|Home | About | Journals | Submit | Contact Us | Français|
Previously, we reported the abrogation of quiescence and reduced survival in lymphocytes from Gimap5sph/sph mice, an ENU germline mutant with a missense mutation in the GTPase of immunity-associated nucleotide binding protein 5 (Gimap5). These mice showed a progressive loss of peripheral lymphocyte populations and developed spontaneous colitis, resulting in early mortality. Here, we identify the molecular pathways that contribute to the onset of colitis in Gimap5sph/sph mice. We show that CD4+ T cells become Th1/Th17-polarized and are critically important for the development of colitis. Concomitantly, Treg cells become reduced in frequency in the peripheral tissues and their immune-suppressive capacity becomes impaired. Most importantly, these progressive changes in CD4+ T cells are associated with the loss of Foxo1, Foxo3 and Foxo4 expression. Our data establish a novel link between Gimap5 and Foxo expression and provide evidence for a regulatory mechanism that controls Foxo protein expression and may help maintain immunological tolerance.
The family of GTPase of immune-associated protein (Gimap) genes are predominantly expressed in lymphocytes and regulate lymphocyte survival during development, selection and homeostasis(1). Members of this family share a GTP-binding AIG1 homology domain, which was originally identified in disease resistance genes in higher plants (2, 3). Recent crystallographic studies revealed that GDP-bound or nucleotide-free GIMAP2 exists in a monomeric configuration with an exposed guanine nucleotide binding domain(4). In the presence of GTP, GIMAP2 oligomerizes and shows similarities with the nucleotide coordination and dimerization mode previously observed for dynamin GTPase. In addition, these studies showed that GIMAP2 localized at the surface of lipid droplets, where it is thought to act as a nucleotide regulated scaffolding protein(4). Other members of the GIMAP family appear to be localized to different subcellular compartments(5). Overall, the function of these proteins remains poorly defined.
Gimap5 was recently reported to localize in lysosomes, based on studies in human, mouse and rat lymphocytes(5). Genetic aberrancies in Gimap5 have been strongly linked to reduced lymphocyte survival and homeostasis, but, importantly have also been associated with autoimmune diseases. In humans, polymorphisms in GIMAP5 were associated with increased concentrations of IA2 auto-antibodies in type 1 diabetes (T1D) patients(6) and an increased risk of systemic lupus erythematosus (SLE)(7, 8). Studies using biobreeding (BB) rats— carrying a mutation (lyp/lyp) in Gimap5— show marked lymphopenia and predisposition to the development of T1D(9-11). In addition, BB rats are prone to develop intestinal inflammation on certain genetic backgrounds(12). Together these observations suggest that, beyond lymphocyte survival, Gimap5 is essential for maintaining immunological tolerance. Interestingly, impaired lymphocyte survival and consequent lymphopenia may be linked to the loss of immunological tolerance. Specifically, CD4+ T cells in a lymphopenic environment can undergo thymic-independent expansion in the periphery. This process—also referred to as homeostatic or lymphopenia-induced proliferation (LIP) — is accompanied by marked alterations in T cell phenotype and is linked to auto-immunity(13-15). Most notably, T cells undergoing LIP acquire a memory-like phenotype, exemplified by high surface expression of CD44 and low surface expression of CD62L. In addition, under lymphopenic conditions CD4+ T cells more readily adopt an effector phenotype, including the ability to robustly produce cytokines upon stimulation through the TCR. The downstream consequences can be severe and a number of pathological conditions have been associated with CD4+ T cells undergoing LIP, including colitis. Classic studies involving the adoptive transfer of naïve CD45RBhigh CD4+ T cells into lymphopenic SCID mice resulted in T cells acquiring a LIP phenotype and rapidly driving colitis when recipient mice were colonized by intestinal bacteria(16-18). Importantly, colitis could be prevented if CD4+ CD25+ regulatory T (Treg) cells were co-transferred, suggesting that the presence or absence of Treg cells is an important determinant of immune-mediated sequela induced by CD4+ T cells undergoing LIP.
Our laboratory previously described an ENU germline mutant, designated sphinx, which contained a recessive mutation in Gimap5 that disrupted both lymphocyte survival and normal hematopoiesis(19). Similar to Gimap5 knockout mice, these mice lack peripheral NK cells and CD8+ T cells, and exhibit dynamic changes in immune homeostasis, marked by the progressive loss of CD4+ T cells and B cells and neutrophilia(19, 20). Following the collapse of lymphocyte populations, CD4+ T cells in Gimap5sph/sph mice acquire a LIP phenotype similar to CD4+ T cells transferred into lymphopenic hosts(18). Around 10-12 weeks of age, Gimap5sph/sph mice develop wasting disease and colitis, limiting their survival(19). Interestingly, adoptive transfer of Rag-sufficient splenocytes into Gimap5sph/sph mice around 5 weeks of age could restore lymphocyte homeostasis and prevent colitis and wasting(19).
In this report, we show that CD4+ T cells are required for development of colitis in Gimap5sph/sph mice. Whereas CD4+ T cells exhibited impaired proliferation, they remained highly capable of producing proinflammatory cytokines, including IL-17A and IFNγ. Importantly, CD4+ T cells in Gimap5sph/sph mice exhibited a LIP phenotype and exhibited a progressive and complete loss of full-length Foxo1, -3 and -4 expression. This loss of Foxo expression was associated with a progressive reduction in the numbers and suppressive capacity of Foxp3+ Treg cells. The development of colitis in Gimap5sph/sph mice could be prevented by transferring wild-type Treg cells into 3-week-old Gimap5sph/sph mice. Since Foxo-deficient mice exhibit many of the phenotypes observed in Gimap5sph/sph mice, including impaired Treg cell activity and colitis, our data suggest that the loss of immunological tolerance in Gimap5-deficient mice may be critically linked to the loss of Foxo expression in CD4+ T cells.
All experiments were performed according to US National Institutes of Health guidelines and were approved by the IACUC of The Cincinnati Children’s Hospital. C57BL/6J, Rag1−/−, CD45.1 congenic and CD90.1 congenic mice were obtained from Jackson. Gimap5sph/sph mice were generated as previously described(19) and bred in the vivarium of the Cincinnati Children’s Hospital. All mice were maintained under specific pathogen-free conditions.
All antibodies used for flow cytometry were purchased from eBioscience or Biolegend. Antibodies for western blotting [anti-Foxo3a (#2497), Foxo1 (2880), pFoxo1(Thr24)/pFox03a(Thr32) (9464), p27, p-pRB (S807,S811), p-pRb (S780) and pan-Actin antibody] were purchased from Cell Signaling. Purified CD3ε (145-2C11) and CD28 (37.51) antibodies (eBioscience) were used for T cell activation. PMA and ionomycin was obtained from Sigma.
CD4+T cells were isolated from spleen and lymph nodes of 4 week-old and 7 week-old Gimap5sph/sph and C57BL/6 mice using L3T4 MicroBeads (Miltenyi Biotech). RNA isolation was done with RNeasy Micro Kit (Qiagen) and reverse transcription was performed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). cDNAs were amplified with Lightcycler 480 SYBR Green I Master (Roche) and quantified by Roche Light Cycler 480 II instrument using the following primer pairs: Foxo1: Fwd: TTCGGAATGACCTCATGGATG Rev TGGACTGCTCCTCAGTTCCTG Foxo3: Fwd; AGTGGATGGTGCGCTGTGT, Rev: TCTGAACGCGCATGAAGCFoxo4: Fwd: GAGAACCTGGAGTGCGACATG, Rev: TGTGTTGCCACCAAT
To quantify T cell proliferation, MACS (Miltenyi Biotech)-purified CD4+ T cells were labeled with CFSE (5 μM CFSE) in PBS with 0.1% FCS for 10 min. Cells were cultured in supplemented IMDM media containing 10% FCS and 1% Penicillin/Streptomycin and were either left unstimulated or stimulated with PMA(50ng/mL)/ionomycin(1μg/mL). After 3 days of incubation, proliferation was measured by analyzing CFSE dilution using flow cytometry. To assess the capacity of T cells to produce cytokines, MACS (Miltenyi Biotech) purified CD4+ T cells were incubated for 6 hours with/without PMA(10ng/mL)/ionomycin(10μg/mL) and subsequently fixed and analyzed for intracellular IL-17A, IL-4 or IFNγ production using flow cytometry. To measure surface markers ex vivo, CD4+ T cells from spleen, MLN or lamina propria were isolated and stained with fluorochrome-labeled antibodies specific for mouse CD44, CD62L and CD69. Foxp3 expression was analyzed by intracellular staining.
T cell-specific BrdU incorporation was measured as follows: during a 24-hr interval, wildtype or Gimap5sph/sph mice received three i.p. injections with 100 μL of a 10 mg/ml BrdU solution in sterile PBS. Incorporation of BrdU in CD4+ T cells was measured 8 hours after the last injection using flow cytometry.
The Treg cell suppressor assay was performed under conditions previously described(21, 22). Briefly, spleens were isolated and Treg cells were MACS purified using the CD4+CD25+ regulatory T cell isolation kit (Miltenyi). Subsequently, Treg cells were harvested and co-cultured at indicated ratios with 5 × 104 MACS purified CFSE-labeled CD8+ T cells or CD4+ T cells. Also included were 1 × 105 T cell-depleted, γ-irradiated (1,500 rad) splenocytes as bystander cells and 0.5 μg/mL soluble CD3 antibody. CFSE dilution was assessed by flow cytometry after three days of co-culture.
Colon tissue was collected and immediately fixed in 10% buffered formalin solution overnight, followed by routine paraffin embedding. Hematoxylin and eosin staining was performed on 4 μm sections from the paraffin-embedded tissue blocks for conventional light microscopy analysis. Histological scoring was performed as described before (23) Briefly, scoring parameters included quantitation of the area of distal colon involved, edema, erosion/ulceration of the epithelial monolayer, crypt loss/damage, and infiltration of immune cells into the mucosa. Severity for the area involved (erosion/ulceration and crypt loss) was graded on a scale from 0 (normal), 1 (0–10%), 2 (10–25%), 3 (25–50%), and 4 (>50%). Immune cell infiltration was scored as: 0, absent; 1, weak; 2, moderate; and 3, severe. Total disease score was expressed as the mean of all combined scores per genotype.
For adoptive transfer studies, Gimap5sph/sph mice at 25-35 days of age were injected i.v. with 3 × 105 Treg cells isolated from C57BL/6J mice using a Treg cell isolation kit (Miltenyi). Purity was confirmed by Foxp3 staining using flow cytometry and cells were >90% Foxp3+. Mice were monitored and weighed every week following cell transfer.
Data were analyzed using GraphPad Prism4® software (GraphPad Software, San Diego, CA). Unless indicated otherwise, statistical significance of the differences among groups was determined from the mean and standard deviation by Student’s two-tailed test or by ANOVA followed by Dunnett’s test for three or more groups. Data were considered statistically significant if P values were <0.05.
In previous work, we determined that NK, NKT, CD8+, CD4+ and B lymphocyte survival are impaired in Gimap5sph/sph mice. In addition, they developed spontaneous colitis that required the presence of microbiota and survived poorly, with most mice succumbing by 150 days of age(19). As several mouse models have linked impaired lymphocyte function with colitis development, we further explored the contribution of lymphocytes to the immunopathology observed in Gimap5sph/sph mice. Firstly, we investigated the survival and functional capacity of CD4+ T cells at different ages. By 4 weeks of age, a reduced number of CD4+ T cells were found in the spleen, and a further decline in T cell numbers was observed in 6- and 10-week old Gimap5sph/sph mice (Figure 1A). Six-week-old Gimap5sph/sph CD4+ T cells had a CD44highCD62Llow phenotype characteristic of T cells undergoing LIP(19) and showed increased incorporation of BrdU (Figure 1B). To assess whether the loss of CD4+ T cells in the spleen and lymph nodes was also observed in gut-associated lymphoid tissue, we isolated lamina propria (LP) cells from the colons of 6-week-old Gimap5sph/sph mice and quantified the number of CD4+ T cells. Similar to the spleen, reduced numbers of CD4+ T cells were observed in the LP (Figure 1C). Further analysis revealed that close to 100% of the colonic CD4+ T cells were CD44highCD62Llow, resembling the LIP phenotype of the CD4+ T cells in the peripheral lymphoid tissues (Figure 1C). Together these data suggest that CD4+ T cells are present in the GALT and exhibit a LIP phenotype similar to what is observed in the spleen.
We next investigated the functional capacity of CD4+ T cells in Gimap5sph/sph mice, and their potential for contributing to the development of colitis. Our previous work indicated that 8-week-old Gimap5sph/sph CD4+ T cells were unable to proliferate ex vivo following stimulation with PMA/ionomycin or αCD3, even though lymphocytes exhibited normal activation of NF-κB and MAP kinase pathways(19). Because of the latter observation, we investigated whether CD4+ T cells were capable of producing cytokines after such stimulation and, if so, were Th1-, Th2- or Th17-polarized. We isolated total lymphocytes from spleen and MLNs from C57BL/6J control or Gimap5sph/sph mice and incubated cells for 6 hours with or without PMA/ionomycin in the presence of brefeldin. Interestingly, a higher percentage of CD4+ T cells derived from Gimap5sph/sph spleen or MLNs produced IFNγ or IL-17A, or both cytokines following PMA/ionomycin stimulation (Figure 1D). Notably, T cell cytokine production was observed even in the absence of PMA/ionomycin in Gimap5sph/sph MLN cells (but not splenic leukocytes), suggesting constitutive activation of T cells in gut lymphoid tissue in these mice. Overall, these data indicate that, despite their inability to proliferate normally ex vivo, CD4+ T cells derived from Gimap5sph/sph mice become Th1/17 polarized and effectively produce cytokines.
Because of their LIP phenotype and spontaneous production of IL-17A and IFNγ, we hypothesized that MLN CD4+ T cells may support the development of colitis in Gimap5sph/sph mice. We tested this hypothesis by depleting CD4+ T cells in Gimap5sph/sph mice using weekly injections of anti-CD4 (GK1.5) antibodies, starting at 3 weeks of age—before the CD4+ T cells “collapse” and the subsequent intestinal inflammation normally occurs in Gimap5sph/sph mice. Importantly, GK1.5 treatment, but not isotype treatment, prevented wasting disease (Figure 2A) and significantly decreased intestinal inflammation as determined by histology in 15-week-old Gimap5sph/sph mice (Figure 2B-H). These data support our hypothesis that the development of colitis in Gimap5sph/sph mice requires CD4+ T cells.
Colitis induced by naïve CD45RBhigh T cell transfer into SCID recipients does not occur when Treg cells are co-transferred. Therefore, even though Treg cell development in the thymus of Gimap5sph/sph mice appeared to occur normally(19), we considered that Treg cell function may be impaired in the peripheral tissues of these lymphopenic mice and contribute to the development of colitis. Thus, we examined the presence and immunosuppressive capacity of Foxp3+ Treg cells in Gimap5sph/sph mice. Although relatively normal numbers of Foxp3+ Treg cells were observed in 3-week-old mice (Supplemental figure 2A), Treg cells became significantly reduced in the MLNs of 6- to 8-week old mice, both as a percentage within the CD4+ T cell compartment as well as in total numbers of cells (Figure 3A,B). In the spleen, the number of Treg cells were reduced but the percentage of Foxp3+ CD4+ T cells within the CD4 T cell population remained similar to the percentage observed in wildtype mice (Figure 3A,B). To assess their functional capacity, we purified Treg cells from 4- or 6-week-old C57BL/6J or Gimap5sph/sph spleens, and co-cultured Treg cells with CFSE-labeled wild-type CD8+ T cells that were stimulated with soluble anti-CD3-antibodies. Treg cells from 4-week-old Gimap5sph/sph mice showed a slight, but significant reduction in their ability to suppress CD8+ T cell proliferation in vitro, whereas Treg cells isolated from 6-week-old Gimap5sph/sph mice were incapable of suppressing CD8+ T cell proliferation (Figure 3C). Similar results were obtained for suppression of wildtype CD4+ T cell proliferation (Supplemental Figure 1A). These findings suggest that both Treg cell survival and functional capacity become impaired as Gimap5sph/sph mice age.
We next questioned whether reduced peripheral Treg cell accumulation in Gimap5sph/sph mice resulted from a cell-intrinsic phenomenon. We injected CD4+ splenocytes from 3 week-old wildtype and/or Gimap5sph/sph mice into Rag-deficient recipients, either as a mixture, or alone, and quantified the presence of Foxp3+ Treg cells 5 weeks after injection (Supplemental Figure 2A). Whereas no differences in the percentage of Treg cells within the CD4+ T cell compartment were observed at the time of injection, after 5 weeks the Gimap5sph/sph Treg population was lost, regardless of whether wild-type cells were cotransferred or not (Figure 3D). Overall, these data indicate that cell-intrinsic expression of Gimap5 is required to allow normal Treg cell survival.
Our data indicate that the Gimap5sph/sph CD4+ T cell population collapses around 5 weeks of age and that the remaining CD4+ T cells undergo LIP thereafter. At the same time, they fail to maintain a functional Treg cell population. Interestingly, these T cell phenotypes show striking similarities with those seen in mice with T cells deficient in the family of Forkheadbox group O (Foxo) transcription factors. The family of Foxo transcription factors contain 4 members of which three (Foxo1, Foxo3 and Foxo4) have overlapping patterns of expression and transcriptional activities(24-26). They play an essential role in the regulation of cell cycle progression, apoptosis, glucose metabolism and the regulation of life span(27). Foxo1 expression is critical for maintaining naïve T cell quiescence. Foxo1-deficient CD4+ T cells exhibit a CD44highCD62Llow LIP or effector memory phenotype(28-30). In addition, Foxo expression has been reported to be essential for Treg cell development and function(29, 31). We therefore analyzed expression of Foxo1, Foxo3 and Foxo4 in Gimap5sph/sph CD4+ T cells. Strikingly, immunoblot analysis of CD4+ T from 6-week-old Gimap5sph/sph mice revealed a near absence of full-length Foxo1, -3a and -4 protein (Figure 4A). At the mRNA level, a reduction in Foxo1 but not Foxo3 and 4 could be observed in CD4+ T cells isolated from Gimap5sph/sph compared to wildtype mice (Figure 4B), suggesting that regulation of Foxo3 and Foxo4 protein expression occurred at the post-transcriptional level. Since many of the T cell-specific phenotypes observed in Gimap5sph/sph occur after 4 weeks of age, we next quantified the temporal progression of changes in Foxo expression in lymphocytes. Immunoblot analyses revealed that Foxo expression was normal at 3 weeks, somewhat reduced after 4 weeks, and almost absent after 6 to 10 weeks of age (Figure 4C and supplementary Figure 1A). Concordant with the loss of Foxo expression, we detected reductions in the abundance of the cyclin-dependent kinase (Cdk) inhibitor, p27kip1, a downstream target of Foxo proteins and an important regulator of cell cycle entry (Figure 4C)(32, 33). As p27kip1 inhibits Cdk4, we measured Cdk4 activity and detected increased phosphorylation of its substrate, retinoblastoma protein (pRb), in Gimap5sph/sph cells (Figure 4C). Due to the progressive nature of this phenotype, we considered that lymphocytes isolated from young Gimap5sph/sph mice with intact Foxo expression might respond normally to mitogenic stimuli. Indeed, CD4+ T cells isolated from 4 week-old, but not 8-week-old, Gimap5sph/sph mice were able to proliferate after TCR stimulation (Supplemental Figure 3A and(19)). Finally, we assessed whether the loss of Foxo expression was also observed in regulatory T cells from Gimap5sph/sph mice. Indeed, Foxo1 and Foxo3 expression was mostly absent in CD4+CD25+ T cells from 6-week-old Gimap5sph/sph mice (Figure 4D).
Interestingly, the progressive loss of Foxo expression appeared to correlate with a progressive increase in the number of CD4+ T cells undergoing LIP (CD44hi, CD62Llo) in 4- to 10-week old Gimap5sph/sph mice as previously reported(19). Subsequent analysis of CD44lo, CD62Lhi and CD44hi, CD62Llo CD4+ T cells from 5-week-old Gimap5sph/sph and wildtype mice revealed loss of FoxO expression specifically in CD4+ T cells undergoing LIP but not naïve CD4+ T cells in Gimap5sph/sph mice (Figure 4E), suggesting that the loss of Foxo expression follows T cell activation. Overall, these data link the loss of full-length Foxo expression with the onset of lymphopenia in Gimap5sph/sph lymphocytes, the impaired cell cycle control and proliferative capacity in such lymphocytes, and the reduced Treg cell survival and function in Gimap5sph/sph mice.
Our previous data show that colitis can be prevented in Gimap5sph/sph mice through adoptive transfer of normal, but not Rag-deficient splenocytes(19), indicating that a lymphocyte population is responsible for the rescue. Given the impaired Treg cell survival and function observed in Gimap5sph/sph mice, we next examined whether adoptively transferred wild-type Treg cells could prevent the development of colitis. Gimap5sph/sph recipients of 3×105 wild-type CD4+CD25+ T cells showed prolonged survival, delayed wasting disease (Figure 5A) and, importantly, did not develop colitis (Figure 5B-D). Characterization of lymphocyte populations 25 weeks after transfer of CD45.1 congenically marked CD4+CD25+ Treg cells revealed that Treg cell reconstitution of the spleen of Gimap5sph/sph mice achieved ~50% of the level observed in wild-type C57BL/6J mice (Figure 5E). Notably, the Foxp3+CD4+ Treg cell population constituted 40% of the overall CD4+ T cell population and was entirely congenic, whereas the Foxp3−CD4+ population was predominantly Gimap5sph/sph–derived (Figure 5E). Functional analysis of isolated CD4+ T cell from spleen and MLN of 15-week-old treated Gimap5sph/sph mice revealed no background cytokine production and a similar activation as observed for wildtype CD4+ T cells following stimulation with PMA/Ionomycin (supplementary Figure 2D). Around 25 weeks of age, Treg cell-recipient mice still developed wasting disease. Necropsy at this time revealed severe inflammation in the lung and infiltration of macrophages in a number of mice (Supplemental Figure 2C). Interestingly, colitis could be prevented by the transfer of Il10−/− splenocytes (data not shown), suggesting that IL-10-independent regulatory pathways are more perturbed by Gimap5-deficienty. Together, these data link impaired Treg cell survival and function to the development of colitis in Gimap5sph/sph mice. In addition, they reveal that the Gimap5sph/sph environment is capable of supporting a functional Treg cell population.
Genetic aberrancies in Gimap5 have been linked to lymphopenia and the loss of immunological tolerance(6, 7, 9, 10). Although we found no evidence of autoimmune responses in Gimap5sph/sph mice, we observed severe and spontaneous inflammation in the gut(19)—an environment where homeostasis critically depends upon maintaining tolerance to exogenous antigens and bacterial stimuli. Similar to Gimap5sph/sph mice, loss of immunological tolerance in the lyp/lyp rat has been associated with reduced Treg cell survival and function, as well as the polarization of T helper cells towards a Th17 pattern of differentiation(34, 35). However, the molecular pathways underlying the loss of immunological tolerance in rat and mouse models of Gimap5-deficiency have remained elusive. Here, we explored the pathways that contribute to the loss of tolerance observed in Gimap5sph/sph mice. We show that the development of colitis in Gimap5sph/sph mice is critically dependent on CD4+ T cells. Around 8 weeks of age, CD4+ T cells lose their capacity to proliferate ex vivo, yet they remain capable of producing proinflammatory cytokines, including IL-17A and IFNγ, and contain a population of IL-17+IFNγ+ Th1/17 cells that have been associated with IL-23 signaling and more severe colitis(36). At the same time, Treg cell numbers and function decline. Most importantly, we found that these phenotypes are directly associated with a progressive loss of protein expression of Foxo1, -3 and -4—important transcription factors that regulate both quiescence and survival of lymphocytes(37). Mice with T-cell-specific deletions of Foxo1 and/or Foxo3 mimic many of the immunological and pathological phenotypes observed in Gimap5sph/sph mice(28-31). For example, Foxo-deficient CD4+ T cells have impaired proliferative capacity and adopt CD44highCD62Llow LIP or memory-like phenotype(28, 30). In addition, reduced Treg cell numbers and function were observed in mice lacking Foxo1 or both Foxo1 and -3 in T cells. Similar to Gimap5sph/sph mice, mice lacking Foxo1 and -3 in T cells develop spontaneous colitis, and furthermore purified Foxo1−/−Foxo3−/− Treg cells were unable to prevent colitis in Rag1−/− mice when coinjected with naïve wildtype CD4+ T cells(31). Interestingly, the impaired Treg cell development and function in T cell-specific Foxo1-deficient mice caused exaggerated T follicular helper cell accumulation, which contributed to B cell-mediated auto-immunity(29). In Gimap5sph/sph mice we found no evidence of autoreactive B cells(19), but it is important to note that, similar to CD4+ T cells, Gimap5sph/sph B cells progressively lost Foxo expression (Supplementary Figure 3B) and were unable to proliferate following stimulation with IgM(19). Thus, B cell expansion and differentiation may be severely hampered in Gimap5sph/sph mice, preventing the development of auto-reactive antibody responses.
The mechanisms by which Foxo transcription factors control Treg cell development, homeostasis and function have been studied in some detail. Foxo proteins have been shown to serve as coactivators downstream of the TGFβ signaling pathway by interacting with SMAD proteins, ultimately fine tuning the TGFβ-induced transcriptional program(38, 39). This pathway is also critical for the development of inducible Treg (iTreg) cells(40), which develop extra-thymically and have been suggested by many studies to comprise an important population of regulatory T cells in the gut(21, 41-43). Indeed, the loss of Treg cells within the CD4+ T cell compartment is most evident in the MLN of Gimap5sph/sph mice (Figure 3), suggesting the iTregs in particular are impaired. In addition, Foxo1 and Foxo3 can cooperatively control the differentiation of Foxp3+ Treg cells through the regulation of a number of Treg cell associated genes, including Foxp3 itself(31). Furthermore, conditional deletion of Foxo1 in T cells resulted in reduced surface expression of CTLA-4 and CD25 in Foxp3+CD4+ T cells(29). Analysis of the Ctla-4 gene showed that the promoter region contained a conserved Foxo binding site 193 bp upstream of the transcription start site(44). Thus, the impaired function of Foxp3+ Treg cells is likely the result of an incomplete transcriptional program in the absence of Foxo expression. In summary, the loss of Foxo expression affects multiple pathways that regulate Treg cell development, homeostasis and function, as well as the generation of iTreg cells.
In Gimap5sph/sph mice, an absence of Foxo expression was observed in all lymphocyte populations examined, including peripheral Foxp3+ Treg cells, conventional Foxp3−CD4+ T cells and B cells. Although conventional Foxp3−CD4+ T cells lack Foxo expression, our experiments reveal that colitis can be prevented by treatment with competent wild-type Treg cells, suggesting that colitogenic CD4+ T cells remain capable of being regulated when they lack Gimap5. Although we link the sphinx mutation in Gimap5 to progressive loss of Foxo expression in lymphocytes, it is unclear to what extent these genes directly interact with each other. Given the progressive nature of the loss of Foxo expression, it is unlikely that Gimap5 directly interacts with Foxo proteins. One possibility that we considered is that the loss of Foxo expression may drive a secondary phenotype resulting from the constitutive proliferation cues associated with LIP, something that may be driven by self-antigens or antigens derived from the microbiota. While we cannot exclude that LIP may contribute to the loss of Foxo expression in Gimap5sph/sph CD4+RB45high T cells, wild-type CD4+RB45high T cells transferred into a lymphopenic host retained normal levels of Foxo expression (Supplemental Figure 3C), suggesting that LIP alone is insufficient to cause loss of Foxo expression.
Our data show that Gimap5-deficiency affects Foxo3 and Foxo4 expression at the protein level, not at the mRNA level. Regulation of Foxo proteins has previously been reported to occur via ubiquitination and proteosomal degradation(45). Moreover, loss of Foxo1 expression has been observed in mouse lymphomas, which served as a mechanism to remove the tumor suppressor activity of Foxo1(46). Foxo degradation was inversely correlated with increased expression of S-phase kinase-associate protein-2 (Skp2)—an E3 ubiquitin ligase that targets numerous cell cycle proteins. Foxo degradation could be reversed following down-regulation of Skp2 via shRNAs, and therefore increased Skp2 expression could pose a potential mechanism by which loss of Foxo-expression in Gimap5sph/sph T cells occurs. Alternatively, the localization of Gimap5 in the lysosomal compartment(5) and the presumed scaffolding function of Gimap family members(4) suggests that Gimap5 may be necessary for optimal lysosomal function. Lysosomes are essential for the catabolic turnover of intra- and –extracellular macromolecules, but also can release lysosomal enzymes (such as cathepsins) that can initiate programmed cell death once in the cytosol(47). Intriguingly, lymphocytes containing large numbers of cytotoxic granules, such as CD8+ T cells and NK cells do not survive in Gimap5- deficient mice, perhaps supporting the hypothesis of a deregulated lysosomal compartment. Although the link between loss of Gimap5 and Foxo protein expression remains to be established in human cells, understanding the molecular pathways that lead to the degradation of Foxo proteins could provide important therapeutic targets, not only in the context of tumor growth, but potentially in the context of autoimmune or chronic inflammatory disorders.
Our data provide evidence that Gimap5 is essential for maintaining lymphocyte quiescence and immunological tolerance. In the absence of functional Gimap5, Foxo expression in lymphocytes is progressively lost, with loss of Foxo3 and Foxo4 most likely involving a proteolytic mechanism. This progressive loss of Foxo expression is associated with concomitant decline in Treg numbers and function, which ultimately leads to the loss of immunological tolerance in the gut. Thus, not only do we establish a critical link between Gimap5 and Foxo protein levels, we also provide evidence for a novel regulatory mechanism controlling Foxo protein expression that may be involved in the development of immune-mediated diseases such as SLE, T1D and colitis.
This research was funded by grants from the NIH, including NIH/NCI R21 Grant 1R21CA133649, NIH/NIAID RO1 Grant 1R01AI074743 and PHS Grant P30 DK078392 (Integrative Morphology Core of the Cincinnati Digestive Disease Research Core Center).