Role of diacylglycerol kinases in T cell development and function

I. INTRODUCTION

Diacylglycerol kinases (DGKs) are a family of enzymes that catalyze the conversion of lipid second messenger diacylglycerol (DAG) to phosphatidic acid (PA). Work from several groups, including ours, has shown that DGKs serve as a braking mechanism in immune cell signaling, dampening DAG levels after receptor stimulation and preventing hyper-activation of immune cells.^1–4 Ten isoforms of DGK have been identified in mammals, many of which are expressed in cells of the immune system.

Notably, both the substrate and product of the DGK-catalyzed reaction, DAG and PA, are bioactive lipids that can act as second messengers.^5–8 DGK activity therefore serves as a switch to simultaneously dampen DAG-mediated signals and boost PA-mediated signals. In T cells, DAG recruits RasGRP1 and PKCθ to the cell membrane, leading to signaling via the RasGRP1/Ras/ERK and PKCθ/IKK/NF-κB pathways.^9,10 Previous studies have shown that PA, on the other hand, can bind to signaling molecules such as mammalian target of rapamycin (mTOR), SHP-1, RasGAP, Sos, PI5Kα, and p47(phox).^8,11–17

All mammalian DGKs contain a catalytic kinase domain, consisting of a conserved motif and an accessory domain, and at least two cysteine-rich DAG-binding C1 domains. However, DGKs also possess other distinct structural domains, based on which they are classified into five types (Figure 1). Two DGK isoforms, the type-I α isoform and the type-IV ζ isoform, are highly expressed in T cells.^18,19 The type-I DGK isoforms α, β and γ possess an N-terminal recoverin homology domain and two Ca²⁺-binding EF hand motifs. While the recoverin homology domain is related to the N terminal region of the recoverin family of neuronal calcium sensors, the EF hands are involved in auto-inhibition.²⁰ Type-IV DGK isoforms, ζ and ι, contain a myristoylated alanine rich C-kinase substrate (MARCKS) motif, four ankyrin repeats, and a C-terminal PDZ binding domain. As its name suggests, the MARCKS domain can be phosphorylated by PKC isoforms. Studies in cell lines have shown that PKCα can phosphorylate the DGKζ MARCKS motif to negatively regulate both DGKζ’s catalytic activity and its ability to interact with other proteins.^21,22 In addition, the MARCKS motif contains a nuclear localization sequence.²³ The ankyrin repeats and PDZ binding motif are thought to play a role in protein-protein interactions, with the latter binding to PDZ domains on proteins such as syntrophins.^24,25 Similar to some other DGKp isoforms such as β, δ, and η, DGKζ contains several alternative splicing isoforms.^26,27 The functional differences among these DGKζ isoforms in the immune system remain to be clearly defined.

Members of the DGK family show substantial diversity in the cell types they are expressed in and their localization within those cells. Notably, DGK isoforms are highly expressed in cells of the hematopoietic and nervous systems. Multiple DGK isoforms are often expressed simultaneously in a given cell type. Though DGK α and ζ isoforms predominate in T cells, we can be detected multiple DGK isoforms by reverse transcriptase-PCR in T cells (α, γ, δ, ζ, θ), macrophages (α, β, γ, δ, ζ, ι) and mast cells (α, γ, δ, ε, ζ, ι)(References 28, 29 and our unpublished observations). Due to their distinct structural domains, different types of DGKs tend to localize to specific subcellular compartments and are regulated by unique cues in the intracellular milieu.³⁰ Six DGK isoforms – α, γ, δ, ζ, ι, and θ - have been observed to reside in or move to the nucleus upon stimulation in different cell types.³¹ For instance, stimulation via the TCR leads to the nuclear translocation of DGKα and its binding to the nuclear matrix in primary rat T cells.³² Immunohistochemical analyses have also revealed that DGKζ localizes to the nucleus in neurons in various parts of the rat brain, and that this subcellular distribution is specifically disrupted in hippocampal pyramidal neurons in a model of forebrain ischemia.³³ Nuclear localization of DGKζ has also been demonstrated in other cell types.^23,24 Nuclear translocation of DGKs could control nuclear DAG and PA concentrations and/or prevent DGKs from terminating DAG in the cytoplasm membrane. Whether DGKζ localizes to the nucleus and the functional importance of nuclear localization of DGK isoforms in T cells remain to be defined.

From an organismal standpoint, it is interesting to note that DGKα protein expression appears to be restricted to certain cell types such as T-lineage cells and oligodendrocytes,^34,35 while DGKζ is expressed more ubiquitously in the brain, lungs, heart, hematopoietic system and skeletal muscles.^27,36,37 However, as stated previously, our unpublished data suggests that DGKα, at least at the mRNA level, may also be expressed in other hematopoietic cells including mast cells and macrophages. In this review, we discuss the varied roles played by DGKs in T cell development and function, while emphasizing recent advances that have helped move the field forward.

II. ROLE AND REGULATION OF DAG IN TCR SIGNALING

Engagement of the TCR by a cognate peptide-MHC complex triggers a multitude of signaling pathways that, in the presence of additional signals, cooperate to turn on a transcriptional program of T cell activation (Figure 2). Following TCR engagement, Lck, a Src family tyrosine kinase that associates with the cytoplasmic tails of CD4 and CD8, is activated by CD45-mediated dephosphorylation.^{38, 39} Active Lck, in turn, phosphorylates immunoreceptor tyrosine-based activation motifs (ITAMs) on CD3, leading to the recruitment of the kinase ZAP-70.^{40, 41} Lck phosphorylates and activates ZAP-70, which then phosphorylates the adaptor protein LAT.^{42, 43} Phosphorylated LAT recruits a number of signaling molecules, including adaptor SLP-76 and phospholipase PLCγ1, to the cell membrane.^44,45

Active PLCγ1 hydrolyzes membrane phospholipid PIP₂ to produce two second messengers ^- DAG and inositol triphosphate (IP₃). IP₃ binds to its receptors on the endoplasmic reticulum (ER) to trigger the release of intracellular ER calcium stores. Depletion of calcium in the ER lumen causes a conformational change in the ER-associated calcium sensors STIM1 and STIM2.⁴⁶ This conformational change activates the calcium-release activated calcium (CRAC) channel Orai1 on the cell membrane, leading to an influx of calcium from the extracellular milieu.⁴⁷ Increasing calcium levels in the cytosol activate the phosphatase calcineurin, which dephosphorylates the transcription factor NFAT to trigger its nuclear translocation.⁴⁸ Signaling via the NFAT pathway is important for T cell development, activation, anergy, and in the function of T_reg, T_FH, T_H17, and CD8 T cells.^49,50

Membrane-associated DAG, on the other hand, recruits PKCθ, PKD1, and Ras family guanine nucleotide exchange factor RasGRP1 to the cell membrane through their C1 domains.^10,51–53 Activated RasGRP1, along with Sos, helps convert the small GTPase Ras from its GDP-bound inactive form to a GTP-bound active form, and active Ras then activates the kinase Raf.^{54, 55} In turn, Raf activates MEK1/2, which subsequently activate MAP kinases ERK1/2. ERK activity increases the expression of transcription factor c-Fos and also phosphorylates c-Fos, leading to its dimerization with c-Jun to form the transcription factor AP-1.^{56, 57} NFAT and AP-1 interact with each other and bind cooperatively to composite binding elements on the promoters of several genes, including IL-2.^58–60 The RasGRP1-Ras-ERK1/2 pathway has been shown to play a critical role in positive selection during intrathymic T cell development and in the activation of peripheral T cells.^61–64

Active PKCθ phosphorylates the adaptor CARD11, leading to the formation of a signalosome with Bcl10 and MALT1.⁶⁵ This signalosome activates the three-subunit IKK complex, which phosphorylates IκB. Phosphorylation of IκB eventually leads to its degradation, allowing active NF-κB dimers to translocate to the nucleus.^{66, 67} The PKCθ-IKK-NFκB pathway is not essential for conventional T cell maturation, but is critical for iNKT cell and regulatory T cell development and for peripheral T cell activation,^68–71 particularly for effective T_H2 immune responses.^72–74 It is important to note here that DAG-independent mechanisms may also play important roles in PKCθ activation. For instance, CD28 can directly recruits PKCθ to the immunological synapse to promote its activation.^75,76

DAG-activated PKD1 phosphorylates the transcriptional repressor HDAC7, leading to its export from the nucleus and the de-repression of its target genes, such as Nur77.^77,78 Membrane-localized and cytosolic forms of PKD1 may have distinct functions during thymocyte development.⁷⁹ The individual DAG-mediated pathways outlined above have also been shown to cooperate with each other. For example, PKCθ phosphorylates RasGRP1 and PKD1 to enhance their activation.⁸⁰ Thus, by recruiting PKCθ^, RasGRP1, and PKD1, DAG controls signaling via a number of interconnected pathways in response to TCR engagement.

DAG-mediated activation of the Ras-ERK1/2 and NFκB pathways also indirectly activates signaling via the PI3K-mTOR pathway. The mTOR pathway and its tight regulation play an important role in T cell development, homeostasis, activation, and differentiation,^81–88 while PI3K isoforms function redundantly to promote T cell development, activation, survival, and self-tolerance.^{89, 90} Studies in cell lines have shown that active Ras can bind to and activate PI3K,^{91, 92} which catalyzes the conversion of PIP₂ to PIP₃. While it was known that PI3K activity can recruit the kinases PDK1 and Akt to activate the Akt-mTOR complex 1 pathway,⁹³ recent studies using primary mouse thymocytes and T cells have provided direct evidence that DAG-mediated activation of the Ras-MEK1/2-ERK1/2 pathway can also initiate signaling through mTOR complex 1 and mTOR complex 2.^94,95

Given the nature and number of signaling pathways activated by DAG, it stands to reason that DAG levels must be tightly regulated to prevent T cell hyper-activation. This hypothesis is supported by findings that dysregulation of individual DAG-mediated signaling pathways can have profound effects on T cell function and iNKT cell development.^{71,96, 97} DGKs play a critical role in regulating intracellular DAG levels, removing DAG through phosphorylation to produce PA. DGKα and DGKζ are the predominant isoforms expressed in T cells, as mentioned previously. Two splice variants of DGKζ, DGKζ1 (130 kDa) and DGKζ2 (115 kDa), are expressed in thymocytes and matureT cells.¹⁹ Splicing of the first coding exon directly to the third generates the smaller splice variant, while transcription from an alternative promoter at the second exon and subsequent splicing to the third generates the longer variant.^26,98 Interestingly, the splicing isoforms are expressed in a complementary fashion, with higher expression of DGKζ1 in CD4⁻ CD8⁻ and CD4+ CD8+ thymocytes, and higher expression of DGKζ2 in mature (CD4SP and CD8SP) thymocytes and peripheral T cells. The mechanisms that control the differential expression of DGKζ1/ζ2 during T cell development are currently unclear, as are the functional differences between these two isoforms. Experiments with deletion mutants in Jurkat cells showed that the N-terminal end of DGKζ, but not the C-terminal end, is essential for optimal inhibition of TCR signaling.¹⁹ Other studies have shown that DGKα is expressed in thymocytes and peripheral T cells,^{34, 99,100} but its expression levels at different stages of T cell development remain to be examined. As discussed below, synergistic regulation of DAG signaling by DGKα and DGKζ is essential for normal T cell development and function.

III. ROLE OF DGKα AND DGKζ IN T CELL DEVELOPMENT

Lymphoid progenitor cells generated in the bone marrow migrate to the thymus, where they travel through the cortex and medulla, developing into mature T cells.^{101, 102} Successive developmental stages of a thymocyte can be distinguished by the combination of CD4 and CD8 co-receptors expressed on its surface. Early committed T cells do not express TCR, CD4, or CD8 on the cell surface and are called CD4⁻ CD8⁻ double-negative (DN) cells. DN cells rearrange V, D, and J gene segments at the TCRβ locus, leading to the expression of a pre-TCR. Cells that express a functional pre-TCR pass through the so-called “β-selection” checkpoint, while others undergo apoptosis. The β-selected DN cells undergo several rounds of proliferation, maturing into CD4⁺ CD8⁺ double-positive (DP) cells that constitute about 90 percent of all thymocytes. DP cells rearrange V and J gene segments at the TCRa locus, leading to the expression of a unique TCR on the cell surface.

Following TCR expression, DP thymocytes are subjected to processes called positive and negative selection,^103,104 that ensure the generation of a functional, non self-reactive T cell repertoire. In order to be positively selected, DP cells must express a TCR that is able to recognize self-peptide-MHC complexes expressed by thymic epithelial cells or bone-marrow-derived dendritic cells in the thymus. In general, DP cells with TCRs that fail to recognize self-peptide-MHC complexes are eliminated at this stage, as they fail to receive survival signals. On the other hand, DP cells with TCRs that recognize self-peptide-MHC with high affinity also undergo apoptosis, a process referred to as negative selection. Thus, only DP cells with TCRs that recognize self-peptide-MHC molecules with low affinity survive positive and negative selection, developing further into mature CD4⁺ CD8⁻ single positive (CD4SP) or CD4⁻ CD8⁺ single positive (CD8SP) cells.

Several signaling pathways, including MAP kinase, NF-κB, and NFAT pathways, are known to play critical roles in thymocyte selection. PLCγ1 deficiency in thymocytes impairs both positive and negative selection processes, suggesting a potential role for DAG-mediated signals in T cell development.¹⁰⁵ Numerous studies have shown that defects in DAG-effector pathways profoundly impact thymocyte development, lending further credence to this notion. For instance, thymocytes deficient in RasGRP1 display severely impaired positive selection, with a marked paucity of mature single positive cells¹⁰⁶ Expression of a dominant negative form of Ras or MEK1 inhibits positive but not negative selection,^{107, 108} leading to a block in thymocyte development at the DP stage. While thymocytes lacking ERK1 experience a partial developmental block at the DP stage with a concomitant reduction of mature thymocytes,¹⁰⁹ combined deficiency of ERK1 and ERK2 has been shown to impair positive but not negative selection.^110,111 The Ras-MEK1/2-ERK1/2 pathway is thus thought to play a critical role in the positive selection of thymocytes. On the other hand, signaling via the p38 and JNK MAP kinase pathways is thought to play an essential role in negative selection.¹¹² Though deficiency of PKCθ or IKKβ does not appear to affect conventional αβ T cell maturation,^{70, 113} a recent study has revealed a differential role for NF-κB in the selection and survival of CD4 and CD8 thymocytes.¹¹⁴ Moreover, PKCθ-mediated signaling is pivotal for natural regulatory T cell and iNKT cell development.¹¹⁵ The importance of DAG-triggered Ras-ERK and PKCθ-NF-κB pathways in thymocyte selection processes thus suggests that tight regulation of DAG levels by DGKs may be critical for normal T cell development.

Previous studies have shown that signaling via the pre-TCR increases DGKα expression in thymocytes.¹¹⁶ Although pharmacological inhibition of DGKα activity (with the inhibitor R59949) suggested that DGKα could promote DP thymocyte survival via a Bcl-xL mediated pathway,¹¹⁶ other studies have revealed that genetic deficiency of either DGKα or ζ does not obviously alter thymocyte populations.^117,118 Additional studies should determine whether this type-I DGKα inhibitor may possess off-target activities that are yet to be identified or if other type-I DGKs are expressed in developing thymocytes that may compensate for DGKα deficiency. More recent work from our group has provided genetic evidence that DGKα and DGKζ synergistically regulate T cell development.¹¹⁹ Combined deficiency of the DGKα and ζ isoforms led to a severe block in murine thymocyte development at the DP stage, with a dramatic reduction in the number of mature CD4SP and CD8SP cells. Crossing with HY TCR transgenic mice revealed that combined DGKαζ deficiency was associated with impaired positive selection, but not negative selection. Reduced DGK activity in the DGKαζ double knockout (DKO) thymocytes was associated with increased levels of intracellular DAG after TCR stimulation, and enhanced signaling via the Ras-ERK pathway. However, the developmental blockade was partially overcome by PA treatment, suggesting that DGKs play a critical role in thymocyte development not only by terminating DAG-mediated signaling but also by initiating PA-mediated signals. How PA promotes T cell maturation is a critical question that remains to be addressed.

A novel role for DGKs in regulating mTOR activity in thymocytes has also emerged recently.⁹⁴ Upon TCR stimulation, DGKαζ DKO thymocytes showed elevated levels of S6K1 and 4E-BP1 phosphorylation, suggestive of increased mTOR complex 1 activity. This was correlated with increased ERK1/2 and Rsk1 activation. Phosphorylation of Akt at S473 was also increased in DKO thymocytes, indicating enhanced mTOR complex 2 activity. Inhibition of MEK1/2 dramatically reduced TCR-induced mTOR activation in both wild-type (WT) and DGKαζ DKO thymocytes, indicating that DGK activity inhibits TCR-induced mTOR activation by attenuating signaling via the RasGRP1-Ras-ERK1/2 pathway. In non-T cell lines, ERK1/2 phosphorylate and inactivate TSC2, a negative regulator of mTOR complex 1 activation, to promote signaling via mTOR complex 1.¹²⁰ Whether the RasGRP1-Ras-ERK1/2 pathway activates mTOR complex 1 signaling via similar mechanisms in T cells needs to be confirmed. In addition, the mechanisms by which DGK (α and ζ) activity inhibits and RasGRP1-Ras-ERK1/2 signaling promotes mTORC2 signaling remain to be explored.

IV. ROLE OF DGKα AND DGKζ IN iNKT CELL DEVELOPMENT

Natural killer T (NKT) cells are a rare subset of lymphocytes that express both NK family receptors (such as NK1.1 in mice) and a semi-invariant TCR.^121–123 Unlike conventional αβ T cells, the TCR on NKT cells recognizes glycolipids presented on MHC-like CD1d molecules. Capable of producing an array of cytokines within minutes to hours of stimulation, NKT cells have been shown to modulate several important immune phenomena including responses to infection and cancer, allergy, and autoimmunity. A majority of NKT cells in humans and mice are characterized by their unique usage of TCR Vα24 (human) or Vα14 (murine) and Jα18 segments and a limited TCR Vβ repertoire. Such NKT cells are called type I or invariant NKT (iNKT) cells. Developing iNKT cells are classified into successive developmental stages from 0 to 3, based on the surface expression of CD24, NK1.1, and CD44¹²⁴ Invariant NKT cell development and function remain actively investigated.

In the thymus, DP cells that express appropriate TCRs to enter the NKT lineage are thought to undergo selection processes analogous to those of conventional αβ T cells, which require signaling from the iVα14TCR.^{125, 126} However, unlike conventional αβ thymocytes that are selected on thymic epithelial cells, NKT thymocytes are selected on fellow CD1d-expressing DP thymocytes.^127–129 Results from previous studies have also suggested that developing iNKT cells may differ from αβ thymocytes in certain signaling requirements for proper development. For example, a SLAM/SAP/Fyn/PKCθsignaling pathway is critical for iNKT cell ontogeny but exerts minimal impact on conventional αβT cell development.^130–134 Several recent reports have demonstrated that DAG-mediated signaling and its proper regulation are pivotal for normal iNKT cell development and homeostasis. Absence of RasGRP1 or expression of a dominant negative Ras impairs iNKT cell development at the earliest stage.^97,135 In contrast, hyper-activation of this pathway by the expression of constitutively active K-Ras in thymocytes causes defective iNKT cell terminal maturation, correlating with decreased T-bet expression.⁷¹ Similarly, absence of PKCθ impairs iNKT cell development and overactive IKKβ severely reduces iNKT cell numbers.^{69, 71} Recent studies have also provided genetic evidence that individual deficiency of DGKα or DGKζ does not significantly affect iNKT development. However, simultaneous deficiency of both isoforms led to a severe iNKT cell-intrinsic developmental blockade/homeostasis defect and a concomitant paucity of iNKT cells in the thymus, spleen and liver. In DGKαζDKO thymocytes, both Ras-ERK1/2 and PKCθ-IKK signaling are elevated. These observations have not only revealed the importance of DAG-effector signaling pathways in iNKT cell development but also elucidated the requirement of DGKαζ activity for normal iNKT cell ontogeny via tight control of these pathways. It remains unclear whether DGKα and ζ promote iNKT cell development solely by terminating DAG signaling or also by initiating PA-mediated signaling. Further studies are required to determine how dysregulation of DAG-mediated signaling might affect iNKT cell function. The generation of mice that allow for conditional deletion of DGKα or ζ isoforms is likely to prove instrumental in defining the role of DAG-mediated signaling in mature iNKT cell homeostasis and function.

V. ROLE OF DGKα AND DGKζ IN T CELL FUNCTION

VI. SUMMARY

Recent studies have revealed a host of new functions for DGK isoforms in T cell development and function (Figure 3). Apart from their role in the development of conventional αβ T cells, newer work has unveiled a previously unappreciated requirement for synergistic DGKα and ζ activity during invariant NKT cell development. Several modulators of DGK activity, including Lck, SAP, and c-Abl, have been identified. In addition, the importance of DGK activity in promoting MTOC polarization and directional secretion, and in restraining CD8 T cell responses against LCMV infection and tumors, has come to the fore. However, a number of fundamental questions about DGKs remain unanswered. Transcriptional control and regulation of DGK activity via post-translational modifications and protein-protein interactions during T cell development and immune responses are poorly understood. The spatial and temporal regulation of DAG by individual DGK isoforms in T cells remains to be defined. The importance of DGK-generated PA and the downstream effector pathways controlled by PA in T cells need to be explored. A better understanding of the role and regulation of DGK activity and DAG signaling in T cells can enable us to modulate immune responses, producing better outcomes during vaccination, tumor responses, and autoimmunity.