Mice homozygous for a Gnb2l1 conditional allele (Gnb2l1^F/F) were previously described and were all backcrossed to the C57BL/6 background (15). Mice expressing a transgene encoding Cre recombinase driven by the lymphocyte-specific protein tyrosine kinase proximal promoter (Lck-Cre mice) (17) and under the control of CD4 promoter (CD4-Cre mice) (18, 19) were gifts from Dr. Hua Han (The Fourth Military Medical University, Xi’an, China) and Dr. Chen Dong (Tsinghua University, Beijing, China), respectively. Specific inactivation of RACK1 in T cells was achieved by crossing Gnb2l1^F/F mice with Lck-Cre mice or CD4-Cre mice. Gnb2l1^{F/F; lck-Cre} mice were kept in the animal facility of Institute of Biotechnology (20 Dongdajie, Beijing, China), whereas Gnb2l1^{F/F; CD4-Cre} mice were kept in the animal facility of Institute of Basic Medical Sciences (27 Taiping Road, Beijing, China). All mice were maintained under specific pathogen-free conditions and animals were handled in accordance with institutional guidelines. Genotyping was performed as described previously (15).

Fluorescence-labeled antibodies against CD3, CD4, CD8, CD25, CD44, CD62L, CD69, CD19, CD45.1, CD45.2, NK1.1, IFN-γ, and TNF-α were purchased from eBioscience. Fluorescence-conjugated antibody against Foxp3 was from BioLegend. α-GalCer-loaded recombinant murine CD1d tetramers (E001-2A-G) were purchased from ProImmune. Cells from thymi, spleens (SPLs), and mesenteric lymph nodes (mLNs) were depleted of erythrocytes by hypotonic lysis. The cells were washed with FACS washing buffer (2% FBS, 0.1% NaN₃ in PBS) twice and were then incubated with fluorescence-conjugated antibodies against cell surface molecules for 30 min on ice in the presence of 2.4G2 mAb to block FcγR binding. Isotype antibodies were included as negative controls. For Foxp3 and intracellular cytokine staining, single-cell suspensions were stained with fluorescence-conjugated antibodies against cell surface markers, fixed and permeabilized using a fixation/permeabilization kit (eBioscience), and stained with fluorescence-conjugated specific antibodies against Foxp3 and cytokines in accordance with the manufacturer’s instructions. Flow cytometry was performed using a Becton Dickinson FACSCalibur machine.

Antibodies against LC3B (3868S), Ub (3936S), and active Caspase-3 (9664S) were purchased from Cell Signaling. Antibody against β-actin (sc-8432) was obtained from Santa Cruz. Antibody against RACK1 (610178) was from BD Biosciences. Antibody against electron-transferring-flavoprotein dehydrogenase (ETFDH) (A6585) was from ABclonal Biotech. Antibody against very long-chain acyl-CoA dehydrogenase (VLCAD) (14527-1-AP) was from Proteintech. Total protein extracts were prepared and dissolved in SDS sample buffer. Protein extracts were separated on 10% SDS-PAGE gels and transferred to polyvinylidene difluoride membranes (Millipore). The membranes were probed with antibodies and visualized with an electrochemiluminescence kit (Amersham).

Bone marrow cells were depleted of T cells and antigen-presenting cells by complement-mediated cell lysis. Bone marrow cells (2 × 10⁶) from 8-week-old Gnb2l1^F/F and Gnb2l1^{F/F; CD4-Cre} mice were cotransferred into sublethally irradiated recipients (9 Gy). The congenic markers CD45.1 and CD45.2 were used for distinguishing cells from the different donors and recipients.

Splenic CD4⁺ or CD8⁺ T cells were isolated by negative selection from single-cell suspensions of SPLs using a CD4⁺ T cell isolation kit II or a CD8⁺ T cell isolation kit II (Miltenyi Biotec). To get naïve CD4⁺ or CD8⁺ T cells, purified CD4⁺ or CD8⁺ splenic cells were stained with fluorescence-labeled antibodies against CD44 and CD62L. Then CD44^lowCD62L^high subsets were sorted using a FACSVantage (BD Biosciences). To purify thymic double negative (DN), double positive (DP), CD4 single positive (SP), and CD8 SP cells, single-cell suspensions of thymi were stained with fluorescence-conjugated antibodies against cell surface markers CD4 and CD8; different subsets were then sorted using a FACSVantage (BD Biosciences). In each case, purity of the cell separation was verified by flow cytometry analysis. Cell purity was verified to be at least 95%.

After treatment with 10 µM chloroquine (CQ) for 16 h, the cells were fixed and permeabilized using a fixation/permeabilization kit (eBioscience). The cells were then incubated with rabbit polyclonal antibody against LC3B (L7543; Sigma) for 1 h at room temperature, followed by incubation with FITC-conjugated goat anti-rabbit IgG for 30 min at room temperature. After incubation with 1 µg/ml 4′,6-diamidine-2-phenylindole (DAPI) for 5 min, the cells were observed under a laser scanning confocal microscopy (RADIANCE 2100; Bio-Rad).

Naïve CD4⁺ or CD8⁺ T cells were labeled by incubation at the density of 1.0 × 10⁶/ml in RPMI 1640 with 0.1 µM carboxyfluorescein succinimidyl ester (CFSE; Invitrogen) at 37°C for 20 min, washed, and resuspended in the complete culture medium. Cells were stimulated with Dynabeads mouse CD3/CD28 T cell expanders (Invitrogen) in 96-well plates at the density of 2.5 × 10⁴ cells/well. Proliferation was assessed by flow-cytometric analysis of CFSE dilutions after 72 h of culture.

Mice received 1 mg thymidine analog BrdU (Sigma) in 0.1 ml PBS via i.p. injection. BrdU incorporation in CD4⁺ or CD8⁺ splenic T cells was analyzed by flow cytometry 24 h later. Staining of BrdU incorporation followed the BrdU Flow Kit (Becton Dickinson) protocol. Briefly, cells were dehydrated in an alcohol solution, fixed and permeabilized in 1% paraformaldehyde/0.01% Tween 20, treated with 50 U/ml DNase I, and then stained with 10 µl of FITC-conjugated anti-BrdU (Becton Dickinson).

Bone marrow-derived macrophages were obtained by culturing the non-adherent bone marrow cells from 6- to 8-week-old mice in RPMI 1640 medium containing 15% (v/v) FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin, and 50 µM 2-ME with 100 ng/ml M-CSF for 7 days.

Purified CD4⁺ or CD8⁺ T cells were seeded into a 96-well plate at the density of 2.5 × 10⁴ cells/well. The cells were stimulated with Dynabeads mouse CD3/CD28 T cell expanders (Invitrogen) or left untreated. 0 or 24 h after culture without stimulation or 72 h after stimulation, cells were stained with Annexin V-FITC and PI resuspended in 300 µl binding buffer containing calcium ion. Apoptosis was assessed by flow-cytometric analysis.

To stain mitochondria, cells were incubated for 30 min at 37°C with 100 nM MitoTracker Green (Molecular Probes) in RPMI 1640 complete medium before staining of surface markers. Mitochondrial content was assessed by flow-cytometric analysis.

Male Gnb2l1^{F/F; CD4-Cre} mice at the age of 8–12 weeks and their littermates were intravenously injected with Con A (Sigma) dissolved in pyrogen-free saline at a dose of 25 µg/g body weight. Survival was monitored for 120 h after injection.

Mice were euthanized, and livers were removed 48 h after Con A injection. The livers were fixed for another 48 h in 4% paraformaldehyde, dehydrated, infiltrated with paraffin, and sectioned at 5 µm. Slides were stained with hematoxylin/eosin.

Serum alanine transaminase (ALT) and aspartate transaminase (AST) levels were determined using commercially available kits (Abcam) according to the manufacturer’s protocols.

Mice were euthanized, and livers were removed. Then, livers were pressed through a 200-gage stainless steel mesh. After the filtrate was washed once, the cells were resuspended in 40% Percoll (GE Healthcare) and overlaid onto 70% Percoll. After centrifugation at 1,260 g for 30 min, the interphase was collected and washed once.

Results are shown as mean ± SD. Differences were considered significant with a P value of <0.05 using rank sum test, Student’s t-test (paired or unpaired), or one-way analysis of variance (ANOVA).

Specific inactivation of RACK1 in T cells was first achieved by crossing mice homozygous for a Gnb2l1 conditional allele (Gnb2l1^F/F) with mice expressing a transgene encoding Cre recombinase driven by the lymphocyte-specific protein tyrosine kinase (Lck) proximal promoter (17). Genotyping confirmed the presence of LoxP sites and lck-Cre in Gnb2l1^{F/F; lck-Cre} mice (data not shown). These mutants were born at the expected Mendelian ratio and appeared phenotypically normal (data not shown). RACK1 deficiency in thymocytes and splenic T cells was confirmed by IB (Figure 1A). However, CD19⁺ B cells in the SPL and BMDM showed unaltered levels of RACK1 protein (Figure 1A). Comprehensive analyses of T cell production in the thymus revealed that the percentages and absolute numbers of DN, DP, and SP CD4⁺ and CD8⁺ thymocytes appeared to unaltered in Gnb2l1^{F/F; lck-Cre} mice compared with littermate controls (Figure 1B). Furthermore, intrathymic development of Treg cells expressing Foxp3 transcription factor was normal in Gnb2l1^{F/F; lck-Cre} mice (Figure 1C). To further confirm the abovementioned observations, we also generated a T cell-specific RACK1 knockout mouse model (Gnb2l1^{F/F; CD4-Cre} mice) with CD4-Cre transgenic mice (18, 19). IB analysis confirmed the deficiency of RACK1 in thymocytes, especially in CD4 SP and CD8 SP subsets (Figure 1D). Consistent with the data obtained in Gnb2l1^{F/F; lck-Cre} mice, the dispensable role of RACK1 in intrathymic development of conventional T cells and Treg cells was also observed in Gnb2l1^{F/F; CD4-Cre} mice (Figures 1E,F).

We next examined the peripheral T lymphocyte compartment in Gnb2l1^{F/F; lck-Cre} mice and Gnb2l1^{F/F; CD4-Cre} mice. The percentages and absolute numbers of CD4⁺ and CD8⁺ T cells in the SPL and mLNs of Gnb2l1^{F/F; lck-Cre} mice and Gnb2l1^{F/F; CD4-Cre} mice were significantly reduced (Figures 2A,C). The general reduction in peripheral T cells was more profound in Gnb2l1^{F/F; CD4-Cre} mice than that in Gnb2l1^{F/F; lck-Cre} mice (Figures 2A,C). However, the percentages of Treg cells in peripheral CD4⁺ T cells remained unaffected in the absence of RACK1 (Figures 2B,D) even though their absolute numbers decreased proportionally due to the reduced numbers of peripheral CD4⁺ T cells (Figures 2B,D).

In lymphopenic mice, residual peripheral T cells usually express a marker profile characteristic of either activated effector T cells or central memory T cells (20, 21). The lymphopenia upon specific deletion of Atg3, Atg5, Atg7, Atg16l1, or Vps34 in T cells has been demonstrated to be associated with increased percentages of activation/memory-like T cells within both the CD4⁺ and CD8⁺ compartments (2–6). On the other hand, splenic CD4⁺ T cells did not show changed expression profile of CD44 and CD62L upon specific deletion of Beclin 1 in T cells, even though more splenic CD8⁺ T cells were CD44^highCD62L^low in the absence of Beclin 1 (7). In this scenario, we analyzed the expression profile of CD44 and CD62L in the presence and absence of RACK1 in T cells. CD4⁺ T cells in the SPL and mLNs of 6- to 8-week-old Gnb2l1^{F/F; lck-Cre} mice (Figure 3A) and Gnb2l1^{F/F; CD4-Cre} mice (Figure 3B) showed no increased percentage of CD44^highCD62L^low or CD44^highCD62L^high population as compared to their counterparts in littermates. Splenic CD8⁺ T cells of Gnb2l1^{F/F; lck-Cre} mice also demonstrated no enhanced activation/memory, whereas CD8⁺ T cells in the mLNs of Gnb2l1^{F/F; lck-Cre} mice exhibited a slight increase in the frequency of CD44^highCD62L^low population (Figure 3A). On the other hand, CD8⁺ T cells in the SPL and mLNss of Gnb2l1^{F/F; CD4-Cre} mice all showed higher percentages of CD44^high CD62L^low and CD44^highCD62L^high populations (Figure 3B). However, CD8⁺ T cells in the SPL and mLNs of Gnb2l1^{F/F; CD4-Cre} mice showed only slightly enhanced CD69 and CD25 expression under steady-state conditions (Figure S1 in Supplementary Material). Therefore, RACK1-deficient CD8⁺ T cells in vivo may be responding to homeostatic rather than antigen-induced expansion signals, which result in aberrant expression profile of CD44 and CD62L.

To distinguish T cell-intrinsic or -extrinsic mechanism(s) under-lying peripheral T cell lymphopenia upon specific deletion of RACK1 in T cells, we created mice with mixed bone marrow through transferring bone marrow cells from congenically marked Gnb2l1^F/F (CD45.1CD45.2) and Gnb2l1^{F/F; CD4-Cre} (CD45.2CD45.2) mice into sublethally irradiated CD45.1CD45.1 mice. We chose Gnb2l1^{F/F; CD4-Cre} mice because peripheral T cell lymphopenia in these mice is more profound than that in Gnb2l1^{F/F; lck-Cre} mice. After 8 weeks of bone marrow reconstitution, the CD19⁺ B-cell compartment showed a 1.62:1 ratio of CD45.1⁺CD45.2⁺ vs. CD45.1⁻CD45.2⁺ (Figure 4). Similar ratios were observed in DP, CD4 SP, and CD8 SP thymocyte subsets (Figure 4). However, peripheral T cells were overwhelmingly derived from RACK1-sufficient CD45.1⁺CD45.2⁺ cells (Figure 4). Interestingly, RACK1 was more important for the reconstitution of peripheral CD8⁺ T cells than peripheral CD4⁺ T cells (Figure 4). All of these findings are consistent with the phenotype of our Gnb2l1^{F/F; lck-Cre} mice and Gnb2l1^{F/F; CD4-Cre} mice in the steady state and suggest that RACK1 plays a crucial cell-autonomous role in maintaining the peripheral T cell number.

We have shown that RACK1 promotes autophagy in several types of mammalian cells in vivo and ex vivo (15). The impaired peripheral T lymphocyte compartment in RACK1-deficient mice was similar to that in mice lacking autophagy genes (1–9). It is possible that peripheral T cell lymphopenia in Gnb2l1^{F/F; lck-Cre} mice and Gnb2l1^{F/F; CD4-Cre} mice was due to impaired autophagy. In line with this notion, IB analysis revealed that splenic T cells from Gnb2l1^{F/F; CD4-Cre} mice showed diminished basal membrane-bound LC3B-II generation from cytosolic LC3B-I (Figure 5A), an indicator of autophagic activity (22, 23). Consistently, indirect immunofluorescence analysis demonstrated that peripheral T cells from Gnb2l1^{F/F; CD4-Cre} mice exhibited reduced numbers of LC3B-positive dots (autophagic LC3 puncta) after treatment with lysosomal protease inhibitor CQ (Figure 5B). Impaired autophagy onset was associated with the accumulation of polyubiquitinated proteins and enhanced production of active Caspase-3 fragment (Figure 5A).

Previous reports have demonstrated that autophagy limits the number of mitochondria in peripheral T cells (2–4, 6). As expected, RACK1-deficient CD8⁺ splenic T cells from the bone marrow chimeric mice had increased mitochondrial mass, as determined by staining with a cell-permeable mitochondrial dye (MitoTracker) (Figure 5C). RACK1-deficient CD4⁺ T cells also showed an increase in mitochondrial mass, albeit to a lesser extent (Figure 5C). Interestingly, RACK1-sufficient and -deficient CD4 SP and CD8 SP thymocytes exhibited comparable MitoTracker Green staining (Figure 5C), suggesting a specific role for RACK1 in peripheral T cell homeostasis but not T cell development. The same phenomena were also observed in RACK1-deficient T cells isolated from Gnb2l1^{F/F; CD4-Cre} mice as compared to their counterparts isolated from littermate controls (Figure S2 in Supplementary Material). Confocal microscopy also revealed that RACK1-deficient T cells had more MitoTracker-positive structures (Figure S3A in Supplementary Material). Furthermore, the MitoTracker-positive structures in RACK1-deficient cells frequently formed tubular clusters (Figure S3A in Supplementary Material). Ultrastructural analysis confirmed increased numbers of mitochondria and revealed certain elongated mitochondria in RACK1-deficient CD8⁺ T cells. In addition, electron microscopy showed that some membrane structures that are not characteristic of autophagosomes accumulated in RACK1-deficient CD8⁺ T cells (Figure S3B in Supplementary Material), similar to the observations in Atg3-deficient T cells (4). Consistently, IB analysis revealed that mitochondrial proteins ETFDH (electron transfer flavoprotein dehydrogenase) and VLCAD (very-long-chain acyl-CoA dehydrogenase) (24, 25) were accumulated in splenic T cells from Gnb2l1^{F/F; CD4-Cre} mice (Figure 5D).

Autophagy-deficient T cells readily undergo apoptosis because healthy mitochondria are crucial for T cell survival (2–4, 6). Thus, it is of importance to examine whether RACK1-deficient T cells had increased apoptosis. To exclude cell-extrinsic effects (e.g., activation-induced cell death due to lymphopenia), we used mixed bone marrow chimeras. In these mixed chimeras, RACK1-sufficient and -deficient splenic T cells exhibited similar basal apoptosis ex vivo (Figure 6A). However, RACK1-deficient splenic T cells were more vulnerable than their RACK1-sufficient counterparts after in vitro culture in exactly the same system (Figure 6A). Moreover, more RACK1-deficient splenic T cells underwent apoptosis than their RACK1-sufficient counterparts did upon TCR ligation (Figure 6A). These data suggest a cell-intrinsic requirement for RACK1 in maintaining peripheral T cell survival.

Besides increased sensitivity to cell death, the reduced numbers of peripheral T cells in the absence of RACK1 might also result from decreased proliferation potential since previous results have indicated that Atg3-, Atg5-, Atg7- or Vps34-deficient T cells cannot proliferate efficiently (2–4, 8, 9). Indeed, RACK1-deficient CD4⁺ T cells in the chimeras, but not CD8⁺ T cells, showed reduced BrdU incorporation as compared to RACK1-sufficient counterpart in the steady state (Figure 6B). Next, we tested how RACK1-deficient T cells respond to TCR-mediated stimulation. For this purpose, naïve CD4⁺ and CD8⁺ T cells were purified from Gnb2l1^{F/F; CD4-Cre} mice and their control littermates and labeled with the fluorescent-dye CFSE. After 3 days of anti-CD3 and anti-CD28 stimulation, the cells were subjected to flow cytometry analysis of CFSE dilution. To exclude the possibility that the possible proliferation defect is caused by cell death, all cells in the CFSE dilution assay were gated on Annexin V-negative cells. We found that both CD4⁺ and CD8⁺ T cells proliferated less efficiently in the absence of RACK1 (Figure 6C). Despite that, in vitro differentiation of naïve CD4⁺ T cells isolated from Gnb2l1^{F/F; CD4-Cre} mice was not impaired under Th1 polarization conditions as compared with their counterparts isolated from control littermates (Figure S4 in Supplementary Material).

Invariant natural killer T cells are a small population of T cells whose development in the thymus requires expression of the Vα14-Jα18 invariant TCR and its recognition of endogenous glycolipid ligands presented by CD1d expressed on DP thymocytes (8, 26, 27). After their development, these cells populate peripheral lymphoid organs, mainly liver and SPL (8, 26, 27). An essential role for autophagy during iNKT cell development has been disclosed (8, 26, 27). In this scenario, it is of interest to explore how T cell-specific RACK1 deficiency might affect iNKT cell development. As expected, we found that Gnb2l1^{F/F; CD4-Cre} mice exhibited an obvious defect in intrathymic iNKT cell development; only about 20% CD3⁺CD1d-tetramer⁺ cells were observed as compared with control littermates (Figure 7). Gnb2l1^{F/F; CD4-Cre} mice showed more profound defect in the iNKT cell population in the SPL and liver as compared to their control littermates (Figure 7). Thus, thymic iNKT cell development is blocked and mature iNKT cells are absent in peripheral lymphoid organs in the absence of RACK1.

Our data indicate that specific deletion of RACK1 in T cells leads to peripheral T cell lymphopenia and impaired iNKT cell development. In this scenario, we tested how T cell-specific RACK1 deficiency might affect Con A-induced acute liver injury because T and NKT cells play essential roles in this model (28–30). A high dose of Con A (25 µg/g body weight) was injected into Gnb2l1^{F/F; CD4-Cre} mice and their littermates, and their survival was monitored. As expected, Gnb2l1^{F/F; CD4-Cre} mice all survived under the conditions that almost half of their littermates died (Figure 8A). Histological analysis revealed that Gnb2l1^{F/F; CD4-Cre} mice had very limited liver damage, while their littermates had extensive liver necrosis (Figure 8B). Measurement of serum AST and ALT levels also indicated that Gnb2l1^{F/F; CD4-Cre} mice had significantly reduced liver injury (Figure 8C).

Various cytokines, including IFN-γ and TNF-α, are involved in the pathogenesis of Con A-induced hepatitis (28–30), so we wondered whether RACK1 deficiency affects cytokine production. As shown in Figure 8D, Gnb2l1^{F/F; CD4-Cre} mice had reduced percentages and absolute numbers of CD3⁺ NK1.1⁻ T cells in the liver after Con A treatment as compared to their littermates. Furthermore, the percentage of IFN-γ⁺ T cells in Gnb2l1^{F/F; CD4-Cre} livers after Con A injection was significantly downregulated, while Con A-induced production of TNF-α in these cells was only slightly affected in Gnb2l1^{F/F; CD4-Cre} livers (Figure 8E). Surprisingly, Gnb2l1^{F/F; CD4-Cre} mice and control littermates showed comparable absolute numbers of CD3⁺NK1.1⁺ NKT cells as well as CD3⁻NK1.1⁺ NK cells in the liver after Con A treatment (Figure 8D), even though both subsets exhibited reduced IFN-γ expression (Figure 8E). In this scenario, we examined the relationship between CD3⁺NK1.1⁺ subset and CD3⁺CD1d-tetramer⁺ subset in the liver of Gnb2l1^{F/F; CD4-Cre} mice and control littermates. As previously reported (29, 30), most NKT cells in the liver of Gnb2l1^F/F mice were CD1d-tetramer⁺NK1.1⁺ iNKT cells under steady-state conditions when CD3⁺ cells were gated (Figure S5 in Supplementary Material). After Con A treatment, the percentage of CD1d-tetramer⁺NK1.1⁺ iNKT cells promptly dropped, whereas that of CD1d-tetramer⁻NK1.1⁺ NKT cells remained largely unchanged (Figure S5 in Supplementary Material). By contrast, there were very few CD1d-tetramer⁺NK1.1⁺ iNKT cells but significantly increased percentage of CD1d-tetramer⁻NK1.1⁺ NKT cells in the liver of Gnb2l1^{F/F; CD4-Cre} mice in the steady state (Figure S5 in Supplementary Material). After Con A treatment, the percentage of CD1d-tetramer⁻NK1.1⁺ NKT cells further increased (Figure S5 in Supplementary Material), which might account for the equal absolute numbers of total NKT cells in the liver of Gnb2l1^{F/F; CD4-Cre} mice and control littermates.