Cell intrinsic role of NF-κB-inducing kinase in regulating T cell-mediated immune and autoimmune responses

NF-κB inducing kinase (NIK) is a central component of the noncanonical NF-κB signaling pathway. Although NIK has been extensively studied for its function in the regulation of lymphoid organ development and B-cell maturation, the role of NIK in regulating T cell functions remains unclear and controversial. Using T cell-conditional NIK knockout mice, we here demonstrate that although NIK is dispensable for thymocyte development, it has a cell-intrinsic role in regulating the homeostasis and function of peripheral T cells. T cell-specific NIK ablation reduced the frequency of effector/memory-like T cells and impaired T cell responses to bacterial infection. The T cell-conditional NIK knockout mice were also defective in generation of inflammatory T cells and refractory to the induction of a T cell-dependent autoimmune disease, experimental autoimmune encephalomyelitis. Our data suggest a crucial role for NIK in mediating the generation of effector T cells and their recall responses to antigens. Together, these findings establish NIK as a cell-intrinsic mediator of T cell functions in both immune and autoimmune responses.

Given the complexity of NIK function in the development and differentiation of lymphoid organs and immune cells, the study of cell-intrinsic functions of NIK requires NIK conditional KO mice. In particular, the role of NIK in regulating T cell function has been controversial. While some studies suggest a role for NIK in regulating T cell-mediated immunity and autoimmunity, other studies suggest the indirect effect from accessary cells, such as DCs 10-12 . In the current study, we employed conditional KO mice lacking NIK specifically in T cells. We show that NIK has a cell-intrinsic role in regulating the homeostasis and function of T cells. NIK is required for in vivo differentiation of inflammatory T cells and the induction of a T cell-dependent autoimmune disease, experimental autoimmune encephalomyelitis (EAE).

Results
T cell-specific NIK ablation does not affect thymocyte development. Canonical NF-κ B plays an important role in regulating development of both conventional T cells and Treg cells 1 . Although global NIK-KO mice have abnormal T-cell selection, it is likely that the impaired development of thymic epithelial cells may contribute to this phenotype. To examine the cell-intrinsic function of NIK in regulating thymocyte development and peripheral T-cell function, we generated NIK T cell-conditional KO (NIK-TKO) mice by crossing NIK-flox mice with CD4-Cre mice (Fig. 1a). Since NIK level is extremely low due to its continuous degradation 13 , we treated the cells with a proteasome inhibitor, MG132, for NIK detection by immunoblot. NIK was detected in wildtype T cells, but not in NIK-TKO T cells, suggesting efficient ablation of NIK by the CD4-Cre (Fig. 1b). Consistently, the NIK-TKO T cells had a defect in anti-CD3/anti-CD28-stimulated p100 processing (generation of p52) (Fig. 1c). (e,f) Flow cytometry analyses of thymocyte and T cell populations in the thymus and spleen of age-matched (6-8 weeks old) WT and NIK-TKO mice based on CD4 and CD8 staining. Data are presented as a representative plot (e) and mean ± SD values of multiple mice (f, each circle represents a mouse). (g,h) Flow cytometry analysis of the surface expression of CD69 and TCRβ on thymocytes from 5 weeks old WT and NIK-TKO mice, showing four subpopulations (1, TCR lo CD69 lo ; 2, TCR lo CD69 hi ; 3, TCR hi CD69 hi ; 4, TCR hi CD69 lo ) of thymocytes, presented as a representative plot (g) and and mean ± SD values of multiple mice (h). In contrast to the severe reduction of B cell numbers in the global NIK-KO mice, the NIK-TKO mice had normal frequency of B cells in the spleen (Fig. 1d). The T cell-specific NIK deficiency also did not have an obvious effect on thymocyte development, as evidenced by the comparable frequency of different subpopulations of thymocytes, including CD4 − CD8 − double-negative (DN), CD4 + CD8 + double-positive (DP), CD4 + CD8 − and CD8 + CD4 − single-positive (SP) cells (Fig. 1e). Consistently, the NIK-TKO and wildtype control mice had comparable frequency of peripheral CD4 + and CD8 + T cells in the spleen (Fig. 1e,f). Thymocyte development from DP to SP stages involves positive selection, a process that is associated with upregulation of surface TCRβ and CD69 14,15 . Flow cytometry analysis revealed that the wildtype and NIK-deficient thymocytes contained similar percentages of TCR lo CD69 lo (preselection), TCR lo CD69 hi (transitional), TCR hi CD69 hi (postselection), and TCR hi CD69 lo (mature) populations (Fig. 1g,h). Thus, NIK expression in T-cell compartment is dispensable for thymocyte development.

NIK has a cell-intrinsic role in regulating the homeostasis of peripheral T cells and Treg cells.
The maintenance of naïve and memory T cells in the peripheral lymphoid organs involves homeostatic signaling pathways stimulated by both cytokines and self-peptide-MHC ligands for TCR 16 . To assess the cell-intrinsic function of NIK in regulating T-cell homeostasis, we analyzed the frequency of naïve and memory T cells in the peripheral lymphoid organs. As expected, the splenocytes of wildtype mice contain both naïve and memory T cells, characterized by the CD44 lo CD62L hi and CD44 hi CD62L lo surface markers, respectively. Interestingly, although NIK-TKO and wildtype mice had similar frequency of total CD4 + and CD8 + T cells in the spleen (Fig. 1e,f), the NIK-TKO mice had significantly reduced frequency and number of the memory population and concomitantly increased naïve population of T cells in the spleen (Fig. 2a,b). These results suggest a role for NIK in regulating the homeostasis of memory like T cells.
We next examined whether the NIK deficiency altered the generation or maintenance of regulatory T (Treg) cells. In the thymus, the NIK-TKO mice and wildtype control mice had a similar frequency of the Foxp3 + Treg cells, suggesting a dispensable role for NIK in regulating Treg production (Fig. 2c,d). However, the NIK-TKO mice had a significantly lower frequency of Treg cells in the peripheral lymphoid organs, lymph nodes and spleen (Fig. 2c,d). Thus, NIK expression in T-cell compartment is important for the maintenance of both memory-like T cells and Treg cells in the periphery. T-cell homeostasis in vivo prompted us to examine whether NIK is required for T-cell activation. We purified naïve CD4 + T cells from young adult mice and stimulated them in vitro using monoclonal antibodies for TCR (anti-CD3) and CD28 (anti-CD28). As expected, wildtype T cells produced the T cell growth factor IL-2 in response to in vitro stimulation (Fig. 3a). NIK ablation in T cells did not appreciably affect this important molecular event of T cell activation (Fig. 3a). The NIK deficiency also did not influence the induction of Il2 mRNA, as revealed by a real-time quantitative RT-PCR (qRT-PCR) assay (Fig. 3a). Furthermore, the NIK-deficient and wildtype naïve T cells displayed a similar level of proliferative ability when stimulated in vitro with anti-CD3 and anti-CD28 antibodies (Fig. 3b).
Following activation, naïve CD4 + T cells differentiate into various subsets of effector T cells, including the interferon-gamma (IFN-γ )-producing T helper (Th)1 cells, IL-4-producing Th2 cells, IL-17A-producing Th17 cells, as well as the Foxp3 + Treg cells 17 . The CD4 + T cell differentiation is controlled by both the TCR signal and signals stimulated by various cytokines. We examined the role of NIK in regulating T cell differentiation using an in vitro model system. The T cell-specific ablation of NIK had little or no effect on the induction of Th1, Th2, or Treg cells (Fig. 3c). However, the induction of Th17 cells was partially inhibited in the NIK-TKO T cells (Fig. 3c,d). These results suggest that NIK is dispensable for in vitro activation of T cells but plays a role in regulating Th17 cell differentiation.

NIK regulates antigen-stimulated in vivo T cell differentiation and recall responses.
To examine the role of NIK in regulating antigen-stimulated T-cell responses in vivo, we immunized wildtype and NIK-TKO mice with a protein antigen, Keyhole limpet hemocyanin (KLH). After a week of immunization, we isolated draining lymph node cells and splenocytes for in vitro restimulation of the antigen-specific effector T cells with KLH. Both the draining lymph node and the spleen of immunized mice contained antigen-specific Th1 and Th17 effector T cells, which produced IFN-γ and IL-17A upon in vitro restimulation with KLH (Fig. 4a). Importantly, compared to the wildtype mice, the NIK-TKO mice had a significantly lower level of Th1 and Th17 responses, particularly in the draining lymph nodes where the majority of effector T cells reside (Fig. 4a). It appeared that NIK was crucial for antigen-stimulated recall responses of the effector T cells, since the NIK-deficient T cells from the draining lymph nodes, but not the spleen, were deficient in IL-2 production in response to KLH stimulation (Fig. 4a). In further support of this possibility, the NIK-deficient CD44 hi effector T cells derived from the draining lymph nodes of KLH-immunized mice had attenuated proliferation when restimulated in vitro with KLH (Fig. 4b,c). Thus, NIK appears to regulate the in vivo differentiation of Th1 and Th17 cells and the recall responses of the effector T cells.

NIK has a T cell-intrinsic function in regulating EAE pathogenesis.
To further examine the T cell-intrinsic role of NIK in regulating T cell differentiation and effector function in vivo, we employed a T cell-dependent autoimmunity model, experimental autoimmune encephalomyelitis (EAE) 18 . EAE is an animal model of the neutroinflammatory disease multiple sclerosis, and the development of EAE involves peripheral priming of T cells by an central nervous system (CNS)-specific autoantigen and the subsequent migration of autoimmune effector T cells to the CNS [18][19][20] . In particular, both Th1 and Th17 subsets of CD4 + effector T cells play an important role for the pathogenesis of EAE 20 . We induced EAE by immunizing mice with a myelin-specific autoantigen, myelin oligodendrocyte glycoprotein (MOG) peptide MOG 35-55 . As expected, immunization of wildtype mice led to the development of severe EAE clinical scores (Fig. 5a). In sharp contrast, the NIK-TKO mice were largely refractory to the induction of EAE (Fig. 5a). Flow cytometric analysis revealed that the NIK-TKO mice had substantially reduced frequencies of CNS-infiltrating CD4 + and CD8 + T cells as well as CD11b + CD45 hi cells, known to contain infiltrating monocytes and activated microglia (Fig. 5b). Concomitantly, the frequency of the resting CNS-resident microglial cells, characterized by CD11b + CD45 lo surface markers, was greatly increased in the NIK-TKO mice (Fig. 5b). Since the NIK-TKO mice had drastically reduced total CD45 hi infiltrating immune cells, the absolute number of CNS-infiltrating T cells and monocytes in the EAE-induced NIK-TKO mice was strikingly lower than that in the EAE-induced wildtype mice (Fig. 5c).
To examine the role of NIK in regulating the production of inflammatory effector T cells, we analyzed the number of IFN-γ -producing Th1 and IL-17A-producing Th17 cells in the CNS and draining lymph nodes of the MOG-immunized mice. Compared with the wildtype mice, the NIK-TKO mice had a much smaller number of Th1 and Th17 cells in the CNS (Fig. 5d). While this result was obviously due to the reduction in the number of total CD4 + T cells in the NIK-TKO CNS (Fig. 5c), the NIK-TKO mice also had a reduced number of Th1 and Th17 cells in the draining lymph node (Fig. 5d). The lymph node T cells isolated from MOG 35-55 -immunized NIK-TKO mice also produced lower amounts of secreted IFN-γ and IL-17A relative to lymph node T cells derived from MOG 35-55 -immunized wildtype mice, following in vitro restimulation with MOG 35-55 peptide (Fig. 5e). These results, along with the result shown in Fig. 4a, suggest that NIK plays a role in mediating the production of Th1 and Th17 subsets of CD4 + effector cells in vivo.
Since Treg cells are involved in the regulation of EAE pathogenesis 21 , we examined the effect of NIK deficiency in T cells on Treg induction during EAE. Compared to the wildtype mice, the NIK-TKO mice had reduced frequency of Treg cells in both the draining lymph node and the spleen (Fig. 5f). This result, which was in line with the Treg cell reduction under homeostatic conditions (Fig. 2c,d), suggests that the resistance of NIK-TKO mice to EAE induction is not due to overproduction of Treg cells but rather results from a defect in effector T cell generation.

NIK ablation in T cells attenuates T cell responses to bacterial infections. T cells, particularly
IFN-γ -producing Th1 and CD8 + T cells, play an important role in host defense against infections by the intracellular bacterial pathogen Listeria monocytogenesis (L. monocytogenesis) 22,23 . We examined the T cell-intrinsic role of NIK in mediating immune responses against L. monocytogenesis by infecting wildtype and NIK-TKO mice with a modified L. monocytogenes strain encoding chicken ovalbumin (LM-OVA) 24 . Compared to the wildtype mice, the NIK-TKO mice had a significantly higher L. monocytogenes load in the liver, indicating compromised immune response (Fig. 6a). Consistently, the spleens of the infected NIK-TKO mice had a reduction in the frequency and absolute number of CD4 + and CD8 + T cells (Fig. 6b). The frequency and absolute number of CD44 hi CD4 + and CD8 + T cells, representing effector T cells, were also substantially lower in the spleen of NIK-TKO mice (Fig. 6c,d). Since IFN-γ -producing CD4 + and CD8 + T cells play a crucial role in host defense against L. monocytogenes infection, we examined the frequency of antigen-specific effector T cells in the infected mice based on in vitro restimulation by antigens. As expected, a proportion of IFN-γ -producing CD8 + and CD4 + effector T cells was detected in the splenocytes of wildtype mice upon in vitro restimulation with an MHC class I-restricted OVA peptide (OVA 257-264 ) and an MHC class II-restricted listeriolysin O peptide  ), respectively (Fig. 6c). Importantly, the frequency and absolute number of the IFN-γ -producing antigen-specific CD8 + and CD4 + effector T cells were profoundly reduced in the NIK-TKO mice (Fig. 6c,d). These results suggest a T cell-intrinsic role for NIK in regulating in vivo T cell responses to bacterial infections.

Discussion
The role of NIK in regulating immune responses has been complicated by the essential role of NIK in regulating the development of lymphoid organs, which are required for T cell selection in the thymus and peripheral lymphocyte activation. In the present study, we studied the T cell-intrinsic functions of NIK in the regulation of immune and autoimmune responses. Since NIK functions in stromal cells to regulate lymphoid organ development, the NIK-TKO mice did not display any obvious abnormalities in the development of lymphoid organs. The T cell-specific ablation of NIK also did not affect the development of conventional T cells or Treg cells in the thymus. This result suggests that NIK expression in T cells is not required for thymocyte development or thymic Treg production, which is consistent with the previous finding that NIK functions in thymic epithelial cells to regulate thymocyte selection and Treg development 9,25 .
Our present work revealed an important T cell-intrinsic role for NIK in regulating homeostasis of T cells, particularly the memory like T cell and Treg cell populations, in the peripheral lymphoid organs. T cell homeostasis is a mechanism that maintains the peripheral pools of mature T cells. This mechanism relies on weak TCR signals, triggered by self-peptide/MHC ligands, and the common γ -chain cytokine IL-7 26 . In addition, the homeostasis of memory-like T cells also requires costimulatory signals from TNF receptor (TNFR) superfamily members, such as OX40 and CD30 27,28 . We have previously shown that NIK is required for the induction of noncanonical NF-κ B activation in T cells by the TCR and CD28 signals 29 . NIK also mediates T cell costimulation by OX40 and likely additional members of the TNFR family of costimulatory molecules cell subsets 11,30 . Thus, NIK may regulate T cell homeostasis via signaling pathways from both the TCR and TNFRs.
Our data suggest that NIK is dispensable for the initial activation of naïve T cells. This finding is consistent with the slow kinetics of NIK induction and noncanonical NF-κ B activation along with T cell activation 29 . It is likely that canonical NF-κ B plays a major role in mediating T cell activation. Activated T cells are induced to express various TNFR family members that mediate induction of NIK and noncanonical NF-κ B signaling, which in turn may contribute to the generation and subsequent function of effector T cells. We found that NIK-TKO mice had a severe defect in producing antigen-specific effector T cells in response to challenge by a protein antigen, KLH, and the CNS-specific self-antigen MOG as well as by the bacterium L. monocytogenes. NIK deficiency only moderately inhibited the induction of Th17 cells in vitro in a system that involves T cell activation by anti-CD3 and anti-CD28 antibodies. One of the differences between the in vitro and in vivo systems is that the in vivo system more critically involves the contribution of TNFR members, whose ligands are abundantly expressed on antigen-presenting cells. Our in vivo model of studies clearly demonstrated a role for NIK in regulating the generation of different subsets of effector T cells, particularly the Th1 and Th17 subsets of inflammatory T cells. In addition, NIK appeared to play a role in the recall responses of antigen-specific effector T cells.
Using conventional NIK-KO mice, we and others have previously shown that NIK regulates EAE induction 10,12 . However, how NIK regulates EAE pathogenesis has been in debate, because NIK functions in different cell types including DCs and B cells, which are also involved in the pathogenesis of EAE 10,12 . Although bone marrow and T cell transfer studies have been performed, these studies were not conclusive. Using T cell-conditional NIK KO mice, we have now provided definitive evidence for a T cell-intrinsic function of NIK in regulating the effector functions of Th1 and Th17 cells and induction of EAE. A previous study proposes that NIK functions in DCs, but not in T cells, in the regulation of EAE induction, since transgenic expression of NIK in DCs rescues the defect of the NIK-deficient mice in EAE induction 12 . We believe that the discrepancy between our present work and the previous study is due to the use of different mouse models. In particular, the expression of NIK in NIK transgenic mice is extremely high, as opposed to the low physiological level of NIK expression. Furthermore, although NIK transgene is mainly expressed in DCs, the transgenic mice also have a low level of NIK expression in other cells including T cells, thus complicating the phenotype 12 . Nevertheless, the previous studies, together with our present work, suggest that NIK may function in both T cells and DCs to regulate inflammatory T cell responses. Future studies will confirm the DC-specific function of NIK using DC-conditional NIK KO mice.
The mice were maintained in a specific pathogen-free facility, and the animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of the University of Texas MD Anderson Cancer Center. Immunoblot (IB) Analysis. Naïve CD4 + T cells were stimulated for 48 h with plate-bound anti-CD3 and anti-CD28 antibodies and lysed in radioimmunoprecipitation assay (RIPA) buffer for IB analysis of p100 processing to p52. For NIK IB analysis, the T cells were activated as indicated above and then further expanded in medium supplemented with IL-2 (10 units/mL) for 72 h. Since NIK is constantly degraded 32 , a proteasome inhibitor, MG132 (10 μ M), was added to the cell culture during the last 5 h to further enhance the level of NIK. Whole-cell lysates were subjected to immunoprecipitation (IP) using anti-NIK antibody followed by detecting the precipitated NIK by IB (using anti-NIK) as described 32 .

Flow cytometry, cell sorting and intracellular cytokine staining (ICS). Flow cytometric analyses
and cell sorting were performed 33 using a LSRII FACSFortessa (BD) and FACSAria (BD), respectively. For ICS, cells were stimulated with 50 ng/ml PMA plus 750 ng/ml ionomycin for 6 h in the presence of 10 μ g/ml protein transport inhibitor monensin, and the fixed cells were incubated with the indicated antibodies and subjected to flow cytometry.
Immunization and T cell recall response assays. Age-matched 6-8 weeks old WT and NIK-TKO mice were immunized s.c. (at the base of tail) with Keyhole Limpet Hemocyanin (KLH, 0.5 mg/ml) emulsified in CFA (0.5 mg/ml) at a dose of 100 μ l per mouse. Seven days after immunization, these mice were sacrificed and analyzed individually. To analyze the role of NIK in regulation of T cell responses in vivo, splenocytes and draining lymph node cells from KLH-immunized mice were cultured in 96-well plates with or without the antigen KLH. After 3 days, ELISA was performed using cell supernatant to measure cytokines. by a standard curve method and was normalized to the expression of β-Actin (Actb). The primers used in qRT-PRC assays are shown below. β-Actin forward, 5′ -CGTGAAAAGATGACCCAGATCA-3′; and reverse, 5′ -CACAGCCTGGATGGCTACGT-3′. mIL-2 forward, 5′ -CCTGAGCAGGATGGAGAATTACA-3′; and reverse, 5′ -TCCAGAACATGCCGCAGAG-3′.
Statistical analysis. Two-tailed unpaired t test statistical analysis was performed using Prism software. The p values < 0.05 were considered significant, and the level of significance was indicated as *p < 0.05, **p < 0.01, and ***p < 0.001.