T-bet is a key transcription factor for natural killer cell maturation (12). T-bet expression correlates with NK cell maturation during the process from less mature CD11b^posCD27^pos stage to the more mature CD11b^posCD27^neg stage. Deficiency of T-bet leads to significantly reduced NK cell numbers, as well as the lack of TRAIL⁺DX5^- NK cells. In addition, Eomes–deficient NK precursors could not develop into NK cells if T-bet is also deficient, as NK cells could not be detected in any organs in Tbx21^−/−Eomes^Flox/Flox; Vav-Cre⁺ mice (13). Therefore, T-bet coordinates with Eomes for the development of NK cells, and T-bet is required for NK cell maturation (14–16). T-bet is also important for NK cell effector functions. Although T-bet^-/- NK cells can also rapidly secrete IFN-γ compared with wild-type NK cells, the maintenance of IFN-γ production is impaired in the absence of T-bet. Besides, the cytolytic activity of T-bet^−/− NK cells against target cells was significantly reduced (14). The above aspects will be discussed in detail in the later sections of this review.

T cell differentiation is essential for the outcome of immune responses in diseases (17). T-bet plays an important role in the regulation of differentiation of T cell subsets, which is orchestrated by the activity of a series of transcription factors (18). T-bet promotes Th1 differentiation, but inhibits Th2, Th17 and Tfh cell lineage commitment (6, 19). T-bet regulates Treg cell homeostasis and promoted T-bet⁺ Treg cells to accumulate at sites of Th1-mediated inflammation by inducing expression of the chemokine receptor CXCR3 on Treg cells (20). CD8⁺T cells that have high T-bet levels tend to become terminally differentiated effector cells, whereas the cells that have low T-bet levels have a higher memory cell developmental potential (21).

In addition to NK cells and T cells, T-bet also regulates other lymphocyte populations. T-bet acts as a selective inducer for the expression of Iγ2a transcription, and promotes IFN-γ-mediated IgG2a class-switching in B cells (22). IFN-γ/T-bet–dependent pathway regulates CXCR3 expression in B cells to drive the migration of memory B cells to inflammatory foci (23). On the other hand, T-bet is expressed in DCs at levels comparable with Th1 cells and is necessary for the production of IFN-γ and activation of antigen-specific Th1 cells (24). Moreover, T-bet suppresses intestine IL-17A –producing ILCs, which are potent promoters of ulcerative colitis, by decreasing the expression of IL-7 receptor in these cells (25). IL-2 and IL-15 are structurally related cytokines that share common receptor subunit-CD122 (IL-2 receptor β chain) (26). T-bet is required for optimal CD122 expression on thymic intraepithelial lymphocyte precursors (12), as well as on Vα14i NKT cells (14) for their development or maturation.

IFN-γ plays a critical role in regulation of Th1 cell development (27). After 6 or 24h of anti-CD3 stimulation, IFN-γ -producing Th1 cells, but not Th2 cells, increased T-bet expression. After infection of a prototypical Th1-inducing pathogen, Toxoplasma gondii, splenic Tbx21 mRNA levels increased 4-fold in wild type mice, but not in Ifnγ ^−/−mice, indicating that IFN-γ is essential in inducing T-bet expression (28). Naive CD4⁺ T cells from lymph nodes of 5C.C7 T cell receptor transgenic/Rag2-deficient mice incubated with APCs increased expression of T-bet in T cells in the presence, but not in the absence, of TCR –specific peptide and IFN-γ, indicating that T-bet could be induced in response to TCR stimulation in an IFN-γ–dependent manner (28). Furthermore, IFN-γ stimulation failed to induce T-bet in Stat1-deficient mice (29). The above evidence suggests that T-bet is induced in T cells by TCR through the IFN-γR/STAT1 signaling pathway.

Moreover, three potential T-box binding sites were detected in Ifnγ gene, two of which were located in the proximal promoter region and one in the third intron (5), showing that T-bet is a transactivator of the Ifnγ gene and mediates positive feed-forward regulation of endogenous IFN-γ production.

Besides IFN-γ, IL-12 is also involved in the differentiation of naive T lymphocytes into Th1 cells (30, 31), and could also induce the expression of T-bet. Although IL-12 could be an inducer of IFN-γ, Ifngr ^−/− T cells increased T-bet expression after stimulation with IL-12 (32), indicating that IL-12 induces T-bet expression independent of IFN-γ. In support of this, chromatin immunoprecipitation showed that the conserved T-bet enhancer element 13 kilobases upstream of the transcriptional start site can interact with IL-12–activatedSTAT4 to induce T-bet expression (33). Purified splenic NK cells from wild-type, Ifnγ^−/− , and Stat4^−/− mice stimulated with IL-12/IL-18 for 12h showed that IL-12 induced the expression of T-bet through STAT4 independently of IFN-γ (14). Therefore, early after TCR stimulation, T-bet expression is mainly induced through the TCR/IFN-γ signaling pathway, when the expression of IL-12 receptor β2 subunit is repressed by TCR signaling, whereas in the later phase, this repression is removed upon termination of TCR stimulation, allowing T-bet induction through STAT4 downstream of IL-12R signaling pathway ( Figure 1 ).

MyD88 is a TIR-domain-containing adaptor protein in response to Toll-like receptors in TLR signaling pathway (34). MyD88 is located downstream of the TLR (except TLR3) and IL-1 receptor family, and transmits signals to activate NF-κB and MAP kinases and to induce inflammatory cytokines (35). 3 weeks after naïve CD4⁺ T cells from WT donors were transferred into congenic WT or MyD88 deficient recipient mice, the differentiation of T-bet^high memory-phenotype (MP) cells was significantly fewer in MyD88 deficient recipient mice than in WT mice, which indicates that extrinsic TLR-MyD88 signaling plays a key role in driving MP cells to differentiate into T-bet^high subpopulations (36). B cell TLR signaling has been reported to induce T-bet expression to produce IgG2A/cAb in response to viral infection (37, 38). Immunization with Qβ-VLP of irradiated chimeric mice of mixed bone marrow cells from WT (CD45.1⁺) and MyD88 KO (CD45.2⁺) showed that T-bet expression in MyD88 KO derived Qβ⁺ B cells from chimeric mice was much lower than that in WT-derived cells, which suggests that B cell-intrinsic MyD88 signaling was required for T-bet expression (39). On the other hand, IL-21 can induce STAT3 to bind to MyD88 activated sequence (GAS) and SIE (a well-characterized element binding activated STAT3), leading to up-regulation of T-bet expression in NK cells (40). This indicates that IL-21 may regulate T-bet expression in NK cells through the JAK-STAT3-MyD88 axis.

mTOR binds to Raptor and Rictor to form two complexes, mTORC1 and mTORC2, respectively (41). The activity of mTOR could be stimulated via IL-15 signaling in NK cells (42), which plays an important role in NK cell effector functions (43). The absence of either mTORC1 or mTORC2 led to decreased expression of T-bet by NK cells (44, 45). Tsc1 (a repressor of mTOR) negatively regulated IL-15-triggered mTORC1 activation in NK cells. TSC1 deficiency resulted in higher activity of mTORC1 in Tsc1 ^−/− NK cells than in WT NK cells. Meanwhile, T-bet was expressed at a higher level in Tsc1 ^−/− NK cells compared with WT NK cells (46). Six phosphorylation sites of T-bet were identified to be targets of mTORC1, among which at least three sites were required for recruitment of chromatin remodeling complexes to the Ifng gene promoter by T-bet for a normal level of IFN-γ production (47). These studies indicate that T-bet expression is promoted by mTOR signaling pathway ( Figure 2 ).

The applications of T-bet reporter mice have facilitated the determination of T-bet expression level (6), which, together with intracellular staining for flow cytometry, have enabled the detection of T-bet expression level even in small lymphocyte populations that are difficult for western blot or Q-PCR. However, the activity of T-bet is regulated by posttranslational protein modifications (48), which could have made it more complex to assess T-bet activity. For example, IL-2 –inducible tyrosine kinase (ITK) phosphorylates the tyrosine residue Y525 of T-bet in Th cells after stimulation with TCR and IL-12, which is important for interaction with GATA-3 and for suppression of Th2 cell differentiation (49). In addition to phosphorylation by ITK, tyrosine residues Y219, Y265 and Y304 of T-bet are also phosphorylated by c-Abl, which is essential for IFN-γ production and suppression of Th2 cytokine production, and for control of allergic lung inflammation (50). Moreover, serine residue S508 phosphorylation is important for T-bet to interact with and prevent NF-κB p65 from binding to the Il2 promoter and for T-bet mediated suppression of IL-2 production in T cells (51). While the phosphorylation of tyrosine or serine residues in T-bet regulates the activity of T-bet, ubiquitination of lysine residues in T-bet affects both the activity and half-life of T-bet. Among 16 lysine residues of T-bet, although ubiquitination of the lysine residue K313 leads to proteasomal degradation of T-bet, such modification is important for T-bet’s DNA binding activity and T-bet –mediated suppression of IL-2 and Th2 cytokines (52). Future studies are required to determine the impact of posttranslational protein modifications of T-bet on NK cell biology.

T-bet was proposed to maintain the stability of immature NK cells, since TRAIL⁺DX5^- NK cells were dramatically reduced in Tbx21−/− mice. Deletion of T-bet by treating Tbx21 ^flox/flox NK cells with TAT-Cre ex vivo resulted in the reduction of this population after transferred into Il2rg^−/−Rag2^−/− mice, confirming that T-bet stabilizes this NK cell subset (13). However, lack of T-bet was later shown to significantly reduce TRAIL⁺CD49a⁺CD49b^- tissue-resident NK cells (trNK cells), which, though showing an immature phenotype, represents a lineage distinct from circulating conventional CD49a^-CD49b⁺TRAIL^- NK cells (53, 54). Therefore, the developmental stability of trNK cells is dependent on T-bet.

scRNA-seq revealed that more than 65% of T-bet-deficient NK cells were classified as the least mature iNK cluster, and the expression of immature NK signature genes was highly upregulated in T-bet-deficient NK cells (55). This shows that in the absence of T-bet the overall NK cell population displayed a more immature phenotype, indicating that T-bet is required for NK cell maturation. Mechanistically, T-bet induces and cooperates with Zeb2 to promote the transcription program for NK cell maturation, and induces S1P5 for mature NK cell egress from the bone marrow.

T-bet plays an important role in immune response against acute infections. In the process of infection with the intracellular bacterial pathogen Listeria monocytogenes (LM), T-bet enhances its own expression and IFN-γ production through the STAT1 signaling pathway, while the absence of T-bet may impair sustained secretion of IFN-γ by NK cells and T cells (65–67). In acute infection with LCMV Armstrong, loss of T-bet expression in Tfh cells suppressed Tfh expansion, and reduced the levels of germinal center B cells and plasma cells, as well as lower LCMV-specific IgG and subtype IgG2c titers in serum, indicating that T-bet is essential for Tfh response and antibody IgG2 class switching in acute LCMV infection (68). In addition, T-bet^-/-mice displayed a compromised resistance to Mycobacterium tuberculosis infection, mainly due to impaired IFN-γ production from the CD4⁺Th1 T cell subset to form an effective type 1 immune response (69, 70). During the infection of intracellular parasite Toxoplasma gondii, dendritic cells and macrophages produce IL-12 to promote the activation and expansion of NK cells and T cells populations expressing high levels of T-bet and IFN-γ (71–75).

T-bet is critical for the development of immunopathology in the experimental autoimmune encephalomyelitis (EAE) model of inflammatory demyelination (76, 77). Importantly, T-bet–dependent NKp46⁺ ILCs are required for the initiation of CD4⁺ Th17 –mediated neuroinflammation (78). In Crohn’s disease, a chronic inflammatory disease of the gastrointestinal tract, T-bet protein expressed higher in lamina propria CD4⁺ T cells from patients than from healthy controls. The high expression of T-bet accelerates the development of Th1-mediated colitis, whereas mice lacking T-bet are more likely to develop Th2-mediated colitis (79, 80). In heart failure(HF), TCR –activated PKC-θ induces T-bet, which ultimately leads to the stimulation of immune response and the occurrence of cardiac hypertrophy and fibrosis (81). However, T-bet could act as a double-edged sword in response to inflammatory and autoimmune diseases as well. In cardiovascular disease (CVD), on the one hand, T-bet stimulates and activates the functions of immune cells such as NK and DC to increase inflammation and cardiac injury. On the other hand, T-bet activates regulatory T cells and stimulates angiogenesis to inhibit inflammation and replaces damaged vessels with new vessels (82). Besides, T-bet in B cells can regulate immunoglobulin (IgG) class switching. Increased T-bet expression in B cells inhibits IgE production and alleviates IgE-mediated allergic reactions and asthma (22, 83).

T-bet-deficient mice generated in a transgenic adenocarcinoma mouse prostate model (TRAMP) indicated that lack of T-bet has little effect on the development of primary tumors, but that T-bet is needed to suppress metastasis (84). In addition, T-bet-deficient mice were extremely sensitive to the metastasis and spread of B16F10 melanoma cells (85). On the other hand, T-bet was down-regulated in adoptively transferred NK cells upon exposure to tumors and proliferation, which was associated with down-regulation of activating receptors and IFN-γ (86). T-bet mRNA level expressed by the exhausted NK cells of HBV individuals is lower than that of the NK cells of healthy donors (87). T-bet expression in virus-specific CD8⁺ T cells is also reduced during chronic infections, which negatively correlated with the expression of exhaustion marker PD-1 (88). Therefore, T-bet is critical for the control of tumors and chronic infections, but its expression might be downregulated in effector cells in these contexts.

In NK-based tumor immunotherapy, the following strategies are currently under investigations to promote the anti-tumor effects of NK cells: 1.Blockade of the cell surface or intracellular inhibitory checkpoint molecules of NK cells (89–92); 2. Arming with chimeric antigen receptor (CAR) to improve tumor recognition specificity of NK cells (93, 94); 3. Bispecific antibody to connect NK cell activating receptor to target cells (95). Based on these technologies, as well as the knowledge on T-bet regulation and functions in NK cells, we can design strategies to promote the anti-tumor efficacy of NK –based immunotherapy.

Previous studies have shown that adoptive NK cell immunotherapy failed to effectively control tumors, which was associated with down-regulation of T-bet. Accordingly, we can further dissect the factors (inhibitory receptor-ligand interactions/soluble factors) responsible for the down-regulation of T-bet in tumor-associated NK cells, and design strategies to recover T-bet expression for maintaining the anti-tumor functions in NK cells. In line with this, small molecules potentially enhancing T-bet expression have been reported (96). Alternatively, ectopic expression of specific transcription factors in immune cells for tumor immunotherapy has also been reported (97, 98), T-bet –overexpressing CAR-T cells exhibited a Th1 phenotype with more effective anti-tumor activity, leading to improved survival of tumor-bearing mice (99). These studies suggest that the ectopic expression of T-bet in NK cells could be tested for tumor immunotherapy in future. In addition, based on knowledge on the cytokine receptors and signaling pathways responsible for inducing T-bet, we can rationally design the signaling domain of CAR-NK to trigger the T-bet –inducing pathways to promote the expression of T-bet for enhancing NK cell effector functions. Alternatively, we can design multi-specific antibody to trigger T-bet –inducing cytokine receptors to promote the expression of T-bet.

T-bet and Eomes are highly homologous transcription factors, which play redundant or non-redundant roles as key checkpoints in the regulation of NK maturation (13). The synergistic cooperation between the two can better exert the effector function of NK cells for the control of tumors and infections. The balance between T-bet and Eomes in NK cell maturation, differentiation and effector function still requires further in-depth studies for the spatiotemporal effects of these transcription factors, the transcription elements that regulate the transcription of T-bet and Eomes, and the molecular chaperones that help them fold and assemble at the protein level. On the other hand, the regulation of T-bet expression in NK cells is closely related to the pathogenesis of various diseases. Further studies on the expression of T-bet in NK cells in disease models will better demonstrate the role of T-bet in NK cell biology. T-bet downregulation was shown to be one of the hallmarks of NK cell exhaustion in tumors, whose details still lack further elaboration, and might benefit future design of NK-based immunotherapy. Moreover, the human TBX21 gene has 40 known polymorphisms, of which one has been shown to be related to the susceptibility to type 1 diabetes (100), and the other is related to the incidence of herpes simplex virus 2 (101). However, the effect of T-bet polymorphisms on NK–dependent immune responses in human diseases is largely unknown. Therefore, from the angle of T-bet, our knowledge on NK cell biology is still limited. Further research on T-bet is essential, not only to expand our understanding of NK cells, but also to assist the design of future strategies of immunotherapy.