Autocrine and Paracrine IL-2 Signals Collaborate to Regulate Distinct Phases of CD8 T Cell Memory

Abstract

Differential IL-2 signaling and production is associated with disparate effector and memory fates. Whether the IL-2 signals perceived by CD8 T cells come from autocrine or paracrine sources, the timing of IL-2 signaling, and their differential impact on CD8 T cell responses remain unclear. Using distinct models of germline and conditional IL-2 ablation in post-thymic CD8 T cells, this study shows that paracrine IL-2 is sufficient to drive optimal primary expansion, effector and memory differentiation, and metabolic function. In contrast, autocrine IL-2 is uniquely required during primary expansion to program robust secondary expansion potential in memory-fated cells. This study further shows that IL-2 production by antigen-specific CD8 T cells is largely independent on CD4 licensing of DCs in inflammatory infections with robust DC activation. These findings bear implications in immunizations and adoptive T cell immunotherapies, where effector and memory functions may be commandeered through IL-2 programming.

Introduction

Memory CD8 T cells comprise a critical component of protective immunity against a variety of viral, bacterial, and parasitic infections as well as cancers (Kalia et al., 2016). Hence, a complete understanding of factors governing memory CD8 T cell differentiation is crucial for designing efficacious vaccines and immunotherapies (Wong and Pamer, 2001; Kaech and Ahmed, 2001; Van Stipdonk et al., 2001; Van Stipdonk et al., 2003). Studies involving in vivo fate-tracking of phenotypically distinct effector CTLs have confirmed that cardinal memory properties such as polyfunctionality, the ability to survive long-term after antigen clearance, and robust recall expansion and protection from secondary challenge are programmed into memory precursor effector CD8 T cell subset (MPECs) during early stages of primary CTL responses (Kalia et al., 2010; Sarkar et al., 2008; Joshi et al., 2007; Kaech et al., 2003; Huster et al., 2004; Pipkin et al., 2010; Bachmann et al., 2005). Moreover, several studies have further established that the proportions of MPECs and their short-lived effector cells (SLECs) counterparts can be manipulated by altering the dose or duration of antigen and cytokines such as IL-12, IFN-I, IL-1 and IL-2 (Kalia et al., 2010; Sarkar et al., 2008; Joshi et al., 2007; Pipkin et al., 2010; Sarkar et al., 2018; Kolumam et al., 2005). Thus, a better understanding of all the factors involved in regulating MPEC and SLEC fates will help commandeer CTL differentiation towards memory or short-lived effector lineages based on the specific clinical need.

Studies involving ablation of IL-2R components, or IL-2 supplementation and blockade strategies show that IL-2 signals are critical for optimal memory responses against secondary challenge (Williams et al., 2006; Bachmann et al., 2007; Mathieu et al., 2015; Obar et al., 2010) and also for robust effector differentiation (Kalia et al., 2010; Pipkin et al., 2010; Obar et al., 2010; Mitchell et al., 2010; Mitchell et al., 2013). Differential dose and duration of physiological IL-2 signals are also associated with disparate effector and memory fates (Kalia et al., 2010; Pipkin et al., 2010; Obar et al., 2010). While robust IL-2 production upon antigenic restimulation is a key property of long-lived, polyfunctional, lymphoid resident memory CD8 T cells (Kalia et al., 2010; Kalia and Sarkar, 2018), terminally differentiated effector CD8 T cells are largely incapable of IL-2 production upon restimulation, thus implying a differential role of autocrine and paracrine IL-2 signals in effector and memory differentiation. However, the relative contributions of autocrine and paracrine IL-2 to effector and memory functions, and their timing of action are not clearly defined. Paracrine IL-2 from CD4 T cells has been proposed as a mode of CD4 T cell help for augmenting primary and secondary CD8 T cell responses (Obar et al., 2010). Alternatively, CD4 T cell-mediated DC licensing has also been implicated in promoting memory responses through co-stimulation (CD40 and CD27)-dependent induction of autocrine IL-2 (Feau et al., 2011; Feau et al., 2012). Likewise, conflicting requirements of IL-2 signals during primary and secondary memory responses have also been reported (Williams et al., 2006; Bachmann et al., 2007; Obar et al., 2010; Redeker et al., 2015).

Here we use CD8 T cell-specific germline and conditional IL-2 ablation strategies to discern differential roles of autocrine and paracrine IL-2 in effector and memory responses. Given that memory-fated CD8 T cells retain IL-2 production capability during primary as well as secondary responses, we also queried whether autocrine IL-2 is required during primary or secondary expansion phases. By ablating autocrine IL-2 in post-thymic GP33-specific CD8 T cells prior to primary LCMV infection, or in fully differentiated memory cells prior to secondary challenge, our studies unequivocally demonstrate that autocrine IL-2 signals are critically required during the programming phase to instill the cardinal memory property of robust recall expansion. On the other hand, paracrine IL-2 signals were sufficient to drive optimal priming, primary expansion, and effector differentiation through robust upregulation of glycolysis and oxidative phosphorylation. Our studies further reveal that under strongly inflammatory conditions, autocrine IL-2 production by memory-fated CD8 T cells occurs independently of CD4 T cell help. Together, these findings will help guide the development of next generation of adoptive T cell therapies that are optimized not only for robust primary expansion and therapeutic efficacy, but also for vigorous secondary expansion to prevent future tumor resurgence.

Results

Autocrine IL-2 signals uniquely promote secondary expansion potential of memory CD8 T cells, but not antigen-independent longevity or polyfunctionality

During an immune response, primarily activated CD4 T cells, but also CD8 T cells and DCs produce IL-2 along with NKT cells and mast cells (Kalia and Sarkar, 2018). Thus, IL-2 signals from both antigen-specific CD8 T cells as well as non CD8 T cell sources are available for controlling various aspects of T cell differentiation. While strategies of artificially polarizing IL-2 signals by supplementation or blockade have been utilized to establish the biological relevance of IL-2 in driving expansion and effector differentiation of antigen-specific CD8 T cells (Williams et al., 2006; Mitchell et al., 2010; Mitchell et al., 2013; Cheng et al., 2002; Cheng et al., 2002; Blattman et al., 2003; Hand et al., 2010), these non-physiological manipulations do not permit finer functional distinctions between paracrine and autocrine IL-2 signals.

To distinguish the roles of autocrine and paracrine IL-2 in regulating the quintessential memory properties of recall expansion, antigen-independent persistence and polyfunctionality, we used one of the most stringent tests for autocrine IL-2 requirement – we transferred wild-type and IL-2-deficient TCR transgenic P14 CD8 T cells into wild-type C57BL/6 mice at physiological precursor frequencies (Fig 1A). In this experimental setting, except for a small subset of IL-2^−/− GP33-specific P14 CD8 T cells, all LCMV-specific CD8 T cells and other immune cells are capable of IL-2 production after acute LCMV infection (Fig S1A). Thus, bypassing any confounding issues of Treg dysfunction and autoimmune disorders associated with global il2 ablation (Sadlack et al., 1993; Suzuki et al., 1995), in this experimental system LCMV-primed wild-type and autocrine IL-2-deficient P14 cells are allowed to develop into memory in the same infectious milieu with similar amounts of paracrine IL-2.

Fig. 1.

Autocrine IL-2 is critical for secondary expansion of memory CD8 T cells. (A) Wild-type and IL-2^−/− D^bGP33 specific CD8 T cells were adoptively transferred into naïve B6 mice which were subsequently infected with LCMV_Arm. Mice were followed to at least day 30 post-infection to establish memory. Wild-type and IL-2^−/− memory D^bGP33 specific CD8 T cells were isolated, normalized and transferred into naïve B6 recipients and rechallenged with LM-GP33. (B) Flow cytometry plots from PBMC of infected mice at given times post-secondary rechallenge. (C) Bar graph depicts number of donor cells in spleen at day 8 post-secondary challenge. (D) Bar graph depicts number of donor cells in spleen and liver at day 45 post-secondary challenge. (E-F). Effector markers (E) and transcription factors (F) were analyzed at day 8 post-secondary rechallenge. Bar graphs depict MFI of respective markers (G) Mice were treated with BrdU 12 hours prior to analysis at day 4.75 post-infection to compare cell cycling. Dot plots are gated on donor CD8 T cells and show proportion of BrdU+ T cells in spleen. Bar graph depicts frequency of BrdU+ donor CD8 T cells. (H) Histogram is gated on naïve (gray), WT (Black) and IL-2^−/− (White) T cells at day 5 post-infection. Numbers in plot represent MFI of Bcl-2 expression in WT (Bold) and IL-2^−/− (Plain) antigen specific CD8 T cells in spleen. (I) Caspase 3/7 activity was compared at day 6.75 post-infection in spleen. Bar graph depicts frequency of Caspase 3/7+ CD8 T cells. Data are representative of 5 experiments with n=2-5 mice per group. Bar graphs show mean and SEM. Unpaired Student’s t-test was used with statistical significance in difference of means represented as * (P ≤ 0.05), ** (P ≤ 0.01), ***(P ≤ 0.001). See also Fig. S1 and Fig S5

Memory Recall Responses

To evaluate the “true” recall expansion potential of wild-type and IL-2^−/− memory on a per cell basis, without any confounding factors associated with differential numbers or tissue localization (Williams et al., 2006; Redeker et al., 2015), we purified and adoptively transferred equal numbers of wild-type and IL-2^−/− memory CD8 T cells into congenically mismatched naïve mice prior to heterologous challenge with LM-GP33 (Fig 1A). Consistent with previous reports of impaired recall expansion of memory cells following immunization with non-replicating Listeria monocytogenes, or vaccinia virus (Feau et al., 2011), we observed that LCMV-primed IL-2^−/− memory cells expanded to significantly lower levels following Listeria monocytogenes challenge. Donor cell proportions in blood and absolute numbers in tissues at the peak of secondary expansion are presented (Fig 1B–C, S1B). Following recall expansion, IL-2^−/− T cells contracted significantly more resulting in lower numbers than wild-type memory cells in both lymphoid and nonlymphoid tissues (Fig 1D, S1C–D).

Nonetheless, secondary effector differentiation was similar in wild-type and IL-2^−/− cells as indicated by similar levels of CD62L downregulation, and PD-1, granzyme B (GzmB) and KLRG-1 upregulation (Fig 1E, S1E). Differences in effector expansion could not be explained by differential expression of effector or memory associated transcription factors Bcl-6, Blimp-1, T-bet and Eomes on wild-type and IL-2^−/− secondary effector cells at day 8 post rechallenge (Fig 1F). Reduced secondary expansion of IL-2^−/− memory cells was instead associated with (i) impaired proliferation as indicated by reduced BrdU incorporation (Fig 1G), (ii) reduced survival as suggested by lower expression levels of Bcl-2 (Fig 1H) and increased apoptosis marked by greater proportion of Caspase3/7+ cells (Fig 1I), resulting in sub-optimal pathogen control compared to wild-type memory cells (Fig S1F). To confirm that the observed recall defect in IL-2^−/− donors was not caused by differences in viral clearance during primary infection, wild-type and IL-2^−/− P14 cells were co-transferred into the same naïve mouse at physiologic precursor frequencies, and memory recall expansion potential was compared to that of wild-type and IL-2^−/− memory cells from single transfer experiments presented in Figure 1. We observed a similar recall defect between single and co-transferred IL-2^−/− CD8 T cells, suggesting that viral clearance and other immune factors were not confounding the recall response data (Fig S1G).

Loss of autocrine IL-2 production specifically impacted recall expansion potential of memory cells. Conversely, paracrine IL-2 signals provided by surrounding IL-2 sufficient cells were necessary to sustain both wild-type and IL-2^−/− antigen-specific CD8 T cell numbers (Fig 2A–C), localization (Fig S2A), homeostatic maintenance (Fig S2B–D) and IFN-γ and TNF-α production upon antigenic rechallenge (Fig 2K). Autocrine IL-2^−/− memory cells, which developed in paracrine IL-2 sufficient environment, accumulated in largely similar numbers as wild-type donors in both secondary lymphoid tissue as well as peripheral sites (Fig 2B–C). Qualitatively, paracrine IL-2 was sufficient for the development of lymphoid and circulatory memory cells as indicated by similar upregulation of lymphoid homing marker CD62L (Fig 2D–F) and similar localization to lymphoid and nonlymphoid tissues (Fig S2A) by wild-type and Il-2^−/− memory cells.

Fig. 2.

Autocrine IL-2 is dispensable for memory homeostasis and polyfunctionality. (A) Wild-type and IL-2^−/− D^bGP33 specific CD8 T cells were adoptively transferred into naïve B6 mice which were subsequently infected with LCMV_Arm. (B) Flow cytometry plots show frequency of donor CD8 T cells in respective tissues at day 75 post-infection. (C) Bar graphs depict number of donor cells at day 40 post-infection in respective tissues. (D) Dot plots are gated on donor CD8 T cells at day 75 post-infection. (E-F) Bar graphs depict proportion of MPEC (CD127⁺ KLRG-1⁻; E) and central memory (CD127⁺ CD62L⁺; F) CD8 T cells in spleens of memory mice. (G) Bar graphs depict MFI of CD127 in donor CD8 T cells at day 75 post-infection. (H) Histograms are gated on donor (black) and endogenous naïve (gray) CD8 T cells at day 38 post-infection in spleens. Numbers and bar graph represent MFI of Bcl-2 of donor CD8 T cells. (I) Transcription factors were analyzed at day 40 post-infection. Bar graphs show MFI of respective markers in wild-type and IL-2^−/− donor CD8 T cells. (J) Dot plots are gated on donor CD8 T cells in spleens at day 38 post-infection. Frequency and bar graph depict percent of donors taking up BrdU. (K) Bar graphs depict frequency of IFN-γ+ only, or IFN-γ+TNF-α+ donor cells at day 40 post-infection. Data are representative of 4 experiments with n=3-5 mice per group. Bar graphs show mean and SEM. Paired and unpaired Student’s t-test was used to determine statistical significance. See also Fig. S2. And Fig S5

Memory Homeostatic Maintenance

Antigen-independent homeostatic self-renewal is a hallmark property of long-lived memory cells. To understand paracrine or autocrine IL-2 requirements for memory homeostasis, we assessed the expression levels of pro-survival memory markers IL-7Rα and Bcl-2 and compared homeostatic proliferation rates of wild-type and IL-2^−/− memory cells by measuring BrdU incorporation in vivo and in vitro in response to homeostatic cytokines. Autocrine IL-2-sufficient and -deficient memory cells expressed similar amounts of IL-7Rα and Bcl-2 (Fig 2D–G, 2H), and upregulated similar levels of T-bet, Eomes, Bcl-6 and Blimp-1 (Fig 2I). In vivo, autocrine IL-2 was dispensable for homeostatic proliferation as indicated by about 8% BrdU+ donor memory cells in both IL-2-sufficient and -deficient subsets (Fig 2J). In vitro stimulation with homeostatic cytokines IL-7 and IL-15 also induced similar division in wild-type and IL-2^−/− CD8 T cells (Fig S2B). In vitro stimulation with the cognate GP33 antigen paralleled the in vivo results of reduced secondary expansion despite similar upregulation of IL-2Rα activation marker (Fig S2B–C). These data suggest that paracrine IL-2 is capable of supporting homeostatic maintenance of memory CD8 T cells in the absence of autocrine IL-2. To further confirm the redundancy of autocrine IL-2 in homeostatic proliferation of memory CD8 T cells, wild-type and IL-2^−/− memory CD8 T cells were co-transferred into irradiated naïve recipients and lymphopenia-induced expansion was assessed. Wild-type and IL-2^−/− CD8 T cells were able to maintain similar cell numbers, thus demonstrating that paracrine IL-2 signals are sufficient to maintain memory T cells in the absence of autocrine IL-2 (Fig S2D) in both lymphopenic and homeostatic conditions.

Memory Polyfunctionality

Coproduction of cytokines, also referred to as polyfunctionality, is also a hallmark of long-lived protective memory cells. In this qualitative aspect also, we noted that IL-2^−/− memory cells were comparable to wild-type cells, with similar proportions of single (IFN-γ only) and double (IFN-γ and TNF-α) (Fig 2K) cytokine producers and induction of similar expression levels of each cytokine upon antigenic restimulation (Fig S2E–F). As expected, IL-2 deficient memory cells did not produce measurable IL-2 upon restimulation (Fig S2G).

Collectively, these findings demonstrate that autocrine IL-2 is specifically required for robust recall expansion potential of memory CD8 T cells, while homeostatic maintenance, central memory differentiation, and polyfunctionality are dependent on the presence of paracrine IL-2.

Paracrine IL-2 signals are sufficient for driving primary expansion and effector differentiation of CD8 T cells in the absence of autocrine IL-2.

To determine whether autocrine IL-2 exerted similar beneficial effects on primary CTL expansion as secondary expansion (Fig 1C), we next conducted a head-to-head comparison of primary expansion of autocrine IL-2-sufficient and -deficient CD8 T cells. We used the experimental strategy of co-transferring wild-type and IL-2^−/− P14 cells at physiological precursor frequencies, followed by acute infection with LCMV (Fig S3A).

As shown in Fig 3, IL-2^−/− donor cells showed similar expansion kinetics as wild-type donors, with largely similar proportions in blood (Fig 3A). At the peak of CTL expansion, IL-2^−/− CD8 T cells accumulated at comparable numbers to their wild-type counterparts in both secondary lymphoid tissues (spleen and inguinal lymph nodes) and peripheral sites of infection (lung and liver) (Fig 3B, S3B). Consistent with similar overall expansion, the proliferative rates of IL-2 sufficient and deficient donor cells were largely similar with about 28% BrdU+ cells (Fig 3C), thus suggesting that paracrine IL-2 signals were primarily responsible for the robust effector CTL expansion. However, a calculation of the fold decline in donor cells from peak of effector expansion to memory phase showed that IL-2^−/− donors exhibited modestly increased contraction (Fig 3D), which was transient, and at later time-points wild-type donors also contracted to similar levels as IL-2^−/− donors (Fig S3C). These findings parallel a similar report in CD4 T cells showing increased contraction of autocrine IL-2-deficient CD4 T cells during early stages (McKinstry et al., 2014).

Fig. 3.

Primary expansion and effector CTL differentiation occurs independently of autocrine IL-2. Wild-type and IL-2^−/− D^bGP33 specific CD8 T cells were adoptively transferred into naïve B6 mice which were subsequently infected with LCMV_Arm. (A) Line graph depicts frequency of donors in PBMC at given times post-infection. (B) Bar graphs depict number of donor cells in given tissues at day 8 post-infection. (C) Mice were treated with BrdU 12hrs prior to analysis at day 8 post-infection. Dot plots are gated on donor CD8 T cells and frequency represent proportion of donor BrdU+ effector cells. (D) Bar graphs depict percent contraction of donor cells from peak effector expansion (day 8) to early memory (day 24) in PBMC. (E) Dot plots are gated on donor CD8 T cells in spleen at day 8 post-infection. Bar graph depicts proportion of MPEC and SLEC donor cells. (F-G) Histograms are gated on donor WT (black), IL-2^−/− (white) and endogenous naïve (gray) CD8 T cells at day 8 post-infection in spleen. Numbers represent MFI in WT (Bold) and IL-2^−/− (Plain) donor cells. Bar graphs depict MFI of effector markers (F) and transcription factors (G). (H-I) Flow cytometry plots are gated on donor CD8 T cells. (H) Bar graphs show frequency of degranulating, IFN-γ+ donor CD8 T cells. (I) Flow cytometry plots are gated on donor CD8 T cells. Bar graphs show percent and MFI of IFN-γ and TNF-α production in donor CD8 T cells. Data are representative of 2-3 experiments with n=2-5 mice per group. Bar graphs show mean and SEM. Paired and unpaired Student’s t-test was used to determine statistical significance. See also Fig. S3 and Fig S5

We also conducted a detailed characterization of effector differentiation of autocrine IL-2 sufficient or deficient donors in a paracrine IL-2-replete environment. We found that paracrine IL-2 signals were principally required to drive robust effector differentiation, as demonstrated by similar expansion of SLECs (Fig 3E–F) expressing high levels of PD-1, as well as potent expression of effector-associated transcription factors T-bet and Blimp-1 and no significant perturbation in memory-associated transcription factors Eomes and Bcl-6 between wild-type and IL-2^−/− cells (Fig 3G). Furthermore, despite lacking autocrine IL-2, effector function was not compromised in IL-2^−/− donors, which showed similar upregulation of GzmB (Fig 3F), and degranulation (Fig 3H) as wild-type effector cells. Consistent with potent effector differentiation, wild-type and IL-2^−/− donors also exhibited similar polyfunctionality following direct ex vivo restimulation with GP33 peptide, and coproduced similar levels of IFN-γ and TNF-α (Fig 3I), with the exception of IL-2, confirming their il2 locus deletion (Fig S3D).

Together, these data demonstrate that paracrine IL-2 signals drive the expansion and effector differentiation during primary CTL response, whereas autocrine IL-2 signals are required for optimal survival during the contraction phase and for mounting a robust recall expansion during rechallenge.

Paracrine IL-2 signals play a dominant role in priming optimal effector expansion and metabolism

During early stages of CD8 T cell priming paracrine IL-2 signals might be limiting, thus necessitating a critical dependence on autocrine IL-2. Hence, we compared early activation and proliferation of wild-type and IL-2^−/− donor cells at 48-66 hours after LCMV infection. Both wild-type and IL-2^−/− CD8 T cells exhibited similar rates of proliferation, as indicated by similar CFSE dilution with similar distribution of cells between distinct divisions (Fig 4A). At the secondary lymphoid sites of priming, similar numbers of wild-type and autocrine IL-2-deficient donor cells were quantified (Fig 4B, S4A). Wild-type and IL-2^−/− donor cells were activated similarly and upregulated canonical activation markers CD69, PD-1, CD25 and CD43 to similar extents (Fig 4C). Following direct ex vivo stimulation with GP33 peptide, IL-2^−/− donor cells clearly lacked autocrine IL-2 (0.2% producing cells compared to the 26.6% of wild-type T cells) (Fig S4B) but produced largely similar levels of IFN-γ and TNF-α, with only a modest trend towards decreased effector cytokine production compared to their wild-type counterparts (Fig 4D–E). These data suggest that early activation, proliferation, and effector differentiation are largely driven by similar perception of paracrine IL-2 signals and activation, as suggested by similar levels of CD25 (Fig 4C).

Fig. 4.

Paracrine IL-2 is sufficient for initial activation and expansion of CTLs. (A) Wild-type and IL-2^−/− CD8 T cells were adoptively co-transferred into naïve B6 mice. Mice were infected with LCMV_Arm and donors were analyzed in respective tissues at day 2.75 post-infection. (A) Histograms are gated on donor CD8 T cells in spleen of infected (Black) or uninfected (gray) mice. Frequency shows proportion of cells past the zero peak of CFSE. Bar graphs depict proportion of cells per cell division. (B) Bar graphs depict number of donors in spleen and iLN. (C) Histograms are gated on donor CD8 T cells in infected (black) or naïve (gray) SPLs. Bar graphs depict MFI of given markers. (D-E) Dot plots are gated on donor CD8 T cells. Bar graphs depict MFI and proportion of TNF-α and IFN-γ production by donors. (F-G) Naïve wild-type and IL-2^−/− CD8 T cells were isolated and stimulated in vitro with anti-CD3 and anti-CD28 for 60 hours. Wild-type CD8 T cells were either untreated (black, bold) or treated with blocking anti-IL-2 antibody (black, non-bold), while IL-2^−/− CD8 T cells were untreated (gray, non-bold), or treated with mIL-2 (gray, bold). Histograms are gated on CD8 T cells, and bar graphs depict MFI of activation markers (F) or transcription factors (G). (H) Heatmap shows MFI fold-change over the control after z-score normalization for the indicated markers (n=4 per group). Data are representative of 2 experiments with n=2 mice per group. Bar graphs show mean and SEM. Paired (A-E) Student’s t-test and one-way ANOVA (F-G) was used to compare more than two groups with significance represented as * (P ≤ 0.05), ** (P ≤ 0.01), *** (P ≤ 0.001). See also Fig. S4.

To confirm the role of paracrine IL-2 in driving early activation and proliferation we used an in vitro system of CD8 T cell priming, where the deleterious effects of systemic in vivo IL-2 ablation are circumvented. Antigen-specific CD8 T cells activated in the absence of IL-2 showed diminished activation (CD25, CD69 and GzmB) and proliferation upon antibody-mediated blockade of IL-2 in TCR-stimulated wild-type P14 cell cultures (Fig 4F, S4C). Notably, IL-2 supplementation led to full recovery of optimal activation and proliferation in in vitro cultures of IL-2^−/− P14 cells stimulated with cognate peptide antigen (Fig 4F, S4D). IL-2-deficient P14 cells expressed significantly lower levels of CD25, exhibited decreased cell cycling (54.9% vs 24.6%, respectively), and were unable to upregulate GzmB production to the same level as their wild-type counterparts (Fig S4D, 4F). Notably, IL-2 supplementation rescued CD25, CD69 expression, and GzmB production in IL-2^−/− cells, albeit not always back to wild-type levels. This suggested that IL-2 from wild-type CD8 T cells was able to further boost perceived paracrine IL-2 signals from neighboring activated cells (Fig 4F). To test if wild-type T cells could produce sufficient IL-2 to rescue their IL-2^−/− counterparts, the wild-type and IL-2^−/− CD8 T cells were cocultured in the presence of cognate antigen. We observed that even in the absence of exogenous IL-2 addition, IL-2 from activated WT CD8 T cells was sufficient to rescue the activation and cell cycling of IL-2^−/− CD8 T cells (Fig S4D). We further queried the differential contributions of autocrine and paracrine IL-2 signals in effector and memory CD8 T cell transcriptional programming during priming by either blocking IL-2 in WT CD8 T cell cultures, or by supplementing with exogenous IL-2 in IL-2^−/− CD8 T cell cultures (Fig 4G). As expected, effector-associated transcription factors Blimp-1 and T-bet were largely dependent on paracrine IL-2 (Fig 4G). However, exogenous administration of paracrine IL-2 in IL-2^−/− CD8 T cell cultures did not fully rescue the defects in Bcl-6, Eomes and cMyc expression, suggesting that autocrine IL-2 signals may collaborate with other inflammatory signals in vivo to fine-tune their expression, as highlighted by largely similar expression patterns in wild-type and IL-2^−/− CD8 T cells isolated at day 8 after infection (Fig 3G, 5C). Together these data show that paracrine IL-2, even from neighboring effectors can sustain the majority of CTL expansion. This lends credence to the notion that paracrine IL-2 is sufficient to sustain the autocrine IL-2 deficient effector CD8 T cell responses in vivo.

Fig. 5.

Potent induction of glycolysis and oxidative phosphorylation by paracrine IL-2. Wild-type and IL-2^−/− CD8 T cells were adoptively transferred into naïve B6 mice. Mice were infected with LCMV_Arm and donors were analyzed during effector expansion (A, D-E) and memory (B) in spleens of infected mice. (A-B) Antigen specific donor cells were purified and sorted from spleens at day 8 (effector; A) and 30 (memory; B) post-infection. (A) Line graphs depict OCR and ECAR. Bar graphs depict basal and spare respiratory capacity (SRC) in oxygen consumption and extracellular acidification. (C) Histograms are gated on donor WT (black), IL-2^−/− (white) and endogenous naïve (gray) CD8 T cells at day 8 post-infection in SPL. Numbers represent MFI in WT (Bold) and IL-2^−/− (Plain) donor cells. Bar graphs depict MFI of cMyc. (D-E) Splenocytes were stimulated direct ex vitro with GP33 (D) or IL-2 (E) and pS6 and pSTAT5 levels were determined in WT (black) and IL-2^−/− (white) antigen specific CD8 T cells at day 6 and 8 post-infection. Bar graphs depict MFI of phosphorylation. Data are representative of 2 (C-E), and 3 (A-B) experiments with n=2-5 mice per group. Bar graphs show mean and SEM. Paired (C-E) and unpaired (A-B) Student’s t-test was used to determine statistical significance. See also Fig.S4.

Effector CD8 T cell state is marked by a unique metabolic program of increased aerobic glycolysis and oxidative phosphorylation (Oxphos) to meet the increased demand for bioenergetics and biosynthetic precursors during rapid proliferation and effector function (Buck et al., 2017). IL-2 is a critical driver of aerobic glycolysis (Oestreich et al., 2014). Given the impact of IL-2 signaling on cMyc expression in vitro (Fig 4G, H), we next sought to distinguish the impact of autocrine and paracrine IL-2 signals on metabolism. Indeed, the early proliferative defects (Fig S4C–D) were associated with a bioenergetic insufficiency caused by decreased aerobic glycolysis in the IL-2^−/− cells as indicated by decreased basal ECAR (Fig S4E). Consistent with this, in vitro IL-2 supplementation augmented basal as well as maximal ECAR of wild-type cells (Fig S4F), thereby suggesting that paracrine IL-2 was dominantly required for CTL metabolism during primary expansion. In vivo as well, paracrine IL-2 signals were able to override lack of autocrine IL-2, resulting in similar glycolytic and Oxphos metabolic programs in wild-type and IL-2^−/− effector (Fig 5A) and memory (Fig 5B) CD8 T cells. Consistent with this, wild-type and IL-2^−/− CD8 T cells engaged the mTOR and STAT5 signaling pathways to similar extents during early and late effector time-points as demonstrated by their phosphorylation levels (Fig 5D–E).

Together, these complementary in vivo assessments of autocrine IL-2-deficient CD8 T cells in an environment replete in paracrine IL-2, and controlled in vitro studies of paracrine IL-2 supplementation, which bypass any confounding effects on Treg cells and other inflammatory mediators, clearly demonstrate that paracrine IL-2 is critical for driving optimal activation, expansion, metabolism, effector differentiation and memory formation, whereas autocrine IL-2 is essential for optimal protective recall expansion potential of memory cells.

Autocrine IL-2 signals are specifically required during primary CTL responses to program optimal recall expansion of memory cells

Having thus established the unique role of autocrine IL-2 in optimal recall expansion of memory CD8 T cells, we next sought to determine when this signaling is critical for memory function. In vivo, IL-2 levels rise systemically following infection as a function of T cell activation during both primary and secondary infections (Kalia and Sarkar, 2018; Liao et al., 2013), and rapidly decline with pathogen control (Kalia et al., 2010). Antigen-specific CD8 T cells capable of IL-2 production alongside other effector cytokines such as IFN-γ and TNF-α following in vitro antigenic restimulation are detectable throughout differentiation (Fig S5D–G), during both primary and secondary CTL responses (Fig S5A–B). We observed that memory-fated CD25^Lo CD8 T cells selectively retained their ability to produce autocrine IL-2 as early as day 3.5 post-infection when CD25 upregulation is the earliest marker of terminal effector fate (Fig S5E). Autocrine IL-2 is correlated to memory fated cells throughout effector expansion and memory differentiation as KLRG-1^Lo CD8 T cells produced the majority of IL-2 (Fig S5C, S5D). To demonstrate that this is a P14 independent phenomenon, we tested several epitopes of LCMV and observed that MPECs produced significantly more IL-2 regardless of antigen specificity (Fig S5C). It is plausible that autocrine IL-2 signals serve to imprint memory-fated CD8 T cells with potent recall potential during primary activation and expansion stages. Alternatively, IL-2 production being a hallmark property of memory CD8 T cells, it is also likely that rapid IL-2 production by memory cells during secondary challenge (Fig S5B) serves as an evolutionarily preserved mechanism to drive rapid recall expansion.

Since the germline il2 gene deletion model precludes any enquiry into the timing of autocrine IL-2 requirement for memory recall expansion, we developed a model of conditional il2 ablation in a subset of antigen-specific CD8 T cells by breeding IL-2^flox/flox Cre-ERT2 mice (Popmihajlov et al., 2012; Owen et al., 2018) to the LCMV D^bGP33-specific TCR transgenic P14 line. We first determined whether antigen-specific CD8 T cells pre-treated with tamoxifen (TAM) prior to infection phenocopied germline il2-deficient CD8 T cells. In vitro priming of untreated (IL-2 sufficient) or TAM pre-treated (IL-2 deficient) CD8 T cells showed that similar to the germline knock-out model, conditional ablation of IL-2 in mature peripheral CD8 T cells led to reduced GzmB and CD25 expression, which was largely rescued by exogenous IL-2 supplementation (Fig S6A). To confirm our in vivo model, we adoptively transferred untreated (IL-2^+/+) or TAM pre-treated (IL-2-ablated) naïve CD8 T cells into naïve B6 mice, which were subsequently infected with LCMV_Arm (Fig S6B). Tissues were analyzed at the peak of effector expansion (day 8 post-infection) (Fig 6A). As expected, TAM treatment effectively ablated IL-2 production by about 6-fold compared to untreated controls (Fig S6C). Recapitulating our observations with germline il2 deficient P14 cells, TAM-treated antigen-specific CD8 T cells accumulated to similar numbers as WT cells at the peak of effector expansion in secondary lymphoid and nonlymphoid tissues (Fig 6A); and expressed similar levels of PD-1, GzmB (Fig 6B), IFN-γ and TNF-α (Fig S6D–E), thus indicating similar TCR signal transduction and effector differentiation. Conditional ablation of il2 also led to comparable levels of SLECs and MPECs at the peak of expansion (Fig 6C), as noted in the germline il2 ablation model. Longitudinal analysis of untreated and IL-2-ablated donor CD8 T cells in the blood also showed largely similar expansion (Fig 6F), transiently increased contraction (Fig S6F), which was followed by largely similar memory numbers as observed in germline il2 ^−/− donor cells (Fig 6G). As in the case of IL-2^−/− P14 cells, tamoxifen-mediated ablation of il2 did not alter the development of CD127⁺ KLRG-1⁻ memory cells (Fig S6G). Importantly, tamoxifen-mediated il2 ablation prior to primary infection also recapitulated the decreased capacity of protective secondary memory responses when rechallenged with LM-GP33 (Fig 6D), albeit the recall expansion differences were slightly lower in the IL-2^flox/flox CD8 T cells compared to the germline deletion setting, possibly due to variable take or downregulation of Cre-ERT2 (Bresser et al., 2020).

Fig. 6.

Autocrine IL-2 signals act during primary CTL responses to program secondary expansion potential of memory cells. Naïve IL-2^fl/fl Cre-ERT2 P14 untreated (UnTx) or tamoxifen (Tam) treated naive CD8 T cells (PreTx) were transferred into naïve B6 mice, which were then infected with LCMV_Arm. (A) Tissues were collected at day 8 post-infection and numbers of donor cells were enumerated in spleen, iLN, lung, and liver. (B) Histograms are gated on donor UnTx (black) or Tam Tx (white) or endogenous naïve (gray) in spleen. Numbers represent MFI of GzmB. Bar graphs depict MFI GzmB and PD-1 at day 8 post-infection. (C) Bar graphs show proportion of MPECs and SLECs on donor UnTx (black) or Tam Tx (white) at day 8 post-infection in spleen. (D) Following memory differentiation donor cells were transferred into naïve B6 mice which were the challenged with LM-GP33. Donors were enumerated in spleens of infected mice at day 8 post-rechallenge. (E) Experiment was repeated at described in A except two additional groups were added where IL-2 was depleted using Tam treatment following day 2 post-infection (Day 2 Tx) or following memory differentiation (memory Tx). (F) Line graph shows frequency of donor T cells in PBMC. (G) Bar graphs depict number of donors in spleen, iLN, and lung following memory differentiation. (H) Memory cells were isolated from spleens of mice in respective groups, normalized and transferred into naïve B6 mice, which were rechallenged with LM-GP33. Bar graphs depict number of donors at day 8 after rechallenge. (I) Bar graphs depict the fold expansion of donors at day 8 after LM-GP33 rechallenge in spleen and liver. (J-K) Bar graphs depict MFI of GzmB (J) and IFN-γ (K) in UnTx, Day 2 Tx and memory Tx donor CD8 T cells from SPLs at day 8 after rechallenge. Data are representative of 2-4 experiments with n=3-5 mice per group. Bar graphs show mean and SEM. Unpaired Student’s t-test (A-D), or One-way ANOVA (F-K) was used to determine statistical significance in difference of means represented as * (P ≤ 0.05), ** (P ≤ 0.01). See also Fig. S6.

We next ablated autocrine IL-2 during distinct times after infection to discern whether autocrine IL-2 signals were necessary during primary or secondary CD8 T cell responses. We utilized timed TAM administration to ablate autocrine IL-2 in GP33-specifc P14 cells either following 2 days of primary infection, or in fully differentiated memory cells prior to secondary challenge (Fig 6E). Tissues were collected from mice following memory differentiation and il2 knockdown was confirmed in the TAM-treated groups (Fig S6I). After primary infection, donor cells expanded similarly, independent of the timing of autocrine IL-2 ablation and accumulated to largely similar numbers at memory in both lymphoid and nonlymphoid tissues (Fig 6F–G). Consistent with similar localization to lymphoid sites, potent upregulation of CD62L (Fig S6H) was noted irrespective of the timing of ablation of autocrine IL-2.

We next compared the recall expansion potential of memory cells when autocrine IL-2 was ablated 2 days after infection, or prior to secondary challenge (Fig 6H). As in Figure 1 for germline il2 deficient P14 cells, we purified wild-type, or IL-2-ablated memory cells and adoptively transferred equal numbers into congenically mismatched naïve mice to determine their recall expansion potential on a per cell basis following heterologous challenge with LM-GP33 (Fig 6H). While we had observed a decrease in pretreated IL-2^fl/fl mice (Fig 6D) as well as germline il2^−/− (Fig 1, 6I) CD8 T cells, day 2 treated and memory treated IL-2^fl/fl memory T cells showed no defect in secondary recall expansion potential in secondary lymphoid tissue (spleen) or peripheral sites of infection (liver), and maintained effector function with unperturbed expression of GzmB and IFN-γ (Fig 6J–K). Wild-type levels of recall expansion when autocrine IL-2 was ablated immediately prior to secondary challenge (Fig 6H–I, S6J) clearly demonstrate that autocrine IL-2 production by memory CD8 T cells during secondary expansion is not required for optimal recall responses. Contrarily, compromised recall expansion when IL-2 was ablated prior to primary infection (Fig 6D), suggests a role of autocrine IL-2 signals in programming of memory recall potential during primary CD8 T cell responses. Normal recall expansion potential of memory cells in which TAM treatment was initiated 2 days after in vivo priming further confirm early programming of recall responsiveness of memory cells (Fig 6I–K).

Collectively, these data demonstrate that autocrine IL-2 is specifically required during priming to program optimal secondary T cell survival.

Induction of autocrine IL-2 production in memory-fated CTLs is independent of CD4 T cell help in inflammatory conditions

CD4 T cell help-mediated licensing of DCs through CD40-CD40L interactions has been proposed to promote memory recall expansion by inducing autocrine IL-2 production, in the poorly inflammatory context of immunization with attenuated Listeria monocytogenes (Feau et al., 2011). Since, the requirement of CD4 T cell help has been shown to vary with the inflammatory nature of the pathogen (Kalia et al., 2016; Swain et al., 2012; Wiesel and Oxenius, 2012), we next sought to determine if CD4 T cell help was necessary to induce autocrine IL-2 production in antigen-specific CD8 T cells during LCMV infection. The effector CD8 T cells generated in the presence or absence of CD4 T cell help were assessed for their ability to produce autocrine IL-2 upon antigenic restimulation. Both TCR transgenic GP33-specific and endogenous CD8 T cells of distinct LCMV specificities were capable of similarly robust IL-2 production upon antigenic restimulation in the presence or absence of CD4 T cell help (Fig 7A–B, S7A). This was associated with robust effector differentiation as indicated by potent production of IFN-γ and TNF-α (Fig 7C, S7B). Considering that optimal CD28 co-stimulation is necessary for IL-2 production, we evaluated the expression levels of MCH II and CD86 on DCs in the presence or absence of CD4 T cell help (Fig S7C). Consistent with similar IL-2 production by CD8 T cells, DCs were similarly activated and expressed similar levels of MHC-II and CD86 in the absence or presence of CD4 T cell help. However, loss of CD28 costimulatory molecule in a subset of antigen-specific CD8 T cells led to reduced autocrine IL-2 production even in a highly inflammatory setting (Fig 7D) and compromised secondary expansion of memory cells as reported previously (Suresh et al., 2001). To further validate the necessity for CD80/86 signals in IL-2 production, we blocked CD28 signals using CTLA-4-Ig during infection and observed a significant decrease in IL-2 production by multiple immunodominant and immunorecessive epitopes of LCMV (Fig 7E). Together, these data show that in the context of inflammatory infections, CD4 T cell help is dispensable for programming of autocrine IL-2 production, owing to effective DC activation and costimulation, thus mediating optimal recall expansion of memory cells (Fig S7D).

Fig. 7.

CD4 T cell help is not critical for autocrine IL-2 production by antigen-specific CD8 T cells under inflammatory conditions. (A, C) Naïve wild-type P14 CD8 T cells were adoptively transferred into naïve wild-type and CD4^−/− mice. Mice were infected with LCMV_Arm and analyzed at day 7 post-infection. Splenocytes were restimulated with GP33 peptide and IL-2, TNF-α and IFN-γ production were compared. Dot plots are gated on donor CD8 T cells. (B) Wild-type and CD4^−/− mice were infected with LCMV_Arm and cytokine production was assessed in spleens of infected mice at day 8 post-infection with indicated LCMV epitopes. Bar graph depicts proportion of antigen specific CD8 T cells capable of producing IL-2 and IFN-γ. (C) Dot plots depict IFN-γ and TNF-α in donor CD8 T cells. (D-E) Wild-type and CD28KO P14 CD8 T cells were co-transferred into naïve B6 mice, which were subsequently infected with LCMV_Arm. Splenocytes were analyzed for cytokine production at day 32 post-infection. (D) Dot plots are gated on donor CD8 T cells. Bar graph depicts proportion of IL-2 producing cells of total IFN- γ + donor cells. (E) Mice were either untreated (WT) or treated with CTLA-4-Ig starting at day 2 post-LCMV infection and followed to memory. Splenocytes were isolated and stimulated with LCMV peptides to assess intracellular cytokine production. Bar graphs show percent of IL-2 producing of IFN- γ + CD8 T cells. Data are representative of 2 experiments with n=3 mice per group. Bar graphs show mean and SEM. Unpaired Student’s t-test was used to determine statistical significance of means represented as * (P ≤ 0.05), ** (P ≤ 0.01). See also Fig. S7.

Discussion

Using a model of temporal il2 ablation selectively in a subset of antigen-specific CD8 T cells, this study establishes a functional role of autocrine IL-2 in programming optimal recall expansion potential of memory cells. For all other cardinal properties of memory cells, such as preferential survival during contraction, migratory properties through secondary lymphoid and nonlymphoid tissues, antigen-independent homeostatic turnover and polyfunctionality, paracrine IL-2 was able to effectively substitute for the lack of autocrine IL-2. Likewise, primary expansion, effector differentiation and the associated metabolic activity were largely driven by paracrine IL-2.

While high dose exogenous IL-2 supplementation has been used in the past to query how IL-2 signals perturb distinct phases of CD8 T cell responses (Blattman et al., 2003), such non-physiological levels of IL-2 are likely to confound the results through pleiotropic effects on multiple immune cells such as Tregs. In this regard, our studies involving analysis of antigen-specific CD8 T cells at endogenous precursor frequencies (such that systemic IL-2 levels are largely unaffected) is of high physiological relevance in preserving the overall immune status. Since only a small fraction of CD8 T cells specific to one particular epitope of LCMV was ablated for autocrine IL-2 signals, systemic IL-2 levels are largely unaffected. Hence, this is a stringent system for delineating autocrine IL-2 dependent CD8 T cell differentiation events and allows for effectively dissecting the relative contributions of paracrine and autocrine IL-2 signals in vivo without perturbing the Treg compartment. Our observations of decreased recall expansion of LCMV-primed memory CD8 T cells deficient in autocrine IL-2 reported here, are replicated in other infections such as attenuated Listeria monocytogenes and vaccinia virus (Feau et al., 2011). Therefore, the role of autocrine IL-2 in promoting memory recall responses is seemingly universal across inflammatory as well as non-inflammatory immunogens (Feau et al., 2011; Redeker et al., 2015). In support, our data show that this crucial autocrine IL-2 function is instilled in the memory precursor lineage as early as day 3.5 post-infection and is maintained in the MPEC lineage throughout memory differentiation. Collectively, these findings provide functional relevance to the close association of IL-2 production potential and memory fate.

Paradoxically, despite a defect in a hallmark property of memory recall function, autocrine-IL-2 deficient CD8 T cells maintained proper balance of effector and memory transcription factor expression during effector and memory states, as well as through secondary recall in vivo. The in vitro model of T cell priming showed that paracrine IL-2 can correct the majority of transcription factor disbalance associated with an autocrine IL-2 deficiency, and thus may explain why differences in these markers are not evident in the paracrine IL-2 replete in vivo setting.

Temporal requirement of autocrine IL-2 in promoting memory recall responses has not been previously evaluated. A recent study showed that increased autocrine IL-2 production in MCMV-primed memory CD8 T cells helps boost numbers of responding T cells in a dose dependent manner during both primary and secondary expansion (Redeker et al., 2015). Since autocrine IL-2 signals were increased during primary as well as secondary expansion, this study was not designed to discern whether autocrine IL-2 is required during primary or secondary responses. Our studies of timed ablation of autocrine IL-2 either before primary or secondary expansion build on the previous data and demonstrate that autocrine IL-2 signals are needed by antigen-specific CD8 T cells during primary expansion. While autocrine IL-2 is critical during CD8 T cell priming, we also show that paracrine IL-2 is sufficient for driving the metabolic upregulation of glycolysis, expansion, and effector differentiation in cytotoxic CD8 T cells.

A three-cell model involving CD4 T cell help to license DCs through CD40 or CD27 costimulation has been presented for optimal recall expansion, and induction of autocrine IL-2 production by CD8 T cells during attenuated Listeria monocytogenes and vaccinia virus infections (Feau et al., 2011). However, we show a robust effector differentiation despite loss of CD4 T cells (a major source of paracrine IL-2) in inflammatory infections such as LCMV. Hence, in the strongly inflammatory LCMV infection, CD4 T cell help is likely dispensable for autocrine IL-2 production. It is possible that sufficient APC mediated CD80/86 signals are capable of driving robust autocrine IL-2 production by CD8 T cells by providing sufficient TCR, inflammatory and/or CD28 signals (Fig 7). Given the critical role of IL-2 in driving effector differentiation, it is possible that localized concentrations of IL-2 at priming sites are high enough to trigger the necessary effector CTL programs. Thus, in addition to confirming the critical role of autocrine IL-2 in driving secondary expansion potential of memory cells in the acute LCMV infection model, this study uses conditional IL-2 ablation models to clearly establish that autocrine signals act during the priming phase to program recall expansion potential in memory-fated cells. Furthermore, our studies provide insight into how autocrine IL-2 production may be regulated by CD4 help in inflammatory viral infections.

Our study also highlights the importance of considering the experimental context such as ratios and overall numbers of wild-type and IL-2-deficient cells (D’Souza et al., 2002) or mixed bone marrow chimerism (Williams et al., 2006), where localized differences in levels of paracrine IL-2 in anatomic microniches might complicate outcomes. For example, migration of antigen-specific CD8 T cells to extralymphoid sites such as lung, liver and gut has been previously shown to be regulated by autocrine IL-2 (D’Souza et al., 2002). As opposed to high dose TCR transgenic cells, our study using physiologic precursor frequencies shows that migration of memory cells to peripheral sites is largely unaffected by presence or absence of autocrine IL-2. This observation is consistent with the notion that paracrine IL-2 signals, as dictated by infection type and numbers of infection-reactive immune cells, exert a dominant role in tissue migration of antigen-specific CD8 T cells.

In conclusion, our study has a bearing on the development of adoptive T cell immunotherapy of cancers and chronic infections. Future studies into the mechanisms promoting autocrine IL-2 production will guide the therapeutic T cell production process towards optimal induction of IL-2 in therapeutic cells for prolonged persistence and efficacious responses in settings where paracrine IL-2 signals might be limiting.

Limitations of the study:

While we observed efficient deletion of il2 in the TAM-induced conditional gene deletion model, reduced IL-2 production due to “leaky” TAM-independent Cre expression might be a potential limitation, which we tried to mitigate by using younger mice and confirming robust IL-2 production in age-matched vehicle-treated mice. Additionally, whether the recall expansion defects associated with loss of autocrine IL-2 may be recoverable by exogenous administration of IL-2, and whether concentrations of IL-2 in microniches might regulate outcomes, remain to be investigated. Lastly, we noted that in vitro antibody blockade of IL-2 in wild-type CD8 T cells maintained subtle differences compared to their autocrine IL-2 deficient counterparts, suggesting that low levels of autocrine IL-2 signaling were present. Exogenous IL-2 administration in IL-2^−/− cells rescued the expression patterns to WT levels in certain cases, but not all transcription factors were similarly rescued (e.g. Eomes, Bcl-6, cMyc). These data suggest that depending on the relative paracrine IL-2 levels in micro anatomical niches, autocrine IL-2 sufficient and deficient CD8 T cells get differentially programmed for their recall expansion potential. It is also plausible that in vivo other inflammatory signals (possibly CD4 T cell dependent and independent), costimulatory signals and duration of antigenic stimulation serve to further fine-tune the memory responses

STAR METHODS

Resource availability

Lead contact

Further information and requests for supporting data and resources should be directed to and will be fulfilled upon reasonable request by the lead contact Vandana Kalia, sarkarkalia@gmail.com

Materials availability

This study generated IL-2^−/− P14 and IL-2^flox/flox Cre-ERT2 P14 mice. The authors declare that all the results supporting the findings of this study are available within the paper and its figures.

EXPERIMENTAL MODELS AND SUBJECT DETAILS

Mice and infection

C57BL/6 (B6; Thy1.2/1.2+ and Ly5.1/5.1+) and IL-2^−/− mice were purchased from the Jackson Laboratory (Sacramento, CA, USA). Thy1.1+ P14 mice bearing the H-2D^bGP_33-41 LCMV epitope-specific TCR were fully backcrossed to B6 mice and maintained in our animal colony. IL-2^−/− and IL-2^flox/flox Cre-ERT2 mice (Mitchell et al., 2010) were crossed to P14 mice and used around 4-5 weeks of age. To generate day 8 effector and memory P14 chimeric mice, 2-4x10³ P14 cells were adoptively transferred intravenously (IV) into naïve mice, ~12 hours prior to infection. To study early priming events in vivo (days 2.75 and 3.5 post-infection), 1x10⁶ P14 cells were transferred into naïve mice. LCMV_Arm was propagated, titered, and used as previously described (Redeker et al., 2015). For primary infections, mice were injected intraperitoneally (IP) with 2x10⁵ PFU LCMV_Arm. Heterologous secondary challenge was performed by transferring 1x10⁴ memory CD8 T cells from LCMV memory mice into naïve congenically mismatched B6 mice, which were then infected with 3x10⁴ CFU Listeria monocytogenes expressing GP33 epitope of LCMV (LM-GP33).

METHOD DETAILS

In vivo treatments

To induce il2 deletion in IL-2^flox/flox Cre-ERT2 mice, tamoxifen (Tam) was administered via oral gavage at 20mg/kg. Tamoxifen (in corn oil) was administered to naïve mice (PreTx), Day 2 infected mice (Day 2 Tx) or memory mice (memory Tx). To confirm il2 deletion, freshly isolated splenocytes or PBMCs were stimulated with GP33 peptide and analyzed for intracellular IL-2. To block CD28 signals, 400μg of CTLA-4-Ig was administered IP on day 2 post-infection and continued every third day.

Flow cytometry

All antibodies were purchased from Biolegend (San Diego, CA, USA). MHC class I tetramers were made as described in NIH tetramer core facility protocol. Cells were stained for surface proteins in 1xPBS containing 1% FBS and 0.05% sodium azide. Intracellular proteins and cytokines were stained using BD intracellular flow kit (BD biosciences) (Redeker et al., 2015). For analysis of intracellular cytokines, 2x10⁶ lymphocytes were stimulated with 0.2μg/mL GP_33-41 peptide in the presence of Brefeldin A (BFA) for 5h, followed by surface staining for CD8, Ly5.1, Thy1.1, and Thy1.2, and intracellular staining for IFN-γ, TNF-α, or IL-2. Staining of intranuclear transcription factors was performed using eBiosciences Foxp3/Transcription Factor Staining protocol. Phosphorylation staining was performed by stimulating antigen specific CD8 T cells with GP33 (0.2μg/mL GP_33-41) or IL-2 (10 ng/mL) in RPMI 1640 with 10% FBS at 37°C in 5% CO2 for 15 min. Cells were immediately fixed with 1.6% formaldehyde, permeabilized, rehydrated and stained with anti-pSTAT5 and anti-pS6. Single cell suspensions of spleen (SPL), inguinal lymph nodes (iLN), lungs (LNG), livers (LVR) or PBMCs were prepared and direct ex vivo staining was carried out (Redeker et al., 2015). LSRII Fortessa (BD Biosciences, San Jose, CA) was used for flow cytometry analysis.

Homeostatic proliferation assay

Wild-type (WT) and IL-2^−/− P14 cells isolated from memory mice were adoptively co-transferred (1x10⁶) into naïve syngeneic recipients. Lymphopenia was induced by sublethal irradiation (450 rad) one day before donor cell transfer. 20 days post-transfer mice were euthanized and donor T cells were enumerated in the spleen.

Seahorse Assay

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured using the XFe96 well Seahorse Analyzer (Agilent). Thy1.1+ cells were purified from spleen at indicated days post-infection and adhered to plates using poly-L-lysine (Sigma) at 1x10⁵ (in vitro activated) or 1.2x10⁵ (direct ex vivo acquired) cells per well. Mitostress test was performed as indicated, in the presence of 10mM glucose (Agilent) with an addition of 2-Deoxyglucose (20mM) to obtain the baseline value of ECAR.

In vitro Activation

Antigen-specific CD8 T cells were purified from naïve or memory spleens using CD8 T cell negative selection Mojo Sort kit (Biolegend). WT, IL-2^−/− or TAM treated IL-2^flox/flox Cre-ERT2 CD8 T cells were stimulated in vitro using GP33 loaded splenocytes or GP33-antigen presenting beads (SA-Beads coated with GP33-MHC I and αCD28 antibody) for 48-60hrs at 37°C. Where indicated, cells were given paracrine mIL-2 at 10ng/mL (PeproTech), or IL-2 signals were blocked using a combination of JES61A12 and S4B6 αIL-2 antibodies at 0.1μg/mL (Bio X Cell).

QUANTIFICATION AND STATISTICAL ANALYSIS

Paired or unpaired Student’s t-test was used as indicated to evaluate differences between means of two groups. One-way ANOVA analysis with a Tukey post-test was used when comparing more than two groups. Nonparametric t-test was used to evaluate differences between two groups with n=3-5 sample size. All statistical analyses were performed using Prism 5 and P values of statistical significance are depicted by asterisk per the Michelin guide scale: * (P ≤ 0.05), ** (P ≤ 0.01), *** (P ≤ 0.001) âd (P > 0.05) was considered not significant (ns).

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Brilliant Violet 650 ™Anti-mouse CD8 α (53-6.7)	BioLegend	RRID:AB_11124344; Cat# 100741
Brilliant Violet 785™ anti-mouse CD45.1 (A20)	BioLegend	RRID:AB_2563379; Cat# 110743
PerCP anti-rat CD90/mouse CD90.1 (Thy-1.1) (OX-7)	BioLegend	RRID:AB_1595487; Cat# 202512
Alexa Fluor® 700 anti-mouse CD90.2 (30-h12)	BioLegend	RRID:AB_493725; Cat# 105320
PE/Cy7 anti-mouse CD127(IL-7Rα) (A7R34)	BioLegend	RRID:AB_1937265; Cat# 135014
PE/Cy7 anti-mouse CD279 (PD-1) (RMPI-30)	BioLegend	RRID:AB_572017; Cat# 109110
Pacific Blue™ anti-mouse CD62L (MEL-14)	BioLegend	RRID:AB_493380; Cat# 104424
PE anti-mouse/human KLRG1 (2F1)	BioLegend	RRID:AB_10574313; Cat# 138408,
APC anti-mouse CD25 (PC61)	BioLegend	RRID:AB_312861; Cat# 102012
Pacific Blue® anti-human/mouse Granzyme B (GB11)	BioLegend	RRID:AB_2562196;Cat# 515408
PE anti-mouse IL-2 (JES6-5H4)	BioLegend	RRID:AB_315302; Cat# 503808
Pacific Blue® anti-mouse IFN-g (XMG1.2)	BioLegend	RRID:AB_893526; Cat# 505818
APC anti-mouse TNF-α (MP6-XT22)	BioLegend	RRID:AB_315429; Cat# 506308
Alexa Fluor® 488 anti-Bcl-2 (BCL/10C4)	BioLegend	RRID:AB_2028390; Cat# 633506
Alexa Fluor® 488 anti-EOMES (Dan11mag)	Thermo Fisher Scientific	RRID:AB_10854265; Cat# 53-4875-82
Alexa Fluor® 647 anti-mouse Blimp-1(/5E7)	BioLegend	RRID:AB_2565618; Cat# 150004
Alexa Fluor® 647 anti-mouse T-bet (4B10)	BioLegend	RRID:AB_1595466; Cat# 644804
PE anti-mouse/human Bcl-6 Antibody (IG191E/A8)	BioLegend	RRID:AB_2561375; Cat# 648304
Rb mAB to c-Myc Alexa Fluor® 647 (Y69)	abcam	RRID:AB_2876372; Cat# ab190560
Alexa Fluor® 488 Anti-Stat5 (pY694) Clone 47/Stat5(pY694)	BD Biosciences	RRID:AB_399881; Cat# 612598
JES6-1A12 Anti-mouse IL-2	BioXcell	RRID:AB_1107702; Cat# BE0043
S4B6-1 Anti-mouse IL-2	BioXcell	RRID:AB_1107705; Cat# BE0043-1
Bacterial and virus strains
LCMV Armstrong	Gifted by Dr. Rafi Ahmed	N/A
Recombinant LM-GP33	Gifted by Dr. Rafi Ahmed	N/A
Biological samples
Chemicals, peptides, and recombinant proteins
Mouse IL-2, recombinant	Peprotech	Cat# 212-12
GP33-41 Peptide	GenScript	Cat# RP20257
Tamoxifen	Sigma-Aldrich	Cat# T5648
Brefeldin A	Sigma-Aldrich	Cat# B7651
Poly- L- lysine solution	Sigma-Aldrich	Cat# P4832
Phosphate Buffered Saline	Thermo Fisher Scientific	Cat# 10010072
Premium Grade Fetal Bovine Serum (FBS)	VWR® Life Science Seradigm	Cat# 97068-091
RPMI Medium 1640	Thermo Fisher Scientific	Cat# 22400-089
Critical commercial assays
Seahorse XF Cell Mito Stress Test Kit	Agilent	Cat# 103015-100
MojoSort™ Mouse CD8 T Cell Isolation Kit	BioLegend	Cat# 480035
FoxP3/Transcription Factor Staining Buffer Set	eBioscience	Cat# 00-5523-00
Fixation/Permeablization Kit	BD Biosciences	RRID:AB_2869008; Cat# 554714
Zombie Aqua− Fixable Viability Kit	Biolegend	Cat# 423102
Deposited data
Experimental models: Cell lines
Experimental models: Organisms/strains
Mouse: C57BL/6	The Jackson Laboratory	Strain: 000664
IL-2flox/flox Cre-ERT2 mice	Mouse model Dr. Kendall A Smith	N/A
IL-2flox/flox Cre-ERT2 P14 mice	This paper
IL-2−/− mice	The Jackson Laboratory	Strain: 002252
IL-2−/− P14 mice	This paper	N/A
CD4−/− mice	The Jackson Laboratory	Strain #002663
CD28−/− P14mice	Mouse model kindly gifted by Dr. Pamela Ohashi	N/A
Oligonucleotides
Recombinant DNA
Software and algorithms
Prism, v 5	GraphPAD	https://www.graphpad.com/scientific-software/prism/
FlowJo, v 9.96	BD Biosciences	https://www.flowjo.com/
Other

Data and code availability

Data reported in this paper will be shared by the lead contact upon request. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.