|Home | About | Journals | Submit | Contact Us | Français|
Caspase recruitment domain-containing membrane-associated guanylate kinase protein-1 (CARMA1) is a critical component of the nuclear factor κB (NF-κB) signaling cascade mediated by T cell receptor (TCR) engagement. In addition to activation of naïve T cells, TCR signaling is important for the development of agonist-selected T cell subsets such as regulatory T cells (Treg), natural killer T cells (NKT), and CD8ααT cells. However, little is known about the role of CARMA1 in the development of these lineages. Here we show that CARMA1-deficient mice (CARMA1−/−) have altered populations of specific subsets of agonist-selected T cells. Specifically, CARMA1−/− mice have impaired natural and adaptive Treg development, while NKT cell numbers are normal compared to wild-type mice. Interestingly, CD8αα T cells, which may also be able to develop through an extrathymic selection pathway, are enriched in the gut of CARMA1−/− mice, while memory-phenotype CD4+ T cells (CD62Llow/CD44high) are present at reduced numbers in the periphery. These results indicate that CARMA1 is essential for Treg development, but is not necessary for the development of other agonist-selected T cell subsets. Overall, these data reveal an important but differential role for CARMA1-mediated TCR signaling in T cell development.
T cells develop from a common bone marrow progenitor into diverse populations through a regulated selection process in the thymus . While most T cells emerge from the thymus as CD4+ or CD8+ naïve T cells, some T cells adopt special nonconventional lineages with different functions in the periphery. In the majority of thymocytes, T cell receptor (TCR) αβ signaling during T cell development selects T cells via two different interactions: 1) positive selection of thymocytes with TCR that recognize major histocompatibility complex (MHC) proteins (with resulting death of non-interacting cells by neglect), and 2) negative selection of T cells with TCR that interact strongly with agonist self antigens displayed on MHC proteins on the thymic epithelial cells, thus deleting autoreactive T cells. However, it is now clear that some thymocytes are not deleted even though they have a strong interaction with antigens displayed on MHC proteins on the thymic epithelial cells. This non-deletional form of TCR-mediated positive selection of thymocytes can induce the development of several alternative lineages of T cells (so-called agonist-selection), including forkhead box p3 (Foxp3) + regulatory T cells (Treg), natural killer T cells (NKT), and TCRαβ CD8αα T cells . Thus, the development of these T cell subsets would appear to be highly dependent on the TCR signalosome and its components.
CARMA1, a member of the membrane-associated guanylate kinase family of kinases, encodes a protein that contains a CARD domain, coiled-coil domain, PDZ domain, Src homology 3 domain and C-terminal guanylate kinase domain . CARMA1 is specifically expressed in lymphocytes and is essential for lymphocyte activation via TCR or B cell receptor signaling [4–7]. Upon TCR engagement, CARMA1 is phosphorylated by protein kinase C θ (PKCθ) [8, 9] and recruited to the immune synapse [10–12] where it forms a complex with the proteins B-cell CLL/lymphoma 10 (Bcl10) and mucosa-associated lymphoid tissue 1 (MALT1). This complex then initiates a signaling cascade that ultimately leads to the activation of the T cell, in part via the inhibitory NF-κB (IκB)kinase complex (IKK) and NF-κB activation [6, 7, 13, 14], and possibly through other signaling pathways . Recently, we have demonstrated that CARMA1 expression is essential for the development of allergic airway inflammation in a murine model of asthma . Given its role in TCR signaling and the potential for a therapeutic role for inhibition of CARMA1, we were interested if CARMA1 participated in non-deletional agonist selection of T cell subsets. Here we demonstrate that CARMA1-deficient (CARMA1−/−) mice have a complete defect in Treg development, but not in other agonist selected T cell subsets. These data suggest an important but differential requirement for CARMA1-mediated TCR signaling in T cell development.
Previous work by others has demonstrated that CARMA1−/− mice have normal T cell development and normal peripheral T cell numbers and ratios [6, 13], although these studies did not specifically look at Treg populations. It is commonly accepted that natural Treg develop in the thymus , although the mechanisms that induce their development remain controversial. Monoclonal TCR transgenic studies suggest that thymocyte encounters with self antigen are necessary for Treg development [18–20], and more recent studies have demonstrated that Foxp3 expression in CD4+ thymocytes is dependent on TCR signaling and cytokine stimulation [21, 22]. In addition, multiple components of TCR induced NF-κB signaling have been shown to be necessary for formation of Treg. Specifically, deficiency in PKCθ , IKK2 [23, 24], Bcl10 , and p50/cRel [25, 26] have all been shown to impair Treg development. These data suggest that CARMA1 might also be necessary for Treg development. Consistent with these data, we could not identify significant numbers of Treg (as measured by CD4+/Foxp3+ staining) in the lymph nodes, lung or spleen of naive CARMA1−/− mice (Figure 1A and Table 1). We also evaluated the presence of CD4+/Foxp3+ T cells in the thymus of these mice and CARMA1−/− mice did not have CD4+/Foxp3+ T cells in the thymus (Figure 1B and Table 1) consistent with a developmental defect in the formation of these cells. Interestingly, despite an absence of Tregs, CARMA1−/− mice do not develop clinically obvious autoimmune disease. We suspect that the T cell activation defect in CARMA1−/− mice  protects these mice from T cell mediated autoimmune diseases.
Two types of Foxp3+ regulatory T cells have been defined: those that develop in the thymus (natural Treg) and those that are induced from naive T cells in the periphery in the setting of inflammation (adaptive Treg). To determine if CARMA1−/− mice can produce adaptive Treg, we induced allergic airway inflammation in these mice by adoptive transfer of OVA-specific T cells followed by OVA challenge as previously described . As seen by others, after induction of inflammation, wild-type mice accumulate Treg in the lung and draining lymph nodes . However, in the CARMA1−/− mice there were almost no CD4+/Foxp3+ T cells in the lungs and thoracic lymph nodes (Figure 1C), consistent with a defect in generating adaptive Treg.
Tregs are one of several nonconventional T cell subtypes that have been shown to undergo non-deletional agonist-selection . These include CD1d-restricted Vα14+ invariant NKT cells, which are a subset of T cells that express a TCR and certain NK cell receptors, and recognize glycolipids bound to the nonclassical MHC class I-like molecule CD1d. In mice, these cells are derived from double-positive thymocytes, and are selected by self-lipids presented on CD1d in the thymus [2, 28]. NKT cell development is dependent on IKK2 expression and RelB expression [23, 29, 30], suggesting that TCR induced NF-κB activation is critical for NKT cell development. However, CARMA1−/− mice have a similar percentage of NKT cells (identified by a tetramer loaded with an analogue of the αGalCer ligand, PBS57 ) in the lung and spleen compared to wild-type mice (Figure 2A and Table 1). Consistent with this, there was also no difference in the absolute number of NKT cells in the lung and spleen of wild-type and CARMA1−/− mice (lung: 5 ± 1 × 104 cells vs. 11 ± 3 × 104 cells, wild-type vs. CARMA1−/− mice respectively; spleen: 4.1 ± 0.5 × 105 cells vs. 5.3 ± 0.9 × 105 cells, wild-type vs. CARMA1−/− mice respectively). We also evaluated the presence of NKT cells in the thymus of these mice and did not detect any difference in the percentage (Table 1) or absolute cell numbers of NKT cells in wild-type and CARMA1−/− mice (2.4 ± 0.4 × 105 cells vs. 1.5 ± 0.4 × 105 cells, wild-type vs. CARMA1−/− mice respectively). These results are similar to data reported in previous work in PKCθ−/− mice, but different from results in Bcl10−/− mice and IKK2−/− mice [6, 23]. These data suggest that there may be compensatory mechanisms or redundant pathways for proximal elements of the TCR cascade that preserve enough signaling to stimulate NKT cell development in CARMA1−/− or PKCθ−/− mice, but not enough for Treg development. Downstream components such as IKK2 or Bcl10, however, remain essential for both NKT cell and Treg development.
TCRαβ CD8αα T cells are a unique T cell subset that is most abundant in the intestine among the intraepithelial lymphocyte population. The cells express an activated phenotype and seem to have an innate-immune signature [32–34]. Studies have suggested that these cells may develop either with or without a process of thymic selection [33, 35–37]. Thymic selection may occur via agonist-selection of cells with TCR against self-peptides similar to Treg and NKT cells [38–40]. In the periphery, effector memory CD8αβ T cells that migrate to the intestine may be induced to express CD8αα . Thus, TCR signaling seems to be important for CD8αα T cell development and we would expect that CARMA1 may have a role in this process. However, we found an increase in the CD8αα/CD8αβ T cell ratio in the intraepithelial lymphocyte population in the small intestine of CARMA1−/− mice compared to wild-type control mice (Figure 2B and Table 1). This reflected a 3-fold increase in the absolute number of these cells in the intestines (0.69 ± 0.15 × 106 cells vs. 2.27 ± 0.61 × 106 cells, wild-type vs. CARMA1−/− mice respectively, p = 0.04). The mechanism for this finding is unknown, but it is possible that the reduction in TCR signaling in CARMA1−/− T cells may shunt T cells destined to be Tregs into other agonist selected T cell populations such as CD8αα T cells. Alternatively, CARMA1 may negatively regulate CD8αα T cell development.
Normal naive mice contain a small but significant population of circulating spontaneously arising memory-phenotype T cells (defined by low expression of CD62L and high expression of CD44), although whether these cells are representative of true antigen-induced memory T cells is uncertain [41, 42]. It is felt that some of these cells may develop after exposure to various environmental antigens, however, recent data suggest that at least some of these memory-phenotype T cells can be generated during thymic development (so called innate memory T cells) . The importance of TCR signaling in the development of these cells is unclear, but TCR engagement is critical for antigen-induced memory T cell formation . Studies on IKK2−/− mice, p50−/−/cRel−/− mice, PKCθ−/− mice, and Bcl10−/− mice have demonstrated that these mice all have reduced numbers of peripheral memory-phenotype T cells [23, 25]. These data suggest that TCR-induced NF-κB activation is necessary for the formation of some of these T cells and that CARMA1 may also have a role in their formation. Consistent with this, CARMA1−/− mice have a reduced percentage of CD62LlowCD44high CD4+ T cells in the spleen compared to wild-type control mice (Figure 2C and Table 1). This reflects an over 2-fold reduction in the absolute number of memory-phenotype T cells in the spleen of CARMA1−/− mice compared to wild-type mice (1.75 ± 0.35 × 106 cells vs. 0.77 ± 0.15× 106 cells, wild-type vs. CARMA1−/− mice respectively, p=0.03). Our data therefore confirm the findings in other TCR signaling protein deficient mice. What drives the development of the remaining memory-phenotype T cells is unclear.
In summary we demonstrate that CARMA1 has a critical role in the development of Tregs, but is not necessary for the development of NKT cells and CD8αα T cells, while memory-phenotype T cell development is only partially impaired. Interestingly, the percentage of CD8αα T cells in the intestines is increased in the CARMA1−/− mice. These data partially resemble findings in other mouse strains deficient in proteins involved in TCR induced NF-κB signaling; however there are some important differences in the CARMA1−/− mice compared to mice deficient in downstream signaling mediators such as IKK2, p50, Bcl10, and RelB. These data suggest a differential role for TCR signaling proteins in nonconventional T cell subset development. We can only speculate on the mechanisms behind our findings. We know that TCR signaling is complex and involves multiple proteins and pathways which may be modulated by the strength of the antigen-TCR interaction. A recent paper has demonstrated that signaling molecules can have qualitatively distinct effects on cell responsiveness and that heterogeneity in protein levels can alter responsiveness . Based on this, we suspect that Tregs, NKT cells, CD8αα cells, and memory-phenotype T cells all require a different amount of TCR signaling during development. One could surmise that very strong antigen-TCR interactions are needed for Treg development, and must utilize CARMA1 to initiate high-level cellular responses. However, the formation of other subclasses does not require this level of signaling. Deficiency in Bcl10 or IKK2 may have more dramatic effects on TCR signaling which then leads to the broader array of T cell subset deficiency. An alternative explanation for the differences between the changes seen in CARMA1−/− mice and other TCR signaling protein deficient mice may be interactions of CARMA1 with other TCR-induced signaling pathways or related proteins [15, 44–46]. We also cannot completely rule out an effect from CARMA1 in non-TCR-induced signaling pathways [47, 48], although these have not been well described. Overall, our results contribute to the growing body of data suggesting a complex role for TCR signaling in the development of T cell subsets [1, 3]. In addition, our data may have implications for therapeutic interventions that target CARMA1 or other proteins involved in TCR signaling.
Mice with a deletion in the CARMA1 gene were generated using modified bacterial artificial chromosome (BAC) technology as previously described . Littermate wild-type mice were used as controls in all experiments. All protocols were approved by the MGH Subcommittee on Research Animal Care – OLAW Assurance # A3596–01, protocol #2003N000336 to Andrew D. Luster, MD, PhD.
Allergic airway inflammation was induced in mice as previously described [49, 50]. Briefly, naive wild-type or CARMA1−/− mice received intravenous transfer of 5 × 106 in vitro–differentiated OVA-specific Th2 cells obtained from the OT-II transgenic mouse strain. 24 hours after injection, mice underwent aerosol challenge with OVA (50 mg/mL in PBS) daily for 4 days. Mice were sacrificed 20–24 h after the last aerosol challenge.
Mice were euthanized by ketamine injection (100 mg/kg) and the thymus, spleen, lymph nodes, lungs, and small intestine were removed. Single cell suspensions were made from the recovered tissues and cells were blocked and then stained with fluorescently labeled antibodies to the indicated antigens (BD Biosciences, San Jose, CA). Isolation of intraepithelial lymphocytes from the intestines was performed as previously described . Foxp3+ cells were identified using intracytoplasmic staining with an antibody against Foxp3 according to the manufactures’ protocol (eBiosciences, San Diego, CA). NKT cells were identified with a tetramer loaded with an analogue of the αGalCer ligand, PBS57 that was developed in Dr. Paul B. Savage’s laboratory at Brigham Young University  (NIH Tetramer Core Facility, Atlanta. GA). Flow cytometry was performed after gating on the lymphocyte population utilizing a FACS Caliber analytical flow cytometer (BD Biosciences) and analyzed using Flowjo software (Treestar, Ashland, OR).
Data are expressed as mean ± SEM. Differences in results were considered to be statistically significant when p<0.05 using a student’s t test.
This work was supported by grants from the National Institutes of Health (HL088297 to BDM and RJX, DK64351 to AM, CCIB and DK043351 to RJX)
Conflict of interest: The authors declare no financial or commercial conflict of interest.