Th cell differentiation is regulated by the coordinated functions of distinct cytokines and transcription factors. A recent study has demonstrated that increased IRF4 promotes the differentiation of CD4+ CD25^low Teff cells, including Th1, Th2 and Th17 cells, at the expense of Tfh cells (26). In fact, the development and differentiation of Tfh cells only needs an appropriate amount IRF4 in addition to specific TFs, including Bcl-6 and signal transducer and activator of transcription 3 (STAT3) (26, 39). B-lymphocyte-induced maturation protein 1 (Blimp1) is a critical antagonist for Tfh cell differentiation, but it is an important TF for other Th cells, including Th1, Th2, Th17 and Treg cells (40). It has been found that IRF4 cooperates with STAT3 to activate Blimp1 (41), and lack of IRF4 in CD4+ T cells reduces binding to STAT3, resulting in Tfh deficiency (41, 42).

Increasing studies have shown that IRF4 regulates Th17 cell development (43–45). IRF4 knockout decreases the expression of RORγt, a specific TF in Th17 cells (45, 46), which leads to a decrease in Th17 counts, in line with a reduction in serum IL17 and IL21 (47). Likewise, IRF4 deficiency also results in the impairment of Th2 cell differentiation and function by reducing GATA3 and IL4, as well as growth factor independence 1 (Gfi1), a transcriptional repressor required by Th2 cells (48, 49), instead, can promote the T-bet expression and skew toward Th1 cells (48), suggesting that IRF4 plays a pivotal role in the development of Th2 cells rather than Th1 cells. Additionally, IRF4 deficiency inevitably impairs the development of Th2 cells (49). Collectively, IRF4 regulates the differentiation and function of diverse Th subsets that mainly depend on its expression level as well as lineage-specific TFs (26).

Treg cells are indispensable for maintaining immune tolerance (37, 50); nevertheless, they also impair anti-tumor capability and promote tumor growth, particularly tumor-infiltrating Treg cells (51). Foxp3 is a lineage-defining TF for Tregs and the key regulator of its development and function (52, 53). IRF4, which acts downstream of Foxp3, can physically and functionally interact with Foxp3 and cooperate with BATF3 to regulate Foxp3 expression (54, 55), which instructs effector Treg cell differentiation and immune suppression (56). Moreover, Blimp1 is a target of Foxp3 in Treg cells, and it is directly induced by IRF4 (57, 58). Accordingly, lack of IRF4 in Treg cells suppresses Blimp1 expression, and more intriguingly, leads to decreases in multiple Treg-related molecules, such as inducible T cell costimulatory (ICOS), IL10 and IL1 receptor 11 (IL1RL1), confirming that IRF4 cooperates with Blimp1 to regulate the differentiation and function of Treg cells (56, 58).

Additionally, compared with IRF4-deficient Treg cells, IRF4+ Treg cells overexpress BATF, IKAROS family zinc finger 2 (IKZF2), Ki67, ICOS and inhibitory molecules, such as programmed cell death protein 1 (PD1) and T cell immunoreceptor with immunoglobulin and ITIM domain (TIGIT) (38), exhibiting a highly activated phenotype and strong inhibitory effects in several tumors (59–61). In particular, an increase in intratumoral IRF4+ Treg cells with superior suppressive activity was significantly correlated with early tumor recurrence and poor disease-free survival (DFS) and overall survival (OS) (38). Accordingly, inhibition of IRF4 severely impaired the development and function of Treg cells at the tumor-infiltrating sites and significantly repressed tumor growth in a mouse model (38, 51). Collectively, growing evidence implicates IRF4 plays a central role in the differentiation and immunosuppressive activity of Treg cells in the TME, and IRF4+ Treg cells definitely inhibit anti-tumor immunity. Therefore, specifically targeting IRF4 in Treg cells may reverse the tumor microenvironment from immunosuppression to immune activation against tumor cells, which may become an effective anti-tumor therapeutic strategy.

Following antigen stimulation, naïve CD8+ T cells are differentiated into a large number of antigen-specific short-lived effector cells (SLECs) (62, 63), exerting cytotoxic activity ( Figure 3A ). Mechanically, antigen stimulation drives the expression of TCR responsive factor IRF4 (68). Next, IRF4 combined with AP1 family TFs form an activating IRF4/AP1 complex, which integrates TCR and costimulatory signals to induce the production of a series of effector cytokines. After antigen clearance, the expression of IRF4 decreased, followed by an increase in expression of stemness-like gene T cell factor 7 (Tcf7; encoding TCF1) ( Figure 3B ), and further producing memory precursor cells (MPECs) and TCF1+ memory-like T cells to rapidly function in the secondary response (64, 65) ( Figure 3A ).

The intensity of TCR signaling regulates the expression of IRF4 (66, 70). High levels of IRF4 in CD8+ T cells contribute to the clonal expansion of SLECs, which are critical for maintaining effective anti-tumor immunity (71) and acute pathogen control (64). Interestingly, ectopic expression of IRF4 remarkably enhances the clonal expansion and effector cytokine production of T cells induced by low-intensity TCR signaling (69). Conversely, selective knockout of IRF4 in peripheral CD8+ T cells leads to progressive loss of the effector function of CD8+ T cells (72–74). The RNA-binding protein Roquin1, a key target upstream of IRF4, can effectively inhibit the expansion of CD8+ T cells (75). Accordingly, lack of Roquin1 can significantly promote the proliferation of CD8+ T cells by upregulating IRF4 (71). However, if IRF4 is also deficient, the expansion-promoting effects caused by Roquin1 deficiency is completely abolished (71). Therefore, the Roquin-IRF4 axis may also serve as a potential target for enhancing anti-tumor immunity.

IRF4 also converts TCR affinity into appropriate transcriptional programs, linking metabolic function to T cell clone expansion and effector differentiation (76) by regulating the expression of key molecules required for aerobic glycolysis on effector T cells, including hypoxia inducible factor1 α (HIF1α) and forkhead box protein o1 (Foxo1) (77). Compared with weak or low-affinity TCR stimulation, strong or high-affinity TCR stimulation contributes to increased glucose uptake in an IRF4-dependent manner (78). Taken together, IRF4 regulates the expansion and differentiation of effector CD8+ T cells by translating the TCR signal and converting it to metabolic function.

Not surprisingly, similar to initial antigen stimulation, IRF4 overexpression significantly induces an increase in the cytotoxicity of memory CD8+ T cells (32, 68, 79). By contrast, IRF4 deficiency may cause memory CD8+ T cells to produce but not proliferate (68), which results in impairment of the effector function (32, 72, 79). So far, at least three types of memory CD8+ T cells have been defined: central memory T (T_CM) cells, effector memory T (T_EM) cells and tissue-resident memory T (T_RM) cells (80). Compared with T_EM cells, T_RM cells express higher levels of IRF4, and their formation and maintenance are IRF4 dependent (32). IRF4 deletion leads to an increase in T_EM cells and a decrease in T_RM cells, but it does not affect the total number of memory T cells (32). Thus, targeting IRF4 may strongly reduce the number of T_RM cells, thus substantially weakening transplant rejection (81).

In addition, recent studies have shown that TCF1 is essential for maintaining CD8+ T_CM cells and serves as a positive biomarker for prolonged survival and effective responses to PD1 inhibitors in various solid tumors and hematological malignancies (82–85). Undoubtedly, high-level IRF4 is beneficial to the initial effector function, but sustained overexpression of IRF4 inhibits the expression of TCF1, which further damages the production of antigen-specific T_CM cells and is not conductive to the rapid effect function in recall responses (24). Collectively, accumulating studies have demonstrated that IRF4 is indispensable for robust proliferation and the effector function of memory T cells in recall responses.

High IRF4 is essential for maintaining the differentiation and expansion of effector CD8+ T cells (68, 72). However, too much is as bad as too little. Persistent antigen stimulation due to tumor or chronic viral infection can cause constitutively high expression of IRF4, which in turn induces CD8+ T cell exhaustion (24). There are several characteristics of exhausted CD8+ T cells ( Figures 3A , 4 ): (1) up-regulation of multiple inhibitory receptors (86), (2) progressive loss of effector function and impaired differentiation of potential memory T cells (85, 87), (3) decreased production of cytokines involved in chemotaxis, adhesion and migration, and (4) metabolic deficiency (88). Thus, functional exhaustion is probably due to both active suppression and passive defects in signaling and metabolism.

Studies have demonstrated that the epigenetic and transcriptional programs driving CD8+ T cell exhaustion are triggered by sustained antigen-dependent activation of TCR signaling, leading to two events: (1) the sustained overexpression of TCR-responsive IRF4 and its binding partners, mainly AP1 family members, including BATF, BATF3, JunB and JunD (24, 35, 89–92), as well as nuclear factor of activated T cells (NFAT), a key regulator of T cell activation (93), followed by (2) sustained expression of multiple exhaustion-related molecules (24). Specifically, overexpressed IRF4 binding with AP1 family members or NFAT leads to an overabundance of IRF4/AP1 complexes or NFAT homodimers that are recruited to specific DNA sites to open multiple exhaustion-related chromatin regions, including inhibitory receptors, such as PD1, T-cell immunoglobulin and mucin domain 3 (TIM3) and cytotoxic T lymphocyte antigen 4 (CTLA4) (24, 35, 94, 95), as well as orphan nuclear receptor 4A (NR4A) and thymocyte selection-associated high mobility group box (TOX) family members, which act to impose exhaustion (96, 97), further inhibiting TCF1 expression ( Figure 3C ) (24, 35). These events eventually drive CD8+ T cell exhaustion and limit the development of TCF1+ memory-like T cells and anti-tumor activity ( Figure 3A ). This chromatin binding imbalance due to the accumulation of IRF4/AP1 TF complexes was also found in CAR-T cell therapy (89).

Fortunately, Lynn et al. (89) found that ectopic overexpression of c-Jun in exhausted CAR-T cells can effectively rescue exhaustion and restore anti-tumor activity by disrupting and/or displacing immunosuppressive transcriptional complexes containing IRF4 and AP1 family members (89). Moreover, based on the overexpression of BATF and IRF4 in exhausted T cells (89, 98), knockdown of BATF or IRF4 could remarkably enhance the tumor-killing ability of CAR-T cells by reversing their exhaustion and prolonging their persistence (89, 90). Likewise, Seo et al. (25) found that overexpressed BATF in BATF-transduced CAR-T cells could cooperate with appropriate amount of IRF4 to counteract exhaustion, promoting the expansion of CD8+ CAR-T cells and increasing their effector cytokine production. Nevertheless, inhibiting the interaction between BATF and IRF4 will greatly weaken the tumor control ability of BATF-overexpressing CAR-T cells (25).

Collectively, these findings show that persistent overexpression of IRF4 drives T cell exhaustion depending on the specific microenvironment and the amount and functional status of its binding partners. Therefore, manipulating the formation of IRF4/AP1 complexes may be an inspiring therapeutic strategy to overcome T cell exhaustion. Yet, the core transcriptional network of IRF4 involved in these two opposing programs still needs to be further elucidated.

There are two types of macrophages: M1 (anti-tumor activity) (106, 107) and M2 (pro-tumor activity) (108). Generally, TAMs mainly refer to M2 macrophages, which are characterized by high expression of arginase1 (Arg1), chitinase-like 3 (Ym1/Chil3), found in inflammatory zone 1 (Fizz1) and mannose receptor (MR) (109, 110). IRFs play a key role in macrophage maturation and phenotypic polarization. Of the nine IRFs, IRF1, IRF5 and IRF8 are involved in the commitment of M1 macrophages, whereas IRF3 and IRF4 are crucial for M2 macrophage polarization through regulating the expression of Arg1 and Ym1, which further sufficiently produces Th2 and directly suppresses effector T cell proliferation (111–113).

In addition, it has been reported that Jumonji domain containing 3 (Jmjd3) is essential for M2 macrophage polarization, and IRF4 is a Jmjd3 target gene (110, 114). Phosphatidylserine released by apoptotic tumor cells could induce the polarization and accumulation of M2 macrophages via a STAT3-Jmjd3-IRF4 signaling axis (115); therefore, down-regulation of Jmjd3 by targeting the STAT3-Jmjd3-IRF4 axis may be a candidate approach for inhibiting the accumulation of M2 macrophages in tumor sites and remodeling the TME. Moreover, some miRNAs have been found to promote the transformation of macrophages from M2 to M1 by targeting IRF4 to activate IRF5 (116, 117). Given that IRF4 promotes the polarization of M2 macrophages, targeting IRF4 to reprogram TAM polarization in the TME appears to be a promising therapy for tumors.

DCs, known as professional antigen presenting cells, play a major role in orchestrating immune responses, and can be mainly divided into three subtypes: plasmacytoid DCs (pDCs), classical DCs (cDCs, including cDC1 and cDC2), and monocyte-derived DCs (mo-DCs) (9, 118, 119). However, the differentiation and maturation of DCs are often impaired by the immunosuppressive TME, which leads to DC dysfunction and induces tolerance to tumor cells (8, 9, 118, 119). For instance, mature pDCs exert immunostimulatory function, which is characterized by the production of large amounts of type I IFNs. Whereas, in the TME, pDCs with reduced production of type I IFNs favor the development of Treg cell, exert immunosuppressive effects on CTLs and promote tumor progression (8, 9, 120, 121). Several studies have indicated a role for IRF4 in development of monocytes, pDCs, and cDCs (122–124). IRF4 contributes to the differentiation of pDCs (122). In addition, IRF4 plays a key role in the development of cDC2 and promotes their survival and migration to lymph nodes and is essential TF for cDC2-mediated Th2 induction (122). By contrast, inhibition of IRF4 in DCs represses Th2 and promotes Th17 responses (123).

The monocytes in the TME can prioritize differentiation into monocyte-derived macrophages (mo-Macs) rather than mo-DCs (3, 105). The presence of mo-DCs has been correlated with CD8+ T cell activation and successful anti-tumor therapy (125). IRF4 is essential for human mo-DC differentiation and efficient antigen cross-presentation, whereas IRF4-deficent monocytes are phone to differentiation into mo-Macs (124). Devalaraja et al. found that the TME induces tumor cells to produce retinoic acid (RA) in murine sarcoma models, which drives intratumor monocyte polarization to mo-Macs instead of mo-DCs by inhibiting IRF4 (3). Interestingly, overexpression of IRF4 in human monocytes can sufficiently block RA-mediated mo-Mac differentiation (3, 124). Collectively, these results suggest that IRF4 plays critical and complicated roles in the maturation and differentiation of DCs in the TME.

MDSCs are immature myeloid cells that do not differentiate into mature myeloid cells, and this is a major obstacle to achieving successful immunotherapy in tumors (126, 127). Two major subpopulations, monocytic (M) MDSCs and polymorphonuclear (PMN)-MDSCs, have an immune suppressive function. IRF4 plays a role in the lymphoid cell development. However, IRF4 expression is decreased in immature myeloid cells, such as MDSCs in tumor-bearing mice and chronic myeloid leukemia cells (104, 128). Accordingly, IRF4 deficiency further favors the generation of MDSCs in the TME, and increases the expansion of M-MDSCs and the infiltration of PMN-MDSCs with a strong suppressive capacity, which inhibits the proliferation of CD8+ T cells through IL10 and ROS generation and promotes tumor growth (104, 129). By contrast, an increase in the IRF4 expression in MDSCs from bone marrow cells inhibits the numbers of MDSCs through induction differentiation, and further damages the immunosuppressive function of MDSCs (104). Unfortunately, IRF4 expression is remarkably suppressed during the development of MDSCs and tumor formation in the TME (104).

Altogether, these data show that IRF4 plays a critical role in preventing the generation of MDSCs; nevertheless, IRF4 expression is limited by tumors and MDSCs, which may in turn boost the accumulation and suppressive activity of MDSCs to accelerate the generation of an immunosuppressive TME. Thus far, the exact mechanisms regulating IRF4 in the differentiation of MDSCs remains largely unknown.