IL-2 signaling strength dictates T cell biology and therapeutic outcomes in autoimmunity and cancer

Abstract

Interleukin-2 (IL-2) is a critical cytokine for T cell peripheral tolerance and immunity. Here we review how IL-2 interaction with the high affinity IL-2 receptor (IL-2R) supports the development and homeostasis of regulatory T cells and contributes to the differentiation of helper, cytotoxic and memory T cells. A critical element for each T cell population is the expression of CD25 (Il2rα), which heightens the receptor affinity for IL-2. Signaling through the high affinity IL-2R also reinvigorates CD8⁺ exhausted T (Tex) cells in response to checkpoint blockade. We consider the molecular underpinnings reflecting how IL-2R signaling impacts these various T cell subsets and the implications for enhancing IL-2-dependent immunotherapy of autoimmunity, other inflammatory disorders, and cancer.

In Brief

Our knowledge concerning the contributions of IL-2 in T cell biology, particularly its role in the development and functional activity of regulatory, effector, and memory T cells and re-invigorating exhausted T cells continue to advance. These cells depend on distinctive IL-2R signaling thresholds, providing a pharmacological window for the immunotherapy of autoimmunity and cancer. Here we report on the current knowledge of the functions of IL-2 and its receptor in T cells and how this understanding is transforming therapeutic strategies.

Introduction

Since its discovery as a growth factor for antigen-activated T cells, IL-2 and its receptor (IL-2R) have been under intensive investigation for over 45 years. This cytokine–receptor pair critically regulates tolerance and immunity^1,2, largely by affecting regulatory (Treg), effector (Teff) and memory T ^™ cells. Activated T cells, particularly CD4⁺ cells, are the main source of IL-2. Under most physiological conditions, IL-2 exerts its biological effect through the high affinity IL-2R (Kd 10⁻¹¹M), which consists of three subunits, IL-2Rα (CD25), IL-2Rβ (CD122) and the common gamma chain (γ_c; CD132)¹ (Fig. 1A). IL-2 binding to CD25 (K_d 10⁻⁸ M), leads to a small conformational change in IL-2, and the CD25/IL-2 complex is presented in cis for binding to CD122 and CD132³. The high affinity IL-2R is constitutively expressed on Treg cells, is induced by antigen on Teff and T_M cells and is also found on ILC2s⁴ and a small subset of regulatory NK cells⁵. Alternatively, when IL-2 concentrations are high enough, IL-2 binds to cells, e.g., most NK cells and memory-phenotypic CD8⁺ T cells, that express the intermediate affinity (K_d 10⁻⁹ M) IL-2R, consisting of only CD122 and CD132 (Fig. 1A). However, physiologic concentrations of IL-2 at homeostasis are generally not sufficient to provide adequate signaling through CD122/CD132 alone.

Figure 1. The IL-2 receptor and IL-2 signaling pathways.

A) The subunits of the intermediate and high affinity IL-2R and the main cell types that express them. B) The basic signaling pathways downstream of the IL-2R and their affect on Teff versus Treg cells. STAT5 is predominant in Tregs due to high PTEN, which abrogates PI3K-Akt-mTOR. Created with BioRender.com.

Proximal IL-2R signaling depends upon the cytoplasmic regions of CD122 and CD132, which engage three main pathways in Teff cells: STAT5, ERK-MAP kinase, and PI3K-Akt-mTOR⁶ (Fig. 1B). In Treg cells, STAT5 is the dominant pathway due to high levels of PTEN, which attenuates the PI3K-Akt-mTOR pathway⁷. The activation of these pathways requires tyrosine phosphorylation of CD122 and CD132 by JAK1 and JAK3, which recruit the Shc adaptor and STAT5A and STAT5B to the cytoplasmic tail of CD122. Subsequent phosphorylation of STAT5 leads to dimerization and tetramerization of STAT5 that allows localization to the nucleus, where they bind DNA to regulate gene transcription^8,9. Although STAT5A and STAT5B have extensive structural homology and largely overlap in their DNA binding targets, they have some critical non-overlapping functional roles. STAT5B has a dominant role in Treg and Teff cell responses and the asymmetric expression of STAT5A and STAT5B seems to underpin phenotypic differences between each paralog¹⁰. Yet, IL-2R signaling is more complex than proximal JAK1/3-STAT5 activation followed by downstream gene activation. Evaluation of the phospho-proteome of IL-2-stimulated activated CD8⁺ T cells revealed that over 600 intracellular proteins are modified by tyrosine, serine, or threonine phosphorylation¹¹. Of these modifications, 90% are due to JAK-dependent IL-2R signaling, with the remaining 10% dependent on SRC kinases.

Considerable interest remains in IL-2 as it has high therapeutic potential. Low-dose IL-2 offers a means to re-regulate the immune system in the context of autoimmunity, graft vs. host disease (GVHD), and other inflammatory disorders. Higher amounts of IL-2 still hold promise to boost cancer immunotherapy. Indeed, IL-2 remains an important target of bioengineering to enhance its selectivity toward the high affinity IL-2R on Treg cells vs. the intermediate affinity IL-2R on NK and CD8⁺ memory-phenotypic T cells. Reviews have covered in detail engineered IL-2s and recombinant IL-2 (rIL-2) in clinical trials of autoimmunity and cancer^12,13. This review focuses on the mechanisms by which IL-2 impacts the development, homeostasis, and function of Treg, Teff and T_M cells, including how IL-2R signaling strength controls the development of these cells. This review also discusses how our current understanding of IL-2 and its receptor impact efforts to harness IL-2 as a therapeutic agent.

IL-2 in Treg development

Mice that lack IL-2, CD25, or CD122 develop a lymphoproliferative disorder that results in lethal autoimmunity^14–16, illustrating that IL-2R signaling is critical for self-tolerance¹⁷. Initially, CD25 was recognized as a marker of activated Teff cells and later was found to also mark Treg cells¹⁸. Subsequent findings showed that the autoimmunity seen in IL-2/IL-2R-deficient mice was caused by reduced Treg cells and that thymic-directed reconstitution of IL-2R signaling corrected this Treg cell defect¹⁹. Importantly, the transfer of wild-type (WT) Treg cells into CD122-deficient mice fully corrected their lethal autoimmunity, which established that IL-2R signaling is essential for Treg cells²⁰. The interpretation of these initial findings suggested that IL-2 is essential for thymic Treg development²⁰. Others favored that IL-2 was essential for Treg cell survival in the periphery^21–23. We now know that IL-2 plays an integral role in both Treg cell development and survival, as discussed below.

Treg cells develop in the thymus from CD4⁺CD8⁻ “single positive” (CD4SP) thymocytes due to TCR, CD28 co-stimulatory and IL-2R signaling^24,25 (Fig. 2). Positively selected thymocytes migrate into the medulla where high affinity recognition of self-antigens in the context of co-stimulatory molecules causes negative selection²⁶. Cells exhibiting a moderate affinity will survive negative selection and induce CD25 expression, representing potential precursors for Foxp3⁺ Treg cells^27,28. IL-2 from self-reactive T cells acts through the high affinity IL-2R on developing Treg cells to prevent negative selection^29,30 and then drives functional maturation of Foxp3⁺ Treg cells^31,32 largely through STAT5 activation^33–35. Correspondingly, thymic development is aborted in mice with Treg-specific conditional deficiency of CD25^31,32. These developing Treg cells lack functionality, leading to rapid lethal autoimmunity akin to that seen in Scurfy and Foxp3-deficient mice. These studies indicate a strict requirement for IL-2R signaling for functional Foxp3⁺ Treg cells during thymic development.

Figure 2. IL-2 provides two types of essential signals for Treg development in the thymus.

The parallel Treg development model (A) consists of two CD25⁻Foxp3^lo and CD25⁺Foxp3⁻ Treg progenitors (TregP) that each derive mature Treg cells. The two TregP arise from a common pool of CD25⁻Foxp3⁻ CD4SP progenitor cells that also gives rise to self-reactive CD4⁺IL-2⁺ T cells. The serial Treg development model (B) proposes that TregP and self-reactive CD4⁺IL-2⁺ T cells arise from a common CD25⁺Foxp3⁻ precursor population and that fate is determined by duration of agonist signals. Mature Treg cells are derived from a single CD25^loFoxp3^int precursor population. Notably, Foxp3^int expression is achieved independent of IL-2 in the serial model. In both models, IL-2 secreted by self-reactive CD4⁺ T cells provides two signals to TregP for Treg differentiation; Step #1 upregulates CD25 and Foxp3 and Step #2 drives functional maturation of Treg cells. Created with BioRender.com.

Comparison of Treg development in mice with a germline defect in CD25 (CD25^gKO), versus a conditional CD25 deficiency limited to Treg lineage cells, i.e., Foxp3^Cre CD25^fl/fl mice (CD25^cKO), is informative concerning the role of IL-2 during thymic development³¹. First, even though CD25 was effectively deleted in both models, the overall defect in Treg development is more pronounced in CD25^gKO mice, although Foxp3^lo cells are readily detected and some of these cells show STAT5 activation. Thus, the development of Foxp3^lo Treg cells is likely due to the TCR and IL-7R or IL-15R signaling rather than IL-2R signaling (Fig. 2). However, IL-7 and IL-15 are not sufficient to rescue Treg development in mice lacking IL-2, CD25, or CD122^32,36–38. Second, developing Treg cells from CD25^cKO mice have increased Foxp3 compared to CD25^gKO mice, but lower than in WT mice. These findings reflect that IL-2 acts at two steps of thymic development (Fig. 2). At step 1, IL-2 helps to promote the survival and proliferation of developing Treg cells. Continual IL-2R signaling at step 2 leads to functional programming of Treg cells.

The precise stage(s) during Treg development where IL-2 is required and the molecular consequences of its signaling are unclear. Treg precursors (TregP) are modeled to consist of either CD25⁺ Foxp3⁻ or CD25⁻ Foxp3^lo cells, both having passed negative selection³⁹. These two cell populations represent two parallel pathways of Treg development^40,41 (Fig. 2k). The Foxp3⁻ CD25⁺ subset is more dependent on TCR and co-stimulatory signaling, which induce expression of CD25. Conversely, Foxp3^lo CD25⁻ TregP cells are purported to rely on multiple signals for their development, such as IL-15⁴², NF-κB and/or IL-4⁴⁰. Subsequent signaling through the high affinity IL-2R then generates fully functional Foxp3^hi Treg cells from both progenitors. The finding that the TCR repertoire of CD25⁻ Foxp3^lo TregP cells has little overlap with CD25⁺ Foxp3⁻ TregP⁴⁰ cells provides support for this parallel pathway model.

Another study challenges this parallel model and instead suggests that a common CD25⁺ Foxp3⁻ precursor cell develops into CD25^lo Foxp3^int TregP cells, then into functional CD25⁺ Foxp3^hi Treg cells in a serial, stepwise manner⁴³ (Fig. 2B). The generation of CD25^lo Foxp3^int TregP cells is dependent on interruption of TCR and co-stimulatory agonist signals during negative selection, then IL-2R signaling in CD25^lo Foxp3^int TregP cells subsequently generates CD25⁺ Foxp3^hi Treg cells. This is referred to as the ‘primary development pathway’ and is responsible for the majority of Treg development. Notably, initial Foxp3 induction is not dependent on IL-2 in this model, providing an explanation as to why IL-2 or IL-2R germline knockouts still generate sub-functional Foxp3-expressing Treg cells^21,22,24,31. The authors also argue for an ‘alternative Treg development pathway’ that allows for CD25⁺ Foxp3⁻ precursors to induce Foxp3 expression, bypass the CD25^lo Foxp3^int stage, and become mature Treg cells under supraphysiologic concentrations of IL-2. This observation highlights the importance of IL-2 concentrations in driving Treg development from distinct precursors and may reconcile differences between this and the parallel pathway model.

The expression of Foxp3 is critical for enforcing the Treg suppressive program and requires STAT5³³. A conserved non-coding sequence (CNS0) downstream of the Foxp3 promoter includes a STAT5 response element to allow for Foxp3 induction by STAT5 in developing thymocytes^44,45. CNS0 acts in concert with intronic CNS3 to maintain Foxp3 expression⁴⁵, whereas CNS2 maintains the heritability of Foxp3 and requires demethylation for Foxp3 stability^46–48.

The IL-2-dependent transcriptome of thymic Treg cells includes ~2000 genes, indicating that IL-2R signaling broadly affects developing Treg cells³¹. Foxp3 and other well-known IL-2-dependent genes⁴⁹, e.g., Socs1, Socs2, Cish, and Myc, require IL-2R signaling in thymic Treg cells. In addition, most transcripts related to Treg function are reduced after conditional knockout of CD25, consistent with a role for IL-2 in functional maturation³¹. Since transcripts related to TCR are also impacted after CD25 ablation, IL-2 may integrate signaling induced by TCR and co-stimulation during Treg development to generate functional Treg cells³¹.

Along with mainstream thymic development, peripheral naïve CD4⁺ T cells can be activated to become Foxp3⁺ Treg cells. Treg cells induced in vitro are referred to as iTreg cells. When this occurs in peripheral tissues, they are designated as pTreg cells⁵⁰. IL-2 and TGF-β support the development of iTreg cells, which are immunosuppressive in vitro and in vivo^51–53. However, iTregs display methylation of CNS2, which jeopardizes the stability of Foxp3^46,47,54. Notably, expression of Foxp3 and suppressive function after adoptive transfer of iTreg cells in vivo depends on endogenous IL-2 secreted by CD4⁺ Teff cells⁵⁴.

Many mediators, including IL-2, TGF-β, retinoic acid, short chain fatty acids produced by commensal organisms, and food antigens support the development of pTreg cells⁵⁵. TCR repertoire analysis indicates that the large majority of Treg cells are of thymic origin and show little overlap with naive CD4⁺ T cells⁵⁶. In contrast, the TCR repertoire of pTreg cells by necessity overlaps with conventional CD4⁺ T cells and serves a nonredundant role in expanding the specificities of Treg cells⁵⁷. pTreg cells are thought to support tolerance toward food antigens, commensal organisms and a developing fetus^58,59. ILC3-derived IL-2 supports Foxp3 induction and pTreg formation in the intestinal tract, which is dependent on IL-1β from intestinal macrophages sensing commensal organisms⁶⁰. Thus, pTreg cells adapt to environmental cues to maintain self-tolerance.

IL-2 in the peripheral homeostasis of Treg cells

The essential role for IL-2 during thymic Treg development complicates the extent to which peripheral mature Treg cells depend on IL-2. This issue has been resolved. Mice with conditional ablation of CD25 induced selectively in Treg cells^31,61 gradually lose Treg cells over 3-4 months. Thus, without high affinity IL-2R signaling, all Treg cells fail to persist. The lack of persistence is not due to impaired proliferation but enhanced apoptosis, mitochondrial dysfunction, and impaired glycolytic metabolism. Gene expression profiling also points to a role for cholesterol biosynthesis that may support oxidative phosphorylation for IL-2-driven homeostasis of Treg cells^31,62. As proliferation of peripheral Treg cells in the absence of IL-2R signaling is not obviously impaired, other inputs, such as TCR and co-stimulatory signaling or cytokines like IL-7 may help maintain such Treg cells for some time^37,61.

In addition to controlling survival, IL-2 programs Treg cell functions⁶³. Expression of constitutively active STAT5 in Treg cells enhances their suppressive capacity³². Moreover, suppression of experimental autoimmune encephalomyelitis by CD25-deficient Treg cells is less effective when compared to WT Treg cells⁶¹. Conversely, administration of an IL-2 agonist that delivers persistent IL-2R signaling alters Treg cell gene expression that enhances suppressive function, proliferation, oxidative phosphorylation, chromatin accessibility and recruitment of DNA-binding proteins⁶². This type of IL-2R signaling also supports expression of chemokine receptors that facilities trafficking of Treg cells into inflamed tissues⁶⁴.

The Treg lineage is highly stable and self-renewing at homeostasis, as evidence by fate-mapped Treg cells that persist for over 8 months⁶⁵. Targeted Foxp3 deletion in mature Treg cells causes them to become IFNγ- or IL-2-producing Teff cells, illustrating the central role of Foxp3 in maintaining Treg lineage⁶⁶. Inflammation destabilizes Treg cells, whereby they lose suppressive function and/or Foxp3 expression and become Teff cells, or ‘ex-Treg cells’⁶⁷. During inflammation, the relative amounts of inflammatory cytokines and IL-2 may determine Treg fate. When Treg cells express a hypomorphic variant of CD25, they are less effective in suppressing colitis, become unstable and lose lineage commitment⁶⁸. In contrast, under conditions without obvious inflammation, peripheral Treg cells that lack IL-2R signaling for many weeks readily express near WT amounts (80-90%) of Foxp3 and exhibit minimal signs of conversion to ex-Treg cells in secondary lymphoid tissue and the gut mucosa^31,61. The intestine is an important site of Treg instability, particularly with conversion into T_H17 cells⁶⁹. Thus, the loss of IL-2R signaling does not inherently promote Treg instability, but rather may depend on an accompanying inflammatory response.

Mature Treg cells in the periphery display extensive heterogeneity in their activation, localization, survival, proliferation, and suppressive function. These are shaped by numerous molecular factors, one of which is IL-2^70–73. CD62L^hi CCR7⁺ resting or central Treg (cTreg) cells reside in secondary lymphoid tissues, whereas CD62L^lo ICOS^hi activated or effector Treg (eTreg) cells home to tissues^71,72,74,75. When compared to cTreg cells, eTreg cells exhibit higher proliferation and suppressive function⁷⁰. Although both Treg subsets depend on IL-2 for their maintenance, cTreg cells decline more rapidly than eTreg cells in the absence of peripheral IL-2R signaling^31,32,61,74. This longer persistence of eTreg cells likely reflects signals from the TCR and CD28. Klrg1⁺ eTreg cells are terminally differentiated eTreg cells that develop after 8-9 cells divisions and are highly dependent on IL-2⁷⁶, perhaps due to the direct action of IL-2, synergy of IL-2 and TCR signaling, or IL-2-dependent programming in cTreg cells.

T follicular regulatory (T_FR) cells have a crucial role in germinal center homeostasis by mitigating the generation of self-reactive B cells. IL-2/STAT5 engagement of Blimp-1 promotes IL-10-producing eTreg cells, and suppresses Bcl-6 to prevent T_FR development⁷⁷. Thus, low or absent IL-2 signaling is required for Bcl-6 induction and T_FR differentiation⁷⁷. Bach2 also acts to restrain STAT5 activation and promote T_FR differentiation⁷⁸. Whether IL-2 signaling limits development of RORγt⁺ Foxp3⁺ T_R17 cells⁷⁹, the counterpart to T_H17 cells that develop in environments with low IL-2⁸⁰, remains to be determined.

Cellular sources of IL-2

Treg cells require paracrine IL-2 because Foxp3 represses IL-2 transcription⁸¹. Although antigen-activated CD4⁺ T cells are considered the main producers of IL-2, other cell types, including dendritic cells (DCs), B cells, CD8⁺ T cells and ILCs produce IL-2 at lower amounts⁸². Thymic Treg cells adapt to limited IL-2 by a somewhat higher sensitivity to IL-2 than peripheral Treg cells⁸³. Treg cells are found in a niche within the thymic medulla where they are thought to receive IL-2 signals from DCs⁸⁴. However, conditional deletion of IL-2 in T cells, but not DCs, impairs Treg development^41,85, consistent with most IL-2 secretion coming from newly generated CD4⁺ self-reactive thymocytes⁸⁵. IL-2 production by self-reactive T cells is induced by strong TCR signaling in a niche whereby developing Treg cells constrain the development of additional IL-2-secreting cells, which in turn also limits Treg development⁴¹.

Thymic DCs express CD25, but not CD122, and therefore do not respond to IL-2. Rather, CD25-bearing DCs may bind IL-2 produced by self-reactive T cells. Although IL-2 binding solely to CD25 has fast on and off kinetics and is unstable, this type of interaction of IL-2 with DCs may help to retain IL-2 in the DC-Treg niche and potentially trans-present IL-2 to Treg cells^85,86.

Cell-targeted deletion of IL-2 also demonstrates that the main source of IL-2 contributing to Treg homeostasis in the periphery is largely T cells⁸⁵ (Fig. 3), with a possible limited role for DCs in mucosal-associated draining lymph nodes, such as the mesenteric lymph node⁸⁵. This T cell-derived IL-2 for Treg cells does not rely on activation by foreign antigen, but rather by self-antigens⁸⁷. Imaging studies of lymph nodes show that IL-2-responsive Treg cells, as measured by activated STAT5, are found in discrete clusters with rare IL-2-secreting self-reactive T cells and DCs⁸⁷. Consequently, Treg cells in the cluster express high levels of suppressive molecules and preferentially consume IL-2 to actively suppress the IL-2-secreting autoreactive T cells. Thus, IL-2 from autoreactive T cells contributes to Treg homeostasis, but this co-localization of Treg cells with autoreactive T cells provides a feedback circuit whereby Treg cells suppress and reduce the frequency of those self-reactive T cells (Fig. 3). This feedback circuit represents one important way in which T cell immune homeostasis and tolerance is maintained.

Figure 3. Peripheral homeostasis is maintained through Treg sequestration of IL-2.

Autoreactive PD-1⁺ CD4⁺ T cells secrete IL-2 upon binding self-antigen presented by dendritic cells. Treg cells bind and sequester this IL-2 to support their homeostasis and suppressive activity and limit IL-2R signaling in autoreactive CD4⁺ T cells. When IL-2 concentrations become high enough to sustain IL-2R signaling in autoreactive T cells, tolerance is broken. Created with BioRender.com.

IL-2 in the differentiation of CD4⁺ T helper cell subsets

Naïve CD4⁺ T cells are clonally-restricted precursors that differentiate into a wide array of specialized effectors upon encounter with their cognate antigen, co-stimulation and cytokine cues. When naïve T cells encounter their antigen in the context of co-stimulation, their initial activation is characterized by IL-2 secretion and expression of CD25 and CD122, making them responsive to autocrine and paracrine IL-2. Binding of IL-2 in the context of TCR signaling promotes T cell proliferation, but also contributes to programming the differentiation of helper and cytotoxic subsets⁸⁸. These subsets can be broadly grouped into T follicular helper cells (T_FH), T helper subsets (T_H1, T_H2 and T_H17), and cytotoxic T lymphocytes (CTLs), although this is not comprehensive, and considerable heterogeneity and plasticity exists with each of these groups⁸⁹.

T_H1 cells are important for immunity against viral and intracellular bacterial pathogens. IL-2 drives T_H1 survival and differentiation by increasing expression of IL12RB2 through activated STAT5, allowing IL-12 induction of T-bet via STAT4^90,91. IL-2 also remodels the metabolism of activated cells by engaging HIF-1α and c-Myc to drive glycolytic activity, which subsequently augments IFNγ secretion, the signature T_H1 cytokine^92–94. This is likely mediated through IL-2-dependent activation of mTORC, which contributes to increased T-bet expression through protein kinase C^95,96.

T_H2 polarization is required for protective immunity against parasites while excessive T_H2 activity may predispose one to asthma and allergy. The T_H2 response is engaged by IL-4, which activates STAT6 to induce GATA3, the T_H2 lineage-determining transcription factor^97,98. IL-2 contributes to T_H2 cell differentiation by STAT5-mediated upregulation of IL4RA and by increasing the accessibility of the IL4 locus^99,100.

If IL-2R signaling in recently activated CD4⁺ T cells remains high, the prolonged activation of STAT5 and Blimp-1 suppresses the expression of Bcl-6, which opposes differentiation into T_FH cells^101–105. In contrast, recently activated T cells that receive strong TCR and IL-6 signals decrease CD122 expression and become hyporesponsive to IL-2. This setting, coupled with sequestration of IL-2¹⁰⁶, establishes an environment with low amounts of IL-2 signaling, which favors T_FH differentiation¹⁰⁷. Once Bcl-6 is expressed, it directly represses the HIF-1α - and c-Myc-driven glycolytic activity involved in T_H1 function⁹³.

T_FH cells are critical in the development of germinal centers to shape antibody responses to infectious agents and vaccination¹⁰⁸. Correspondingly, modulation of IL-2 affects this response. For example, the levels of IL-2 are relatively high in neonates from autoreactive CD4⁺ T cells prior to Treg cell egress from the thymus. When IL-2 is transiently blocked with anti-IL-2 during this time, the T_FH response is augmented, leading to improved immunity to respiratory syncytial virus (RSV) and faster clearance of subsequent RSV rechallenge¹⁰⁹. Furthermore, high T_FH/Treg ratios correlate with autoimmunity. Exogenous low-dose IL-2 lowers this ratio, likely by limiting T_FH differentiation¹¹⁰ but does not elevate other T helper responses.

T_H17 cells normally function to protect the host from extracellular bacteria and fungi but upon hyperactivation, are implicated in multiple autoinflammatory processes. T_H17 are characterized by their secretion of IL-17 and expression of the transcription factor RORγt¹¹¹. Like T_FH, the presence of IL-2 also restrains the differentiation of T_H17 cells. Some mutations in IL2RA result in hyperresponsiveness to IL-2 and are associated with inflammatory bowel disease due to increased CD4⁺ T_H1 polarization, higher concentrations of IFNγ and a loss of T_H17 cells in the intestine^112,113. Mechanistically, IL-2R signaling leads to accumulation of STAT5 binding at the IL17 locus, which outcompetes STAT3, inhibiting transcription of IL17¹¹⁴. PTEN inhibits IL-2-mediated suppression of IL17 by enforcing the phosphorylation of STAT3 over STAT5¹¹⁵. Additionally, Batf antagonizes STAT5 recruitment of Ets1-Runx1 to T_H1 and Treg-specific gene loci, thereby favoring Th17 differentiation¹¹⁶. Similar to T_FH, low-dose IL-2 decreases the ratio of T_H17/Treg in patients with autoimmune diseases, which correlates with improved therapeutic responses¹¹⁰. Thus, along with its essential role of IL-2 for Tregs, IL-2 also critically shapes proliferation and differentiation of CD4⁺ Teff cells with specialized functional activity.

IL-2 in differentiation of CD8⁺ Teff, T_M and exhausted (Tex) cells

In the context of primary infection or antigen exposure, naïve CD8⁺ T cells undergo expansion, contraction, and memory phases. The balance between short-lived Teff and memory precursors is regulated by duration of antigen exposure and inflammatory cytokines as well as the concentration and timing of IL-2 (Fig. 4B). CD8⁺ Teff cells with high amounts of CD25 during the primary response exhibit robust proliferation and develop high effector function yet are more prone to apoptosis¹¹⁷. In contrast, activated CD8⁺ T cells that express CD25 at relatively low amounts during the primary response generate moderate IL-2R signaling, resulting in CD62L⁺ CD127⁺ T_M cell development¹¹⁷.

Figure 4. CD25 levels tune immune cell responses to endogenous and exogenous IL-2.

A) The ability of IL-2 to promote or inhibit the functional activity of the indicated cells depends on the concentration of IL-2 and the levels of CD25. B) High IL-2 causes differentiation of short-lived effector CD8⁺ T cells. Moderate IL-2 supports differentiation into memory precursors. Programming memory recall responses requires autocrine IL-2 signaling. Exogenous IL-2 directed to the high affinity IL-2R induces highly functional effector-like cells from the Tex lineage; exogenous IL-2 directed to the intermediate affinity IL-2R leads to proliferation of stem-like Tex cells, which can give rise to terminally differentiated Tex cells.Created with BioRender.com.

IL-2-dependent regulation of Eomes and Blimp-1 contribute to the T_M vs. Teff fate decision during priming^118–121. Moderate IL-2R signaling acts through Akt and STAT5 to support expression of Eomes, and coupled with Bcl-6 and Id3, promotes T_M cells^{120,122–124}. High IL-2R signaling induces expression of Blimp-1, and coupled with T-bet and Id2, supports development of highly functional short-lived CD8⁺ Teff cells^117,124. Since Bcl-6 and Blimp-1 reciprocally repress each other, the T_M fate is enforced by Bcl-6 repressing Blimp-1. At higher IL-2R signaling, greater levels of Blimp-1 are induced, leading to the repression of Bcl-6, to drive development of CTLs. Another consequence of IL-2-dependent STAT5 activation is increased expression of granzyme B and perforin-1, which are mediators of CTL activity^120,125,126. Activated CD4⁺ T cells are not only cytokine-producing helper cells but under some circumstances have the potential to develop into CTLs^127–129, seemly independent of the CD4⁺ T helper program. The same pathways active in activated CD8⁺ T cells, as discussed above, also function in CD4⁺ T cells to generate CTL activity^127,130.

CD4⁺ Teff cells are canonically the main source of IL-2 for CD8⁺ Teff development. However, IL-2-dependent recall expansion by memory cells can form in the absence of CD4⁺ T cell “help”^131,132. A subset of antigen-activated CD8⁺ T cells produce autocrine IL-2 early in the response after initial priming^133–135 (Fig 4B). This autocrine IL-2-dependent STAT5 signaling supports differentiation into stem-like memory cells while protecting these cells from terminal Teff differentiation and eventual cell death^135,136.

The CD8⁺ Tex lineage consist of dysfunctional T cells, where chronic antigen exposure results in high expression of co-inhibitory receptors (IRs), low effector function, and reduced proliferative capacity^137,138. CD8⁺ Tex cells lose the ability to produce IL-2 followed by TNFα and IFNγ, whereas expression of granzyme B is more variable^137–140. The CD8⁺ Tex lineage consists of two major developmental stages (Fig. 4B). TCF1^hi progenitor-or stem-like Tex cells exhibit low effector activity and are the major population that expands to anti-PD-1/PD-L1 therapy^141–145. These cells develop into TCF1^lo terminally differentiated Tex cells, which are short-lived and express higher levels of PD-1, other IRs, and granzyme^141–145.

The strength and duration of IL-2R signaling influence CD8⁺ Tex cell development. In contrast to CD25, relatively high levels of CD122 persist on CD8⁺ Tex cells^117,146. Infrequent exposure to IL-2 and transient STAT5 activation promotes development into the Tex lineage through Blimp-1 and subsequent induction of IRs¹⁴⁶. In mouse tumor models, intermittent IL-2 and transient STAT5 activation causes an accumulation of 5-hydroxytryptophan, leading to nuclear translocation of the aryl hydrocarbon receptor (AhR)¹⁴⁷. Transcriptional regulation by AhR increases IR expression while decreasing cytokine production, both hallmarks of CD8⁺ Tex cells¹⁴⁷. Membrane-spanning sushi domain-containing protein 2 (SUSD2) also acts to promote CD8⁺ Tex development through its action as a negative regulator of STAT5 signaling¹⁴⁸. Correspondingly, after deletion of SUSD2, tumor antigen-specific CD8⁺ T cells and CAR T cells elicit more productive antitumor responses than their WT counterparts¹⁴⁸.

The action of some engineered IL-2s also supports a role for IL-2 in Tex development. When compared to WT IL-2, the IL-2 partial agonist H9T induces lower STAT5 activation but does not affect IL-2R-dependent PI3K-Akt or ERK-MAPK signaling¹⁴⁹. This attenuated IL-2 activity favors development of TCF1⁺ CD8⁺ stem-like cells that express lower levels of exhaustion markers and supports more effective antitumor responses (Fig. 4B). These types of effects were recapitulated in additional studies with other engineered IL-2s with selectivity toward CD122/CD132^150,151. These IL-2 agonists promote low moderate IL-2R signaling and antitumor activity by Tex cells, but they do not lead to transcriptional reprogramming and full recovery of Teff activity.

Persistent IL-2R signaling in combination with anti-PD-1 more effectively reverses the Tex state by inducing highly functional CD8⁺ Teff cells with robust antitumor and antiviral capacity^152–156 (Fig. 4B). These enhanced responses are due to epigenetic reprogramming of TCF1⁺ Tex stem cells leading to a unique transcriptional signature similar to canonical Teff cells. To achieve this effect, IL-2 is delivered to the TCF1⁺ Tex cells through their expression of CD25 or by linking a CD122/CD132-selective IL-2 mutein to anti-PD-1^{152–154,156}. Although anti-PD-1 alone minimally affects the gene program of Tex cells, these IL-2-dependent effects on TCF1⁺ Tex cells synergize with anti-PD-1, leading to expansion of reinvigorated Tex cells. This effect, even in the PD-1-directed IL-2 variant, is a combination of both anti-PD-1- and IL-2-specific effects, but most cellular changes seem to be driven by IL-2. Similar to providing persistent IL-2R signaling, transduction of LCMV-specific CD8⁺ T cells with constitutively active STAT5A antagonizes the Tox-dependent Tex lineage, rescues polyfunctionality, and induces durable effector-NK-like cells from CD8⁺ Tex cells¹⁵⁷. Thus, IL-2R signaling through STAT5 appears to reprogram TCF1⁺ Tex cells to reinvigorate their response to anti-PD-1 to substantially enhance their functional activity.

The molecular basis for distinct IL-2R signaling thresholds in the development of Treg, Teff and T_M cells.

Cell surface amounts of CD25 are one important factor that impacts the sensitivities of T cells to IL-2^158,159. The stimulation of Treg cells and naïve CD4⁺ and CD8⁺ T cells by TCR and IL-2R signaling results in very high levels of IL-2R subunits in vitro. However, the expression of these subunits is much more nuanced in vivo (Fig. 4A). At homeostasis, mouse and human Treg cells express the highest amount of CD25, likely through continual sensing of self-antigen and IL-2, and consequently, these cells express the highest amount of the high affinity IL-2R. In normal mice, very few other T cells express the CD25 whereas in healthy humans, some naïve and memory-phenotypic CD4⁺ and CD8⁺ T cells express low amounts of the high affinity IL-2R. After naïve CD4⁺ and CD8⁺ T cells encounter antigen in vivo, CD25 and CD122 are induced, resulting in the expression of the high affinity IL-2R¹⁶⁰, but these levels are typically lower than detected after stimulation in vitro. Thus, under most circumstances, even during immune activation, Treg cells express the highest amounts of the high affinity IL-2R, although some Teff cells may transiently express similar levels.

High levels of the high affinity IL-2R allow Treg cells to preferentially respond to IL-2. When human PBMCs were examined, proximal IL-2R signaling, as measured by activation of STAT5, was induced with 10-20-fold lower levels of IL-2 in Treg cells than CD45RO⁺ memory-phenotypic CD4⁺ T cells^83,161. Importantly, downstream gene activation, which requires integration of IL-2R signaling for a period of time, required 100-fold less IL-2 in Treg cells^83,171. This high level of sensitivity to IL-2 forms the basis of using low-dose IL-2 to boost Treg cells in autoimmune diseases.

High CD25 expression does not fully account for the low IL-2R signaling threshold in Treg cells¹⁶². Serine/threonine protein phosphatase 2A (PP2A) activity is enhanced in Treg cells, which potentiates proximal IL-2R signaling^159,163. PP2A directly enhances IL-2R signaling by preventing the loss of CD122 through limiting Adam-10-dependent cleavage of cell surface CD122. PP2A also enhances gene expression in Treg cells related to survival, activation, immunosuppressive function, and IL-2R signaling. Similarly, Mst1 and Mst2 are serine-threonine kinases that are essential to maintain Treg sensitivity to IL-2 by promoting STAT5 activation¹⁶⁴.

Mutations of three key tyrosine residues to phenylalanine on the intracytoplasmic tail of CD122, which are critical to the interaction of the Shc adapter and STAT5 to CD122, lead to markedly diminished, but not abrogated, proximal IL-2R signaling in T lymphocytes⁸³. Notably, thymic Treg cell development and peripheral homeostasis is outwardly normal whereas Teff cells are markedly impaired. These mutations affect the expression of 300-400 genes in Treg and Teff cells. However, key IL-2-dependent genes of Treg cells, e.g., Foxp3, Il2ra, and Socs2, are normally expressed. Thus, low proximal IL-2R signaling is sufficient to support some essential elements required for Treg but not Teff cells. Similar results are reported for mice harboring a point hypomorphic mutation in CD25 or treated with anti-IL-2, leading to reduced CD25 expression and proximal IL-2R signaling^68,162. Mice with hypomorphic CD25 show relatively normal Treg development under homeostatic conditions, but their Treg cells do not effectively suppress colitis in the T cell transfer model. This reduced control of autoimmunity may be due to inflammatory conditions and/or autoreactive T cells with normal IL-2R signaling.

Another feature of Treg cells is their low expression of the IL-7R¹⁶⁵, in part due to Foxp3 repression of IL7RA (CD127)¹⁶⁶, which limits their responsiveness to IL-7. The IL-2R and IL-7R share CD132 as a receptor subunit. Surface plasmon resonance showed that CD127 binds to CD132 in the absence if IL-7¹⁶⁷. Thus, IL-7 competition for and/or CD127 sequestering of CD132 may act to impede IL-2R signaling for naïve and memory T cells or memory Treg cells that express a high amount of CD127, while this mechanism is less relevant for CD127^lo Treg cells^166–168. Such competition may also fine-tune the response of Teff and T_M cells, where high vs. moderate strength of IL-2R signaling support CD8⁺ Teff vs T_M cell fate decision, respectively.

Single nucleotide polymorphisms (SNPs) in IL2, IL2RA, and IL2RB are linked to susceptibility to several autoimmune diseases. Most SNPs are found in enhancers, consistent with their effects on altering gene transcription¹⁶⁹. The precise role of SNPs in autoimmunity is difficult to discern as their effects are subtle and likely promote autoimmunity in conjunction with other SNPs and environmental triggers. With respect to IL2RA, one SNP (rs12722495) associated with type 1 diabetes (T1D) lowers CD25 expression in CD4⁺ Teff and Treg cells, resulting in lower IL-2-dependent STAT5 activation and Foxp3 expression by Treg cells¹⁷⁰. Low Treg suppressive activity due to this SNP likely contributes to T1D. Another SNP (rs61839660) associated with Crohn’s disease delays the expression of CD25, which in turn increases inflammatory T_H17 cells^113,171,172.

TCR and IL-2R signaling are major inducers of IL2RA transcription. Although TCR signaling is sufficient to induce CD25 expression, IL-2 further enhances CD25 amounts. TCR regulates IL2RA in part through activation of NFAT, NF-κB, and AP-1 whereas IL-2 regulates IL2RA through STAT5¹⁷³. The IL2RA locus is one of the most STAT5-responsive genes, containing 17 binding regions in humans⁴⁹. In Treg cells, IL2RA transcription is further enhanced and maintained by Foxp3 in coordination with Runx1 and NFAT^174–176. Binding to the locus requires the coordination of STAT5 super-enhancer sequences found upstream and downstream of the promoter region. The extent that these enhancers bind STAT5 dictates the amount of IL2RA transcription and cell surface expression to fine-tune responses to environmental levels of IL-2 (Fig. 4A).

Immunotherapy of autoimmunity and cancer with IL-2

There remains much interest and activity to harness IL-2 to manipulate T cells in patients with autoimmunity and cancer. Two main principles driving this effort are: (1) selective targeting of the high affinity IL-2R on Treg cells with low-doses of IL-2; (2) higher doses targeting the intermediate affinity IL-2R on memory-phenotypic CD8⁺ T and NK cells can boost anti-cancer responses while minimizing the off-target expansion of tumor-infiltrating Treg cells associated with poor prognosis¹⁷⁷. In both situations, IL-2 or IL-2R has been engineered in various ways to enhance selectivity toward the high vs. intermediate affinity IL-2R. The five main classes of engineered IL-2s include: muteins, IL-2/anti-IL-2 complexes, PEGylated IL-2, IL-2/CD25 fusion proteins, and orthogonal IL-2. For details, refer to reviews on these biologics and clinical testing^12,13. Below discusses some of the issues uncovered in applying IL-2 to immunotherapy of autoimmunity and cancer.

Low dose (LD) rIL-2 was first used in clinical trials for chronic GVHD and hepatitis C virus-induced cryoglobulinemic vasculitis^178,179. Such trials have also been completed for patients with T1D, alopecia areata, rheumatoid arthritis, systemic lupus erythematosus (SLE), psoriasis, granulomatosis with polyangiitis, Takayasu’s disease, and inflammatory bowel disease, and many others are ongoing^13,180,181. From these trials, LD rIL-2 is 1-1.5 million IU of rIL-2 administered at frequencies ranging between daily to every other week. However, somewhat higher doses are administered in chronic GVHD¹⁸². These trials indicate that LD rIL-2 therapy is safe and increases Treg cells in all patients but at variable levels through an undetermined mechanism. Many patients show clinical improvement, but not complete resolution of disease. Importantly, self-reactive T cells are not reactivated, reflecting that human Treg cells are at least 100-fold more responsive than CD4⁺ CD45RO⁺ Teff cells to low amounts of rIL-2¹⁶¹. Treg cells from patients undergoing LD rIL-2 therapy exhibit a more activated phenotype and are more suppressive in vitro. Some evidence also suggests increased thymic Treg cell output and lower abundance of T_FH and T_H17 cells^110,183, but the enhancement of the Treg compartment is thought to be the primary therapeutic mechanism.

Several observations indicate that LD rIL-2 may have some limitations. First, although clinical symptoms improved, LD rIL-2 has not shown off-drug efficacy¹⁸². Thus, LD rIL-2 is unlikely to induce long-lasting immune tolerance. Second, clear indication of clinical efficacy has not been robustly established in several phase 2 trials in SLE and alopecia areata^180,184,185. Third, besides Treg cells, ILC2s¹⁸⁶ and the CD56^hi subpopulation of NK cells^187,188 also respond to LD rIL-2. Both populations express the high affinity IL-2R and their expansion is an off-target effect (Fig. 4A). Virtually all trials to date have relied on using rIL-2, which has poor pharmacokinetics and pharmacodynamics that require frequent administration to maintain Treg increases.

The development of novel IL-2 agonists may enhance the efficacy of LD IL-2 and several are in clinical testing^12,13. The IL-2 muteins (efavaleukin, RG-7835, CC-92252), which are fused to Fc, or pegylated IL-2 (NKTR-358), show more selective binding to the high affinity IL-2R by lowering the interaction with CD122 and may reduce off-target effects^189,190. However, RG-7835 and CC-92252 did not meet efficacy criteria perhaps due to altered binding to the IL-2R. The IL-2/CD25 fusion protein targets the high affinity IL-2R through slow and steady dissociation of IL-2/CD25 monomers from more numerous transdimers and achieves selectivity toward the high affinity IL-2R by limiting the availability of active monomers¹⁹¹. IL-2/CD25 (BMS-986326) is currently being tested in normal subjects (NCT04736134) and patients with SLE (NCT06013995).

Off-target responses by ILC2s and CD56^hi NK cells may be difficult to avoid, even with engineered IL-2s. ILC2-derived IL-5 promotes inflammation and is likely responsible for the eosinophilia that accompanies LD rIL-2 therapy¹⁸⁶. However, ILC2s produce amphiregulin, which may facilitate tissue repair¹⁹². In addition, IL-2 in conjunction with retinoic acid and IL-33 induces a regulatory subset of ILC2s that secretes IL-10 that may decrease inflammation and represents another potential beneficial off-target effect of LD rIL-2¹⁹³. CD56^hi NK cells produce IFNγ, which is not desirable during LD rIL-2¹⁸⁷, yet CD56^hi NK cells exhibit anti-proliferative activity on Teff cells, which could also be beneficial¹⁹⁴. The extent that stimulation of these cells is a detriment or a benefit during LD IL-2 needs to be carefully considered.

rIL-2 was the first approved immunotherapy for use in renal cell carcinoma and melanoma^195,196. A few (~5-10%) patients were cured of these cancers, but most did not respond to therapy. Checkpoint blockade is the immunotherapy of choice for these and other malignancies due to improved response rates and relatively fewer side effects. Yet, many patients do not respond to checkpoint blockade, and for those that do, remission does not result in a cure. Thus, there remains considerable interest to use IL-2 in cancer immunotherapy due to its ability to promote development and expansion of highly functional Teff cells. Most efforts are focused on using engineered IL-2s to improve bioavailability, limit toxicities, and direct their activity toward lymphoid cells that express the intermediate affinity IL-2R, largely memory phenotypic CD8⁺ T and NK cells^197,198. These types of engineered IL-2 agonists limit expansion of Treg cells, which have the potential to suppress antitumor responses. Another approach is to target IL-2 or engineered IL-2s to the tumor microenvironment, often by linking it to tumor-targeting antibodies or by inducing the release of bioactive IL-2 due to properties inherent to the tumor, such as tumor-specific proteases or high pH¹³.

Numerous trials are underway testing these novel IL-2-based approaches^13,199. Several trials offer cautionary notes concerning IL-2s with selectivity for the intermediate affinity IL-2R. In a phase 3 trial, patients with metastatic melanoma were treated with anti-PD-1 (Nivolumab) or the combination of Nivolumab and a pegylated IL-2 (Bempegaldesleukin) with selectivity toward the intermediate affinity IL-2R. The combination therapy offered no benefit beyond what occurred with anti-PD-1 monotherapy (NCT03635983)²⁰⁰. A similar phase 3 trial (NCT03729245) in patients with metastatic renal cell carcinoma also did not reveal added benefit from the combination group. Some of these disappointing results might reflect the pharmacokinetics and pharmacodynamics specific to the engineered IL-2s used.

The IL-2-dependent reversal of Tex cells in conjunction with anti-PD-1 in preclinical models (as discussed above) revealed highly effective control of antitumor responses^154,156. In a side-by-side comparison between IL-2/CD25 (selective toward the high affinity IL-2R), and IL-2/anti-IL-2 complexes (selective toward the intermediate affinity IL-2R), the former developed substantially more effective antitumor responses as a monotherapy or in combination with anti-PD-1¹⁵⁴. Although Treg cells may expand in the periphery, notably, the ratio of CD8⁺ T cells to Treg cells in the tumor microenvironment was high¹⁵²–^154,156. Indeed, this might reflect a goldilocks situation, where an IL-2 analog directed to the high affinity IL-2R promotes robust antitumor responses with checkpoint blockade while an accompanying increase in Treg cells in the periphery suppresses off-target autoimmune-like effects that often accompany checkpoint blockade²⁰¹. The critical issue is whether these promising preclinical studies translate when tested in patients with cancer.

Concluding remarks

We have a very detailed understanding concerning the role of varied IL-2R signaling for the development and homeostasis of Treg cells and for the development of various helper and cytotoxic T cell subsets. The IL-2R also provides important signals related to development of CD8⁺ T_M and Tex cells fates. In all these cases, these cells are influenced by the levels of cell surface CD25 and thus, by the ability of T cells to capture typically limiting amounts of IL-2 through the high affinity IL-2R. Despite our current knowledge, we still have a cursory understanding of the molecular basis by which IL-2R signaling controls the above cell populations. This is especially true for thymic development of Treg cells, T_R17 cells, and reprogramming of Tex cells. Little attention has been placed on how IL-2 impacts CD4⁺ T_M^202–204. Moreover, IL-2R signaling is more complex than the current paradigm (Fig. 1B), as emphasized by the plethora of IL-2-induced phosphorylated proteins in CD8⁺ T cells¹¹. The ability to perform high resolution multiplex imaging, as done for Treg cells in lymph nodes⁸⁷, offers the opportunity to examine the physiological and therapeutic contribution of IL-2R signaling on the immune landscape in lymphoid and non-lymphoid tissues.

The current state of the field makes it plain that IL-2 as a therapeutic should target the high affinity IL-2R. This is relatively straightforward for Treg cells, where a significant pharmacological window for selectivity over Teff cells exists. However, LD rIL-2 therapy likely requires more persistent signaling that may be delivered by newer IL-2 analogs. The effectiveness of this therapy may also be enhanced when used in combination with other immune modulators that lower inflammation and/or impair activation of autoreactive T cells. Although LD rIL-2 avoids Teff cells, activation of ILC2s and CD56^hi NK cells often occurs. Whether these cells appreciably hinder or contribute to any therapeutic benefit needs careful consideration.

Targeting the high affinity IL-2R may also be critical when using IL-2 in cancer immunotherapy. IL-2 agonists, designed to selectively target the high amounts of the intermediate affinity IL-2R on memory-phenotypic T cells, are essentially a surrogate for IL-15. This type of IL-2 is less effective in generating antitumor responses than IL-2 agonists that target the high affinity IL-2R on tumor-specific CD8⁺ Teff cells^{153,154,156,205,206}.Targeting the high affinity IL-2R is also effective in reinvigorating CD8⁺ Tex cells in the context of checkpoint blockade to drive antiviral and antitumor responses in preclinical studies^153,154,156. In these type of studies, substantial increases in antigen-reactive T cells with high functional activity occurred in the tumor microenvironment such that the Teff cells outnumbered Treg cells, even though Treg cells are expected to respond to the higher doses of IL-2. Why Treg cells do not interfere with this productive immune response requires further investigation. Moreover, these promising preclinical findings warrant testing in cancer patients. Our aim in this review was to highlight our understanding of how IL-2 critically impacts Treg, Teff, and T_M cells and provide a forward-looking point regarding IL-2 as a therapeutic.

Highlights:

• IL-2 shapes the development and function of regulatory, effector and memory T cells.

• The amount of CD25 determines which cells access and respond to IL-2.

• IL-2 reprograms exhausted CD8⁺ T cells to rescue effector function.

• Targeting the high affinity IL-2R is the key to therapy for autoimmunity and cancer.