcm (52). In P. chabaudi infection, we see a large expansion of CXCR5⁺ effector cells that are Tbet⁺ (V.H.C., M.M.O., R.S., manuscript in preparation), and here we show that CD62L^lo Teff^Int and Teff^Late CXCR5⁺ cells are also PSGL-1^loLy6C⁻ suggesting that the CD62L^lo Teff contain Th1 and Tfh effector cells, but not Pre-Tcm, thereby correlating all three definitions of the late subsets. Therefore, our work suggests that CD62L^hiCXCR5⁺ are Pre-Tcm, while CD62L^loCXCR5⁺ may be Tfh effector cells. Since Teff^Early can generate both Teff^Late and memory cells, it is likely that the Teff^Early subset described here contains the precursors of both effector and memory cells, as also suggested by the representation of Ly6C⁺ population in the Teff^Early. Consistent with this, we observe an enrichment of both effector-associated and memory-associated genes in Teff^Early compared to the Teff^Int and Teff^Late effector subsets. This is consistent with a very early decision point, such as that provided by asymmetrical division (53). Nevertheless, CD62L^hi Teff^Early is clearly the effector T cell subset containing the long-lived precursors of CD4 memory T cells.

By precisely measuring degrees of effector T cell activation in a linear pathway, and neutralizing IL-2 we show that the T cell growth factor, IL-2, while essential for CD4 T cell proliferation, also has a role in regulating the pace of progression through these activation states. Without IL-2, progression through the spectrum of activation is increased, suggesting a more rapid development of a highly activated, but terminally differentiated late effector T cell phenotype in the absence of this important proliferation-inducing factor. This is consistent with previous studies showing that IL-2 plays a role during the T cell expansion phase in generation of functional CD4 memory cells (54, 55).

The role of antigen in the generation and maintenance of protective memory T cells is an active area of research. We had previously shown that treatment of P. chabaudi infection on day 30 had no effect on the ratios of Tcm to Tem (17), however; by treating the infection and exposing the T cells to either three or thirty days of malaria infection, we have shown that only a very short exposure to the infection increases the durable Teff^Early population. Interestingly, elimination of parasite on day 3 led to a higher proportion of Tcm at d53 p.i., while d30 treated mice had more Tem, and higher cell numbers supporting the hypothesis that chronic infection drives Tem generation, and providing the additional insight that the change from Tcm generation to Tem generation depends on very quick control of infection. Similarly, some studies on development of CD8 memory have shown that the cumulative degree of activation, determined by the type and duration of stimuli early in activation can determine effector versus memory fate of the responding cells (56-58). Additional signals such as production of survival cytokines, even later in infection are also likely to affect the subset ratios over time (59). If prolonged antigen or cytokine exposure leads to a loss in these long-lived cells, this could be one reason why people in malaria endemic areas do not quickly develop immunity to infection.

Although there is not a very good protection from malaria infection, there is protection from severe disease, which may be T cell mediated. We have shown that early effector T cells in this model infection can in fact survive to the memory stage, and make both Tcm and Tem phenotype cells, which are likely to balance the pathology of re-infection. Further elucidation of their ability to protect and maintain this protection are an important focus of future studies. Because Tem is the predominant memory subset generated by chronic infections like malaria and tuberculosis (16, 17, 60), and autoimmune diseases (61), and that their differentiation is not currently understood separately from that of effector cells, understanding the development of these potentially longer-lived cells is likely to be critical for the development of successful adjuvants, treatments and vaccines for chronic T cell mediated disease.

Abbreviations used in this article

B5 TCR Tg

B5 T cell receptor transgenic

KI

knock-in

MQ

mefloquine hydrochloride

MSP-1

Merozoite Surface Protein-1

NT

Not treated

p.i.

post-infection

Tcm

central memory T cells

Teff

Effector T cells

Teff^Early

Early Effector T cells

Teff^Int

Intermediate Effector T cells

Teff^Late

Late Effector T cells

Tem

effector memory T cell

Tfh

T follicular helper

Tmem

Memory T cells