Survival of mature T cells depends on signaling through HOIP

T cell development in the thymus is controlled by a multistep process. The NF-κB pathway regulates T cell development as well as T cell activation at multiple differentiation stages. The linear ubiquitin chain assembly complex (LUBAC) is composed of Sharpin, HOIL-1L and HOIP, and it is crucial for regulating the NF-κB and cell death pathways. However, little is known about the roles of LUBAC in T-cell development and activation. Here, we show that in T-HOIPΔlinear mice lacking the ubiquitin ligase activity of LUBAC, thymic CD4+ or CD8+ T cell numbers were markedly reduced with severe defects in NKT cell development. HOIPΔlinear CD4+ T cells failed to phosphorylate IκBα and JNK through T cell receptor-mediated stimulation. Mature CD4+ and CD8+ T cells in T-HOIPΔlinear mice underwent apoptosis more rapidly than control T cells, and it was accompanied by lower CD127 expression on CD4+CD24low and CD8+CD24low T cells in the thymus. The enforced expression of CD127 in T-HOIPΔlinear thymocytes rescued the development of mature CD8+ T cells. Collectively, our results showed that LUBAC ligase activity is key for the survival of mature T cells, and suggest multiple roles of the NF-κB and cell death pathways in activating or maintaining T cell-mediated adaptive immune responses.

lymphangiectasia 23 . However, the precise role of HOIP or LUBAC ligase activity in T cell development is poorly understood.
Here, we demonstrated that T cell-specific T-HOIP Δlinear mice showed impairments of mature T cell development and proliferative responses. Those data highlighted the HOIP-mediated NF-κ B pathway as a crucial pathway in the regulation of T cell development. Furthermore, our data indicated that deficiency of LUBAC ligase activity disturbed the development of mature T cells and their function, suggesting the important role of LUBAC for T cell-mediated adaptive immune responses.

The deficiency of Rnf-31 ligase activity in T cells impaired the development of mature T cells in the thymus.
To evaluate the involvement of HOIP in T cell development, we established Rnf31 Δlinear/Δlinear mice with a CD4-Cre transgene (T-HOIP Δlinear mice). The frequency of TCRβ + cells in the thymus was reduced in T-HOIP Δlinear mice and the relative and absolute numbers of CD4 + CD8 − and CD4 − CD8 + cells were markedly reduced in T-HOIP Δlinear mice whereas CD4 + CD8 + cells were not depressed (Fig. 1a,b). The effect was much stronger in CD4 − CD8 + cells than CD4 + CD8 − cells. The frequency of TCRβ + cells in T-HOIP Δlinear mice was equivalent to that of Rnf31 +/+ mice with CD4-Cre transgene (HOIP +/+ ) mice (Fig. 1a). Mature CD4 − CD8 + cells and CD4 + CD8 − T cells in the thymus downregulate CD24 and CD69 during the final maturation steps 15 . T-HOIP Δlinear mice had relatively higher frequencies of CD24-positive and CD69-positive cells in both CD4 + CD8 − TCRβ + and CD4 − CD8 + TCRβ + fractions than did HOIP +/+ mice (Fig. 1c). These results suggested that HOIP-mediated ligase activity was required for final maturation or survival of mature CD4 + CD8 − and CD4 − CD8 + T cells in the thymus. Data are shown as means ± SEM. *p < 0.05 (c) Thymocytes from T-HOIP Δlinear mice or HOIP +/+ mice were stained with anti-CD4, anti-CD8α , anti-TCRβ , anti-CD24 and anti-CD69 antibodies and the frequencies of cells expressing CD24/TCRβ or CD69 were determined by flow cytometry using gates for CD4 + CD8 + (DP), CD4 + CD8 − (CD4SP) and CD4 − CD8 + (CD8SP) cells. The number indicates the percentage of CD69 + cells in the DP, CD4SP and CD8SP cell fractions. The data in these figures are representative of four independent experiments.
Scientific RepoRts | 6:36135 | DOI: 10.1038/srep36135 T-HOIP Δlinear mice had lower numbers of mature T cells in the spleen and lymph nodes. We next assessed the T cell numbers and phenotypes in the spleen and lymph nodes of T-HOIP Δlinear mice. The relative frequencies of TCRβ + to TCRγ + cells or TCRβ + cells to B220 + cells was markedly reduced in the spleen and lymph nodes of T-HOIP Δlinear mice (Fig. 2a). The total cell numbers of TCRβ + , CD4 + and CD8 + T cells in the spleen and lymph nodes in T-HOIP Δlinear mice were also much less than in control mice (Fig. 2a). The relative frequency of CD8 + cells to CD4 + cells was reduced in HOIP −/− mice in the spleen (Fig. 2a). T-HOIP Δlinear mice possessed higher numbers of CD44 hi CD62L lo CD4 + and CD44 hi CD62L lo CD8 + T cells compared with control mice (Fig. 2b), suggesting that mature T cells from T-HOIP Δlinear mice had undergone activation after being exported from the thymus. The relative frequency of CD4 + Foxp3 + regulatory T cells in CD4 + T cells was not affected in the lymph nodes of T-HOIP Δlinear mice (Fig. 2c), whereas the frequency of CD4 + CD1d tetramer + NKT cells was reduced in the thymus and liver of T-HOIP Δlinear mice (Fig. 2d). Taken together, those data demonstrated that HOIP-deficiency in T cells markedly impaired the differentiation or survival of both mature CD4 + and CD8 + T cells with striking defects in the development of NKT cells.
CD4 + T cells in T-HOIP Δlinear mice proliferated poorly after TCR ligation. We evaluated the in vitro proliferative ability of CFSE-labeled CD4 + T cells from T-HOIP Δlinear mice when stimulated by anti-CD3 mAb Data are presented as means ± SEM. **p < 0.01. Spleen cells or liver lymphocytes from T-HOIP Δlinear or HOIP +/+ mice were stained with a combination of (b) anti-CD4, anti-CD8α , anti-CD44 and anti-CD62L antibodies, or (c) anti-CD4, anti-TCRβ and anti-Foxp3 or (d) anti-CD4 and anti-CD1d tetramer. The frequency of CD44/CD62L cells was evaluated by flow cytometry by gating on CD4 + CD8 − (CD4SP) or CD4 − CD8 + (CD8SP), or CD4/Foxp3 using a primary FSC/SSC gate to identify lymphocytes expressing CD4/ CD1d. The number indicates the percentage of each population within the viable population. The data in these figures are representative of four independent experiments. exposure. Those CD4 + T cells showed less CFSE dilution than did control cells, indicating relatively slower proliferative activity (Fig. 3a). The poorer proliferative activity of CD4 + T cells from T-HOIP Δlinear mice was not rescued by the addition of IL-2 to the culture medium (Fig. 3a). To assess the role of HOIP in the functional differentiation of CD4 + T cells, we immunized T-HOIP Δlinear mice and HOIP +/+ mice with OVA protein and evaluated anti-OVA specific immunoglobulin production and levels of serum cytokine 14 days after immunization. The CD4 + T cells from OVA-immunized T-HOIP Δlinear mice produced lower amounts of IFN-γ than those from HOIP +/+ mice (Fig. 3b) and failed to produce anti-OVA-specific IgG, IgG1 and IgG2c (Fig. 3c).
CD4 + T cells in T-HOIP Δlinear mice failed to phosphorylate IκBα. We assessed the role of HOIP ligase in the activation of NF-κ B in mature T cells. Thus, CD4 + T cells from T-HOIP Δlinear mice and HOIP +/+ mice were stimulated by anti-CD3 mAb and phosphorylation of Iκ Bα was evaluated. Anti-CD3 mAb treatment of CD4 + T cells from T-HOIP Δlinear mice induced less phosphorylation of Iκ Bα than observed in cells from control mice (Fig. 4a). We then analyzed nuclear translocation of NF-κ B (p65) after anti-CD3 mAb-stimulation of CD4 + T cells from T-HOIP Δlinear mice or HOIP +/+ mice. Little nuclear translocation of p65 was found in CD4 + T cells from T-HOIP Δlinear mice compared with efficient translocation of p65 into the nucleus in control T cells (Fig. 4b).
The deficiency of HOIP ligase activity in B cells disturbs CD40 but not B cell receptor-mediated JNK activation 24 . Thus, we analyzed TCR-mediated JNK phosphorylation in T-HOIP Δlinear mice T cells. Stimulation with anti-CD3 mAb induced less phosphorylation of JNK in HOIP ligase-deficient T cells compared with control cells (Fig. 4c). Those data demonstrated that the deficiency of HOIP ligase activity disturbed activation of not only the canonical NF-κ B pathway but also the JNK pathway. or T-HOIP Δlinear (filled gray) mice were stimulated for 3 days on plates coated with anti-CD3 mAb (1 μ g/mL) in the absence or presence of recombinant IL-2 (10 U/mL). HOIP +/+ or T-HOIP Δlinear mice at the age of 8 weeks were immunized by OVA protein (10 μ g/mL) emulsified in CFA. (b) Serum IFN-γ was evaluated by ELISA ten days after immunization. Data show means ± SEM. **p < 0.01. (c) Serum anti-OVA IgG, IgG1 or IgG2c levels were evaluated by ELISA ten days after immunization. Data show means ± SEM. **p < 0.01. The data in these figures are representative of four independent experiments.
Scientific RepoRts | 6:36135 | DOI: 10.1038/srep36135 T cells from HOIP-deficient mice lost viability. We sought to assess whether the loss of mature T cells in T-HOIP Δlinear mice was attributable to increased cell death. Thus, mature CD4 + or CD8 + T cells from the thymus and spleen were stained with Annexin V and 7AAD. Larger percentages of CD8 + T cells in the thymus and CD4 + or CD8 + T cells in the spleen of T-HOIP Δlinear mice were positive for Annexin V than in control cells (Fig. 5a). Those data suggested that the deficiency of HOIP ligase activity increased the frequency of cell death in mature T cells, especially in early developmental stages of single positive cells in the thymus.
To confirm that CD4 + or CD8 + T cells from T-HOIP Δlinear mice did not retain viability, we compared cell survival of CD4 + and CD8 + T cells from T-HOIP Δlinear mice and HOIP +/+ mice. When CD4 + T cells from T-HOIP Δlinear mice (CD45.2) and control (CD45.1) mice were cultured in vitro without any stimulation or after stimulation with anti-CD3 mAb, T cells from T-HOIP Δlinear mice died more rapidly than those from HOIP +/+ mice (Fig. 5b). To determine if the impaired T cell survival also occurred in vivo, CD4 + T cells (CD45.2) from T-HOIP Δlinear mice or HOIP +/+ mice were transferred into recipient C57BL/6 (CD45.1) mice. The number of CD4 + cells from T-HOIP Δlinear mice was much lower than control cells 3 days after transfer into inguinal lymph nodes, (Fig. 5c). Those results suggested that CD4 + T cells from HOIP −/− mice were prone to die compared with control CD4 + T cells.

CD127 expression was lower in T cells from T-HOIP Δlinear mice.
In order to determine the molecular mechanisms for impaired development of T cells from T-HOIP Δlinear mice, we tested the expression of cytokine receptors on T cells. The expression levels of common γ -chain (CD132) were comparable between CD4 + and CD8 + T cells in T-HOIP Δlinear mice and HOIP +/+ mice (Fig. 6a). However, the expression levels of IL-2Rα (CD25) and IL-2Rβ (CD122) were higher in splenic CD8 + T cells from T-HOIP Δlinear mice. Moreover, the expression levels of IL-7Rα (CD127) were relatively high in CD4 + and CD8 + splenic T cells from T-HOIP Δlinear mice (Fig. 6a). In contrast, the expression of CD127 was lower in thymic CD4 + and CD8 + T cells from T-HOIP Δlinear mice (Fig. 6b).
Il7r (CD127) was reported to be a target gene for NF-κ B signaling 25 . As IL-7 is required for CD8 + T cell survival, we examined if impaired development of T cells from T-HOIP Δlinear mice was attributable to low CD127 expression. CD127-encoding retrovirus was infected in fetal thymocytes from T-HOIP Δlinear mice. The GFP-expressing thymocytes were cultured in fetal thymus for 7 days and the development of mature T cells was examined. The overexpression of CD127 increased the frequency of mature CD8 + TCRβ + but not CD4 + TCRβ + T cells (Fig. 6c). Those data indicated that impaired CD8 + T cell survival in T-HOIP Δlinear mice is, at least partly, attributable to low CD127 expression.

Discussion
LUBAC-mediated poly-linear ubiquitination is a crucial event for activating the NF-κ B pathway 21,20 . However, the roles of LUBAC-mediated NF-κ B regulation in T cell activation or in development have been unresolved. In this paper, we show that among the LUBAC components, HOIP ligase activity is required for the development of mature T cells and is crucial for CD4 + T cell proliferation. T-HOIP Δlinear mice T cells failed to upregulate CD127, which was attributable to the impaired survival of thymic CD8 + T cells but not CD4 + T cells in T-HOIP Δlinear mice. These findings demonstrate the crucial contribution of HOIP-mediated linear ubiquitination of NEMO to T cell development. They support a model in which CD4 + and CD8 + T cells have distinct molecular requirements for NF-κ B-mediated molecules downstream.
T cell development in the thymus is controlled by a multistep process utilizing the TCR, costimulatory molecules and cytokine signals, each of which is required during specific stages of development. Given that the TCR and cytokines signaling are crucial for T cell development, with NF-κ B downstream for various receptors in conventional T cells, HOIP could control thymic T cell differentiation at multiple points. Our data demonstrated that mature CD4 + or CD8 + T cells were markedly diminished with reduced expression of CD127 in T-HOIP Δlinear mice, a deficit that was rescued by overexpressing CD127 on CD8 + T cells, at least in an in vitro culture system. IL-7 functions in the survival and development of conventional CD4 + and CD8 + T cells, as evidenced by a markedly reduced number of mature CD4 + and CD8 + T cells in CD127-deficient mice 15 . Therefore, the impaired survival of CD8 + T cells in T-HOIP Δlinear mice could be, at least partially, attributable to the reduced expression of CD127. In contrast, the development of CD4 + T cells could not be rescued by overexpressing CD127, suggesting that the dysregulation of other target molecules downstream from HOIP is responsible for the impaired survival of CD4 + T cells. Those data suggest a model in which CD4 and CD8 T cells require distinct regulation of target molecules downstream of HOIP for their survival in the thymus.
HOIP complexes with HOIL-1L and SHARPIN 26,22 . Mutations in the murine Sharpin gene cause spontaneous chronic proliferative dermatitis (cpdm) that develops into psoriasis-like proliferative skin lesions, splenomegaly, absence of Peyer's patches and low levels of serum immunoglobulin 27 . A recent study reported that patients with a loss-of-function mutation in HOIL-1L suffered from chronic autoinflammation, invasive bacterial infections and muscular amylopectinosis 28 . Furthermore, an inherited mutation in HOIP causes multi-organ autoinflammation, combined immunodeficiency, subclinical amylopectinosis, and systemic lymphangiectasia 23 . These findings suggest a distinct requirement for each LUBAC subunit to control downstream pathways. In contrast to the autoinflammatory phenotypes associated with HOIP-or HOIL-1L-deficiency in humans, the present study revealed that a deficiency in HOIP ligase activity impaired NF-κ B activation leading to the impairment of both CD4 + and CD8 + T cell development without any inflammatory responses. As LUBAC ligase activity was deleted only in T cells in our mouse study, the loss of function of non-T cells might be involved in the development of inflammatory responses.
The CBM (CARMA1-Bcl10-Malt1) TCR adaptor complex regulates TCR-dependent NF-κ B activation [29][30][31] . Despite the important roles of CARMA1 in NEMO activation, CARMA1-deficient mice have normal T-cell development and normal peripheral T-cell numbers and ratios 32,33 . However, they do have a defect in the development of intrathymic CD4 + CD25 + regulatory T cells 34 . In contrast, T cell-specific, NEMO-deficient mice are devoid of mature CD4 + and CD8 + T cells in the thymus 18 , a finding that is similar to T cell-specific, HOIP ligase activity deficient mice. Furthermore, HOI -ligase activity deficient CD4 + T cells have a defect in TCR-mediated proliferation and NF-κ B activation. Those results suggest that the engagement of TCR activates NEMO by utilizing CARMA1-dependent or -independent pathways, and that LUBAC-mediated linear ubiquitination of NEMO through engagement of TCR is essential for the survival of mature T cells. In addition, a recent paper revealed that LUBAC integrates the CBM complex and that NF-κ B reporter activity is stimulated following antigen receptor ligation independent of its catalytic activity 35 . However as this study was performed by evaluating NF-κ B reporter activity in Jurkat cells that had been transfected with siRNA against HOIP and siRNA resistant ligase activity-inactive HOIP, the effect of the residual activity of endogenous HOIP might not be negligible. In future studies, it will be necessary to evaluate which domains of HOIP are crucial for binding with the CBM complex.
In this report, we found that canonical NF-κ B signaling through linear ubiquitination by LUBAC was an essential molecular pathway that regulated CD4 + and CD8 + T cell development. Our data highlight a previously unknown molecular link between LUBAC and mature CD4 + and CD8 + T cell survival. Those data also suggest new approaches for inhibiting HOIP ligase activity and thereby suppressing T-cell-mediated immune responses.

Methods
Mice. Six-to 8-week-old C57BL/6 mice were purchased from Japan SLC (Hamamatsu, Japan). Rnf31 Δlinear/Δlinear mice were previously described 24 . C57BL/6 mice (CD45.1) and CD4-Cre transgenic mice were purchased from Jackson Laboratory (MA, USA). All animal experiments were approved by an animal research ethical committee of Tokushima University and were performed according to its guidelines.
OVA immunization. Mice were immunized with OVA protein (50 μ g) emulsified in CFA (Sigma, Saint Louis, MO, USA) and the titers of OVA-specific antibodies (IgG, IgG1 and IgG2c) were measured by ELISA using HRP-conjugated anti-mouse IgG, IgG1 or IgG2c (Southern Biotech, Alabama, USA) as the secondary antibodies.
ELISA. ELISA for IFN-γ was performed using an ELISA kit from eBioscience.
Confocal laser-scanning microscopy analyses. T cells were isolated with a pan-T cell isolation kit (Miltenyi Biotec) and stimulated for 10 min at 37 °C with anti-CD3 mAb followed by anti-hamster IgG. Cells were then seeded on poly-l-lysine hydrobromide-coated cover glass, fixed with 4% paraformaldehyde and permeabilized with acetone. Staining with anti-p65 mAb (1 μ g/mL) (Santa Cruz Biotech) was followed by Alexa Fluor 546-conjugated goat anti-rabbit IgG (Invitrogen). The nucleus was stained with DAPI. The observations were performed using an FV10i confocal microscope (OLYMPUS, Japan). Several cells were analyzed for each labeling condition, and representative results are presented.
Fetal thymic organ culture. Fetal thymus (fetal age, day 15) from C57BL/6 mice was cultured in the presence of deoxyguanosine (1.35 mM) (Sigma) on Transwell plates for 7 days. Fetal thymocytes (fetal age, day 15) from HOIP +/+ or HOIP −/− mice were isolated and infected with control retrovirus or Il7r-encoding-retrovirus as previously reported 36 . The pKE004 retrovirus vector 37 that encodes IRES-GFP and Il7r was transfected into Plat-E cells 38 to generate retrovirus. The infected thymocytes were cultured for one day in the presence of IL-7 (5 ng/mL) (eBioscience). Thymocytes were hanging-drop cultured with deoxyguanosine-treated thymus using a Terasaki plate for one day. Then the thymus was cultured on a Transwell plate for 7 days.

Statistical analysis.
For all experiments, the significance of differences between groups was calculated using the Mann-Whitney U test for unpaired data. Differences were considered significant when p < 0.05.