T-lymphocyte trafficking is targeted to specific organs by selective molecular interactions depending on their differentiation and functional properties. Specific chemokine receptors have been associated with organ-specific trafficking of memory and effector T-cells, as well as the recirculation of naïve T-cells to secondary lymphoid organs. In addition to the acquisition of tissue-selective integrins and chemokine receptors, an additional level of specificity for T-cell trafficking into the tissue is provided by specific recognition of antigen displayed by the endothelium involving the TCRs (T-cell antigen receptors) and co-stimulatory receptors. Activation of PI3K (phosphoinositide 3-kinase) is a robust signalling event shared by most chemokine receptors as well as the TCR and co-stimulatory receptors, contributing to several aspects of T-lymphocyte homing as well as actin reorganization and other components of the general migratory machinery. Accordingly, inhibition of PI3K has been considered seriously as a potential therapeutic strategy by which to combat various T-lymphocyte-dependent pathologies, including autoimmune and inflammatory diseases, as well as to prevent transplant rejection. However, there is substantial evidence for PI3K-independent mechanisms that facilitate T-lymphocyte migration. In this regard, several other signalling-pathway components, including small GTPases, PLC (phospholipase C) and PKC (protein kinase C) isoforms, have also been implicated in T-lymphocyte migration in response to chemokine stimulation. The present review will therefore examine the PI3K-dependent and -independent signal-transduction pathways involved in T-cell migration during distinct modes of T-cell trafficking in response to either chemokines or the TCR and co-stimulatory molecules.
The transition from bone-marrow-resident haematopoietic stem cells, through the development of T-cell precursors in the thymus, naïve T-cell migration into secondary lymphoid organs for immune response initiation and maturation into circulating memory and TEFF cells (effector T-cells), involves sequentially co-ordinated activity of adhesion and chemotactic receptors, signalling molecules and the cytoskeleton.
Considerable work on mammalian neurophils and Dictyostelium amoeba models has focused in recent years on the contribution of PI3K (phosphoinositide 3-kinase)-dependent signalling mechanisms to cell motility and chemotaxis. These studies led to the positioning of PI3K at the heart of evolutionarily conserved cellular-compass mechanisms, and PI3K has become a popular drug target for the inhibition of leucocyte migration in response to chemoattractants. However, the precise role of PI3K in the regulation of cell migration remains open to refinement, as numerous examples of PI3K-independent leucocyte migration (particularly with respect to T-lymphocytes), have been described. More recently, PI3K has also been implicated as a key regulator in novel mechanisms mediated by the TCR (T-cell antigen receptor) and co-stimulatory molecules CD28 and CTLA-4 (cytotoxic T-lymphocyte antigen-4)/CD152, that guide the access and retention of specific T-cells into antigen-rich non-lymphoid tissue, and this provides a new platform for as yet unexplored therapeutic strategies with respect to the inhibtion of PI3K and T-cell migration. The present review will examine the role of PI3K and other signalling molecules in T-cell migration during distinct modes of T-cell trafficking mediated by either chemokines and/or the TCR and co-stimulatory molecules.
AN OVERVIEW OF CLASS I PI3KS
The class I PI3Ks are composed of a regulatory subunit and a tightly associated catalytic subunit. The class 1A enzymes are represented by five regulatory subunits encoded by three genes: PIK3r1 encodes p85α and its alternative transcripts p55α and p50α, PIK3r2 encodes p85β and PIK3r3 encodes p55γ. These regulatory subunits bind tightly to one of the three catalytic isoforms p110α, p110β and p110δ and recruit the complex to the plasma membrane upon receptor ligation and phosphorylation. Class 1A PI3Ks are activated downstream of immune-cell receptors, including the TCR, BCR (B-cell receptor), co-stimulatory molecules and cytokine receptors that are phosphorylated by tyrosine kinases upon cognate stimulus [1–4]. The most abundant regulatory isoforms in T-cells are p85α and p50α, with lesser amounts of p85β present [1,3,5,6].
A single class 1B catalytic isoform, p110γ, can pair with one of two regulatory subunits p84/p87 or p101 [7,8]. This isoform is activated by Gβγ subunits and signals downsteam of GPCRs (G-protein-coupled receptors), although some GPCRs, including chemokine receptors, activate class IA PI3Ks (reviewed in [5,9]). Some evidence has been presented previously for the coupling of p110β to GPCRs either by in vitro studies that documented activation of p110β by Gβγ subunits [10,11] or in cellular experiments, where p110β function was probed by microinjection of a neutralizing antibody to p110β [12,13], RNAi (RNA interference) against p110β  or expression of p110β [14,15]. Recent evidence using cells derived from mice with conditional genetic inactivation of 110β has revealed that p110β and p110γ can couple redundantly to the same GPCR agonists, including some chemokines .
Class 1 PI3Ks show broad substrate specificity toward PI (phosphatidylinositol), PI(4)P (phosphatidylinositol 4-phosphate) and PI(4,5)P2 (phosphatidylinositol 4,5-bisphosphate). The major products of class I PI3Ks are PI(3,4,5)P3 (phosphatidylinositol 3,4,5-trisphosphate) and its metabolite PI(3,4)P2 (phosphatidylinositol 3,4-bisphosphate). The levels of PI(3,4)P2 and PI(3,4,5)P3 in cells are usually low under resting conditions, but rise sharply after cell stimulation [1,5,6]. The resulting 3′-phosphorylated lipids have important biological functions that rely on the interaction with effector proteins containing lipid-binding domains, such as PH (pleckstrin homology) and Phox homology domains [17,18]. The effects of PI(3,4,5)P3 are counteracted by the lipid phosphatases PTEN (phosphatase and tensin homologue deleted on chromosome 10) and SHIP (SH2-containing inositol phosphatase), which convert this lipid into PI(4,5)P2 and PI(3,4)P2 respectively .
Expression of p110δ and PI3Kγ is restricted largely to leucocytes and therefore they represent promising targets for the selective inhibition of PI3K-mediated signalling pathways involved in inflammatory and autoimmune diseases. Mice in which the genes encoding p110δ or p110γ have been either ablated or altered to encode kinase-inactive versions are viable, fertile and apparently healthy . However, when their immune system is challenged, they exhibit severely altered phenotypes, demonstrating that p110γ and p110δ have non-redundant functions in mast cells, neutrophils, dendritic cells, B-cells and T-cells, and that the activities of these isoforms in immune cells are crucial during the onset, progression and maintenance of chronic inflammatory diseases [5,6]. Importantly, there is growing evidence that p110γ and p110δ act in partnership to regulate immune cell signalling and function. For example, in TNFα (tumour necrosis factor α)-primed human neutrophils, fMLP (N-formylmethionyl-leucyl-phenylalanine) induces a biphasic increase in PI(3,4,5)P3 accumulation, in which the first phase is dependent on the p110γ isoform . The second phase of PI(3,4,5)P3 accumulation is dependent on prior exposure to TNFα and is driven predominantly by p110δ . With regard to T-lymphocyte biology, it is interesting to note that mice deficient in both p110γ and p110δ (unlike mice deficient in single isoforms) display severe impairment of thymocyte development and profound T-cell lymphopenia, as well as T-cell and eosinophil infiltration of mucosal organs, elevated IgE levels, and a skewing toward Th2 immune responses [21,22]. However, since selective inhibition of 110δ revealed that it actually plays a major role in the control of Th1 and Th2 cytokine secretion from mature T-cells, the pathology in the p110γδ-deficient mice is therefore likely to be secondary to a developmental block in the thymus that leads to lymphopenia-associated inflammatory responses. Hence the serious immune developmental defects observed in the p110γδ-null mice prevent serious dissection of the selective roles of these p110 subunits in post-thymic responses.
T-LYMPHOCYTE MIGRATION AND HOMING
Leucocyte migration requires an interaction with the endothelial lining of blood vessels within inflamed organs and involves four distinct adhesion steps during their migration through blood vessels [23–25]. These include tethering/rolling, activation, arrest and transendothelial migration. Leucocyte tethering and rolling are mediated by the interaction of selectin molecules, which are a family of membrane glycoproteins that have a lectin-like domain that binds to proteins containing specific oligosaccharide groups expressed on the endothelium or on leucocytes. The selectin family comprises E-selectin (endothelial), P-selectin (platelet and endothelial) and L-selectin/CD62L (expressed by all myeloid cells, naïve T-cells and some activated and memory T-cells). The main physiological function of all selectins is to mediate initial interactions with the endothelium, that, in conjunction with blood-flow-mediated shear stress, induces a rolling movement over the endothelial cell surface [26,27]. When the leucocyte is rolling on the endothelium, its receptors engage with endothelial-cell- or tissue-derived chemokines immobilized on the endothelial cell surface. Chemokine binding to GPCRs  expressed by migrating leucocytes induce conformational changes, leading to increased affinity of leucocyte integrins for their cognate endothelial-cell adhesion molecules, which, in turn, mediate their firm adhesion to the endothelium . Integrins are heterodimeric proteins consisting of an α- and a β-chain. Different α- and β-subunits can associate, resulting in 24 members of this family, each containing a distinct pattern of subunits which confers ligand selectivity . In lymphocytes, chemokine-induced conformational changes leading to increased integrin affinity can only occur under flow conditions . Members of the integrin family expressed by leucocytes include the β2 integrins LFA-1 (lymphocyte function-associated antigen 1; also referred to as αLβ2 integrin) and αMβ2, as well as the α4 integrins VLA-4 (very late antigen-4; also referred to as α4β1) and α4β7. Integrins bind to members of the endothelial immunoglobulin superfamily, including ICAM (intercellular adhesion molecule)-1, ICAM-2, VCAM-1 (vascular cell adhesion molecule-1) and MAdCAM-1 (mucosal addressin-cell adhesion molecule type 1), as well as components of the extracellular matrix .
Leucocyte adhesion is followed by emigration into the extravascular tissue, which involves crossing the vascular wall and its associated perivascular basement membrane into the tissue . This process, termed transendothelial migration, can occur through endothelial cell junctions (paracellular route), particularly through junctions between more than two endothelial cells where the cellular contacts are relatively more loose . This is generally considered to be the principal route of leucocyte migration. In addition to the paracellular route, migration through the body of the endothelium has been observed, the so-called transcellular route . A number of endothelial-cell- and leucocyte-expressed molecules have been shown to mediate this process, including ICAM-1 (CD54), ICAM-2 (CD102), JAM (junctional adhesion molecule)-A, JAM-B and JAM-C, platelet/endothelial cell adhesion molecule-1 (also known as CD31), CD99 and endothelial-cell-selective adhesion molecule . Most of these molecules are localized to the borders between adjacent endothelial cells. Many of these molecules are also expressed by various types of leucocytes and probably guide migrating cells through the intercellular junctions via homophilic and/or heterophilic interactions [23,25]. Finally, previous reports have shown that T-cell transendothelial migration can be promoted by a combination of TCR and chemokine-induced signals [36,37], as we will discuss later in the present review. An animation highlighting key stages in T-lymphocyte migration and homing can be found at http://www.BiochemJ.org/bj/418/0013/bj4180013add.htm.
CHEMOKINE-DEPENDENT RECIRCULATION AND ORGAN-SPECIFIC TRAFFICKING
Chemokines are small molecules of approx. 7–10 kDa that form a large cytokine family composed of approx. 50 members in the human system. These proteins are defined either according to the pattern of cysteine residues in the ligands (CC, CXC, C and CX3C) and/or their function and pattern of expression (homoeostatic and inflammatory chemokines) [38,39]. So-called ‘inflammatory’ chemokines constitute the vast majority, are inducible and control cell recruitment to sites of infection and inflammation. Certain ‘homoeostatic’ chemokines are also involved in developmental processes, including lymphopoiesis [40,41]. They can influence an array of cellular activities, including adhesion, angiogenesis and proliferation. They are best known for their effects on motility (chemokinesis) and particularly on directional cell movement, either as soluble chemoattractants (chemotaxis) or when immobilized on tissue structures such as interstitial collagens or a stromal cell network (haptotaxis) . Approx. 20 different chemokine receptors have been identified as mediators of chemokine activities. All chemokine receptors belong to the large family of seven-transmembrane-domain GPCRs. Chemokines are functionally important for the progression of many pathological phenomena, including several inflammatory/autoimmune diseases as well as cancer. T-cell migration in response to GPCR-coupled chemokine receptors has, until recently, been the most extensively studied mode of lymphocyte migration, although there is growing appreciation that lipid chemoattractants, such as S1P (sphingosine 1-phosphate) and eicosanoids, which also engage GPCRs, play a prominent role in navigating distinct T-cell subsets at different stages of the immune response to their intended destinations at sites of infection and inflammation [40,43].
Whereas extravasation of most leucocytes is mediated by cell-specific, but non tissue-selective, inflammatory stimuli, specific chemokine receptors have been associated with organ-specific trafficking of memory and TEFF cells, as well as recirculation of naïve T-cells. Naïve lymphocytes continually traffic from blood through specialized endothelium of SLOs (secondary lymphoid organs), such as PLN (peripheral lymph node), MLN (mesenteric lymph node), spleen and gut-associated lymphoid tissue, including PP (Peyer's patches). Within the SLOs, T- and B-cells localize in the T-cell area and B-cell follicles respectively, where they screen antigen-presenting cells for specific surface–antigen complexes. In case naïve T-lymphocytes do not encounter their antigen, they emigrate via efferent lymphatic vessels back into the bloodstream using S1P-dependent pathways . Efferent lymphatics eventually transport T- and B-cells back into the bloodstream, from which they will continue homing to SLOs, thus completing their recirculation. However, upon activation with cognate antigen in the presence of co-stimulatory molecules, T- and B-cells undergo changes in microenvironmental positioning within the SLOs. This allows T-cell–B-cell interactions at the T-cell area–B-cell follicle border and in germinal centre light zones . Activated T- and B-cells are primed against specific antigen and eventually leave SLOs to accumulate at sites of inflammation or other tissue-effector sites . The capability of tissue-selective homing following priming is provided by expression of various combinations of integrins and chemokine receptors, which provide a unique area code that determines the final destination of the lymphocyte where they will exert their targeted effector response.
Unlike naïve T-lymphocytes, which constitutively traffic through lymphoid tissue, memory T-cells acquire the ability to infiltrate non-lymphoid sites where antigen is located. Antigen-experienced T-cells are more diverse than naïve T-cells and are subdivided into TCM cell (central memory T-cell), TEM cell (effector memory T-cell) and TEFF cell (effector T-cell) subsets which are based largely on distinct homing-receptor-expression profiles that allow their access to specific organs, such as the skin and the gut , as summarized in Figure 1. For example, CCR (CC chemokine receptor) 7 has been associated with T-cell homing and recirculation of naïve T-cells into lymphoid tissue during antigen surveillance. CCR7 expression also discriminates between lymph-node-homing TCM cells and tissue-homing TEM cells. TCM cells retain the lymph-node-homing receptors, L-selectin and CCR7 expression and, like naïve T-cells, are well represented in all secondary lymphoid organs . TCM cells can also localize to peripheral tissues and sites of inflammation [48,49]. For example, expression of high levels of CXCR (CXC chemokine receptor) 4 enables them to migrate efficiently to sites where its ligand, CXCL (CXC chemokine ligand) 12, is expressed, including bone marrow  and PLNs .
Cognate ligand–receptor interactions that facilitate and regulate T-lymphocyte recirculation and trafficking
Both CCR7-negative TEFF cells and TEM cells lack the ability to enter PLNs and home preferentially into non-lymphoid tissues, where expression of tissue-selective chemokine receptors determines their final anatomical destination . Hence expression of the chemokine receptors CCR4, CCR8 and CCR10, along with the cell-surface carbohydrate epitope CLA (cutaneous lymphocyte antigen), directs cells to the skin . Similarly, effector/memory T-cell trafficking into the lamina propria of the small intestine requires the interactions of α4β7 integrin and chemokine receptor CCR9 on lymphocyte surfaces with MAdCAM-1 and CCL (CC chemokine ligand) 25 on endothelial cells of gut lamina propria venules respectively [46,52].
In addition to this model of homoeostatic and tissue-specific trafficking, T-lymphocytes must be able to respond to inflammatory situations. In inflammation-driven migratory patterns of TEFF cells, CD4+ T-cell-driven inflammation is characterized by the predominant type of cytokines secreted by infiltrating T-cells. In Th1-type inflammation, IFNγ (interferon γ) predominates; in Th2-type inflammation, IL (interleukin)-4, IL-5 and IL-13 predominate; and in Th17-type inflammation, IL-17 predominates. These cytokines in turn induce specific subsets of IFN-γ-inducible chemokines; IL-4- and IL-13-inducible chemokines; or IL-17-inducible chemokines at the site of inflammation. These chemokines then induce the recruitment of T-cells that express the specific cognate receptors for these inflammatory chemokines. CXCR3, CXCR6 and CCR5 are preferentially expressed on Th1 cells [41,53], whereas CCR3, CCR4 and CCR8 [along with the prostaglandin D2 receptor CRTH2 (chemoattractant receptor homologous molecule on Th2 cells)] are expressed on Th2 cells [41,54,55]. More recently, CCR6 and CCR4 distinguish the Th17 subset, although there is also evidence that Th17 cells may express CCR2 and CCR9 (Figure 2) [56–59].
Summary of chemokine receptor expression profile on CD4+ T-cell subsets
In contrast with Th1, Th2 and Th17 cells, the naturally occurring Tregs (regulatory T-cells) (CD4+, CD25+ and FoxP3+) are not responsible for promoting cytokine-induced inflammation, but rather for regulating inflammation generated by other T-cell subsets. Tregs can traffic as part of immune surveillance homoeostatically, but also respond to inflammation-induced chemokine gradients to contain excessive T-cell expansion and autoimmunity. It is necessary to understand Treg trafficking patterns in the context of immune-mediated diseases, because this subset could play an important role in controlling both the initiation and the progression of the pathogenic immune response. In mouse models of disease, the chemokine receptors CCR4 and CCR7 play critical roles in Treg trafficking and function in the draining lymph node [60,61], whereas CCR5 is important for Treg trafficking and function at sites of inflammation [62,63]. In addition, CCR8 has been observed on human and mouse FoxP3+ Tregs [60,64,65]. Tregs probably function in peripheral tissue as well as in draining lymphoid tissue and utilize different sets of chemokine receptors to accomplish these functions.
ROLE OF CHEMOKINES IN INTERSTITIAL LYMPHOCYTE MIGRATION
Recent studies using multiphoton intravital microscopy have revealed that T- and B-cells move vigorously within their specific microenvironments, following apparently random migration pathways [45,66]. This interstitial lymphocyte migration in vitro and in vivo is integrin-independent, being mediated by actin flow along the confining extracellular matrix scaffold structure, shape change and squeezing [42,67] and probably serves to increase dendritic-cell-screening efficiency, which may accelerate immune response initiation. As chemokines, even in the absence of a gradient, provoke random migration, SLO-expressed homoeostatic chemokines, such as CCL19/21 and CXCL12/13, may be important motility-inducing factors in vivo. Basal motility of T-cells requires CCL19 and CCL21 (CCR7 ligands) that are abundant throughout the T-cell zone, together with adhesion ligands present on stromal cells [68–70]. SLO-expressed homoeostatic chemokines, such as the CCR7 ligands CCL19 and CCL21, the CXCR5 ligand CXCL13, and the CXCR4 ligand CXCL12, fulfill two functions during lymphocyte trafficking. First, they are presented on high endothelial venules where they promote firm arrest of blood-borne lymphocytes and eventual transendothelial migration across the high endothelial venule [44,45]. Secondly, chemokines organize lymphoid tissue by creating specific microenvironments in lymphoid tissues. Transmigrated CCR7high T-cells move to the PLN paracortex correlating with high concentrations of the CCR7 cognate ligands CCL21 and CCL19 that are expressed in this area. In contrast, CXCR5-expressing B-cells accumulate in follicles where high levels of the CXCR5 ligand CXCL13 are present [44,45]. Accordingly, mice lacking homoeostatic chemokines or their receptors show disturbed SLO architecture, with poorly developed T- and B-cell areas.
SIGNALLING EVENTS THAT CONTRIBUTE TO CHEMOKINE- DEPENDENT-LYMPHOCYTE MIGRATION
Cell polarization, whereby the molecular processes at the front (leading edge) and the back (uropod) of a moving cell are different, is a pre-requisite for efficient migration. Hence the uropod has the microtubule-organizing centre at its base and is rich in adhesion molecules, as well as being the site of mitochondrial redistribution during migration that ensures high ATP at this strategic position [71,72]. In other systems, the small GTPases Rho, Rac and Cdc42, have key roles in regulating cell polarity and the morphology of migrating cells through effects on the actin cytoskeleton and actomyosin contraction, and usually involves cross-talk with other signalling elements such as PI3K [73–75]. We will now consider the contribution and potential interplay between these signalling events during chemokine-stimulated T-cell migration.
Role of Rho family GTPases in T-lymphocyte migration
Polarization and migration of T-lymphocytes requires rapid Rac-driven new formation of F-actin at the leading edge [42,73,76]. In naïve T-cells and B-cells, an early wave of chemokine-induced F-actin formation (less than 5 s) is initiated by the Rac GEF (guanine-nucleotide-exchange factor) DOCK (dedicator of cytokinesis) 2 . T-lymphocytes that are DOCK2-deficient, have less chemokine-induced F-actin formation, cell polarity and in vitro migration . This is a cell-specific defect, as similar disruption of F-actin formation is not observed in DOCK2−/− monocytes . Likewise, the interstitial motility of DOCK2-deficient T-cells and B-cells in lymphoid tissue is much lower . DOCK2 has also been implicated as being required for chemokine-promoted human T-lymphocyte adhesion under shear stress mediated by α4β1 integrin . The Rac GEF Tiam1 (T-cell lymphoma invasion and metastasis 1) also acts during chemokine-induced T-cell migration and associates with members of the Par (partitioning defective) polarity complex that include Par3 and PKC (protein kinase C)-ζ. This complex segregates to the leading edge in polarized cells and helps to establish or stabilize the anterior–posterior axis [80,81]. Curiously, Tiam1 activation appears to occur distal to the activation of both Cdc42 and another GTPase Rap1  that has been implicated previously in polarization, integrin activation and the motility of lymphoid cells [80,82,83]. It has been proposed that Tiam1 acts subsequent to DOCK2 during stabilization of the leading edge/uropod formation after an initial up-regulation of F-actin by DOCK2. However, Tiam1-deficient mice have normal lymphoid structure and cellularity, in contrast with DOCK2- or Gαi2-deficient mice. This indicates that this pathway is partially redundant with other promigratory signalling modules, including other Rac GEFs such as Vav, which is known to be involved in integrin activation and adherence of T-cells [79,84,85]. Indeed, overexpression of Vav mutants abolished lymphocyte polarization, actin polymerization and migration in response to CXCL12, yet, curiously, Vav-1-deficient cells exhibit normal migration , although this may reflect compensation by other Vav isoforms. Moreover, evidence also suggests that Vav localization is influenced by interactions with Tec family kinases that can, in turn, be activated by CXCL12 in T-cells [87,88]. Expression of a loss-of-function Itk (IL-2-inducible T-cell kinase) mutant impaired CXCL12-induced migration, cell polarization and activation of Rac and Cdc42 , whereas T-cells purified from Rlk−/−Itk−/− (where Rlk is resting lymphocyte kinase) mice exhibited impaired migration to multiple chemokines in vitro and in vivo [87,88].
There is strong pharmacological and genetic evidence that Rho-dependent signalling is a key component of T-cell migration and adhesion in response to several chemokines in mature T-cells and thymocytes [89–91]. RhoA activation also appears to be necessary for integrin activation induced by Rap1 and Rac in thymocytes . RhoA appears to control the LFA-1 high-affinity state triggering by chemokines, as well as the lateral mobility induced by chemokines . In addition, formins are downstream RhoA effectors, which are actin-assembly factors. In contrast with Arp2/3 (actin-related protein 2/3), which creates branched actin networks, formins generate elongated actin filaments . Lack of the lymphocyte-expressed formin isoform mDia1 (mammalian diaphanous-related formin 1) leads to impaired in vitro migration in T-cells (but not in B-cells), and is consistent with less chemokine-induced F-actin formation in mDia1−/− T-cells [94,95]. In other systems, RhoA-dependent signalling activated by Gα12/13-coupled GPCRs has been demonstrated to drive myosin-based contraction at the back of the rear of the cell and thus help define the polarized ‘back’ of the cell , although this is not defined in T-cells.
Role of PI3K in cell migration: lessons from neutrophils and amoeba
Several GEFs, particular those with specificity for Rac, are regulated by 3′-phosphoinositides, whereas PI3K activity can itself be modulated by Rho GTPases [97–99]. Moreover, another well-characterized downstream PI3K effector, the serine/threonine kinase Akt, has been implicated in F-actin polymerization and myosin assembly [100–103]. Accordingly, PI3K(s) contribute to several aspects of the migratory machinery, including gradient sensing, signal amplification, actin reorganization and hence cell motility [104–106]. Studies in neutrophils and Dictyostelium have produced a body of evidence which indicated that PI(3,4,5)P3-dependent signals were part of a compass mechanism, sensing and responding to extracellular gradients of chemoattractants [97,107–109]. Several studies have demonstrated that PI3K inhibitors or the genetic loss of PI3Ks causes a reduction in the chemotactic responses of neutrophils and amoebae in a variety of in vitro and in vivo migration assays [107,108,110–112]. Furthermore, use of biosensors composed of fluorescent proteins fused to PH domains, which are capable of binding selectively to different phosphoinositides, revealed that, in contrast with the distribution of chemoattractant receptors and G-proteins, the levels of PI(3,4,5)P3 become highly polarized in amoebae, neutrophils and neutrophil-like cell lines, with high levels closest to the leading edge [97,107–110]. Recent findings have led to a re-evaluation of the model that places PI3Ks centrally in a evolutionary conserved cell-navigational mechanism. First, some experiments examining the effects of either the genetic loss of PI3Ks or selective PI3K inhibitors on the chemotactic efficiency of both neutrophils and Dictyostelium amoebae revealed no specific deficiencies [113,114]. Secondly, PI(3,4,5)P3 polarization to the leading edge of migrating cells was initially thought to be facilitated by the exclusion of PTEN from the leading edge and localization to the trailing edge of the migrating cell [107,108,115]. However, it seems that SHIP rather than PTEN provides a critical role in the polarization and motility of neutrophils [116,117]. Instead, PTEN appears to be a key component in the mechanism by which neutophils prioritize migration toward different competing chemoattractants [116–118]. Finally, the genetic loss of p110γ or selective PI3K inhibitors caused reductions in the chemokinetic responses of the cells (rather than the ability to move toward a gradient) that could explain some of the apparent reductions in chemotactic migration reported previously .
PI3K-dependent T-lymphocyte migration
Activation of PI3K is a robust signalling event shared by most homoeostatic and inflammatory chemokine receptors expressed on T-lymphocytes [9,43]. Accordingly, there has been intense interest in exploring the role of PI3K-dependent signalling in the migratory responses of T-cells after chemokine stimulation. Chemokine interaction with GPCRs on lymphocytes in response to either homoeostatic or inflammatory chemokines has been shown to depend predominantly on Gαi proteins . This led to the assumption that these chemokines receptors are coupled to the βγ-dependent p110γ isoform. This is indeed the case, although several chemokine receptors can activate other PI3K isoforms [9,43,119,120]. Early studies revealed that chemokine-stimulated migration of leukaemic T-cell lines and primary T-cells in standard in vitro assays (e.g. across synthetic membranes on transwell permeable supports in the absence of endothelial cells) is abrogated by pan-isoform PI3K inhibitors [9,43]. Genetic and pharmacological approaches have been employed to assess the contribution of individual class I PI3K isoforms to the migratory response of T-lymphocytes to several chemokines. Hence the in vitro migration of p110γ-deficient CD4+ and CD8+ T-cells to CXCL12, CCL19 and CCL21 is significantly decreased compared with cells from wild-type mice . Moreover, migration of freshly isolated human peripheral-blood T-cells is also inhibited by p110γ-targeting inhibitors, but not inhibitors directed toward the α, β or δ isoforms , although it should be remembered that selectivity of most of these inhibitors for their intended targets is not absolute .
Surprisingly, in migration assays that probably better represent physiological conditions, PI3K inhibitors have little effect on T-cell migration. Hence T-lymphocyte arrest and adhesion to high endothelial venules in exteriorized PP  or transendothelial migration in laminar-flow chambers  in response to either CXCR4 and/or CCR7 ligation is unaffected by PI3K inhibitors. One concern, therefore, is that the reported sensitivity of T-cell migration to PI3K inhibitors is a phenomenon restricted to in vitro assays of migration across bare membranes in the absence of endothelial cells. However, this seems unlikely, as these assays have revealed PI3K-independent migratory responses for either activated or Th2-differentiated human T-cells [89,123]. Other lines of evidence also cast doubt on whether the model for PI3K/PTEN polarization in neutrophils can be applied to T-lymphocytes. For example, many studies have been performed in the Jurkat leukaemic T-cell line. These cells polarize and migrate normally in response to several chemokines acting on pertussis- toxin-sensitive Gαi-coupled receptors, despite the fact that they are devoid of both PTEN and SHIP protein expression . In fact, reconstitution of PTEN expression in Jurkat cells down-regulated CXCL12-stimulated cell migration, suggesting a negative regulatory role for PTEN in T-cell migration . Introduction of a constitutively active SHIP mutant into leukaemic cell lines normally deficient in SHIP, abrogates CXCL12- mediated migration . This was somewhat surprising, given the reported role of SHIP in neutrophil polarization . However, this effect probably reflects the fact that this construct is expressed widely throughout the plasma membrane and disrupts polarized accumulation of PI(3,4,5)P3 at the leading edge. Finally, it appears that the activation status of the cell helps to determine whether PI3K is required for migratory responses to chemoattractants. In this regard, in vitro assays have revealed that migration of freshly isolated human T-cells is dependent on PI3K, but after ex vivo maintenance and activation, the migratory response becomes PI3K-independent . The reader is referred to the animation that accompanies this review for further information (see http://www.BiochemJ.org/bj/418/0013/bj4180013add.htm).
Role of DOCK2 versus PI3K in T-lymphocyte homing and migration in vivo
The use of genetically targeted mice in conjunction with in vivo models of homing of T-cells to PLNs or TEFF cells to sites of inflammation/antigen challenge has refined our view of the role of PI3K in T-cell migration further. Analysis of mice lacking DOCK2 and p110γ alone or in combination has revealed that although DOCK-2 is the predominant molecule required for T-cell migration, p110γ can sustain a modest residual migratory response. Hence optimum T-lymphocyte migration in vivo is dependent on the expression of both DOCK2 and p110γ . There is also some indirect evidence that PI3K may contribute to DOCK2-dependent actin polymerization, as in other cell types PI(3,4,5)P3 can bind to the PH-domain-containing adaptor protein ELMO (engulfment and cell motility), which co-associates with DOCK2 homologues and localizes to membrane ruffles . Whether such interactions occur in T-lymphocytes responding to chemokines remains to be explored. Importantly, a more recent study reported no defect in the migration of p110γ−/− cells to lymph nodes in vivo . The reason for this discrepancy with previous studies may reflect differences in the T-cell populations analysed (e.g bulk compared with CD8+ T-cells). Although p110γ−/− T-cells exhibit minimal defects in migration in vitro and in vivo, it is notable that pan-isoform PI3K inhibitors, such as wortmannin or Ly294002, effectively block in vitro and in vivo migration of naïve T-cells [120,121]. This may simply reflect off-target effects of these compounds or the involvement of other PI3K isoforms in cell migration. Certainly, recent evidence has identified p110δ as being required for antigen-driven T-cell localization, as described below , and is the dominant PI3K isoform in B-cell homing . Interestingly, analysis of neutrophil migration in vivo revealed that although p110γ is important in early chemokine-induced emigration, p110δ replaces and maintains the delayed chemokine-induced neutrophil recruitment into inflamed tissues [132,133]. Whether these isoforms fulfil a similar role during T-cell migration in vivo remains to be established.
The reader is referred to the animation that accompanies this review for further information (see http://www.BiochemJ.org/bj/418/0013/bj4180013add.htm).
Role of PI3K in interstitial T-lymphocyte motility
Recent elegant multiphoton and conventional epifluorescence microscopy studies have explored whether PI3K is involved in regulating basal interstitial T-lymphocyte migration/motility within intact lymphoid tissue in vivo. Despite the established contribution of p110γ to T-cell homing to lymphoid tissue and migration , there was no effect on the dynamic movements of p110γ-deficient T-cells or the pan-PI3K isoform inhibitor wortmannin compared with wild-type controls inside the T-cell area [68–70]. The role of other PI3K isoforms in the dynamic movement of T-cells within lymphoid tissue was not addressed. Interestingly, another group, using multiphoton microscopy in conjunction with wortmannin, revealed a modest reduction in mean T- and B-cell velocities compared with untreated controls . Complementary gene targeting strategies, in which class 1A function had been ablated by deletion of the pik3r1 (p855α-, p55α- and p50α-null) and pik3r2 (p85β-null) gene products showed a significant decrease in velocity and a marked loss of cell polarization . However, these experiments do not distinguish whether reduced motility results from impaired class IA PI3K signalling function or from the loss of adaptor functions of the regulatory subunits independently of their role in activating the catalytic subunits. The reduced motility in wortmannin-treated cells supports at least some role for PI3K enzymatic subunits, but could also be due to the inhibition of other PI3K subclasses or non-PI3K targets of wortmannin [6,43,66,120].
Contribution of phospholipase C and PKC signalling to chemokine-simulated migration
In addition to PI3K and GTPases, chemokine receptor stimulation also leads to activation of several other biochemical signals including PLC (phospholipase C) and DAG (diacylglycerol)-dependent PKC isoforms, as well as elevation of intracellular calcium levels. Given the evidence for PI3K-independent T-lymphocyte migration, it is important to note that these signals have also been proposed as regulators of cell adhesion and migration [134,135]. Curiously, studies with mice deficient in PLCβ2 and PLCβ3 appeared to suggest that the PLC pathway is not required for migratory responses in neutrophils . Subsequent analysis of T-lymphocyte migration in the presence of pharmacological or genetic disruption of PLCβ function revealed a key role for PLC isoforms in T-cell migratory responses in vitro [123,137], once again highlighting the heterogeneity of signalling molecules involved in the migration responses of different leucocytes. The exact role of PLC and its substrate PI(4,5)P2 in cell migration is unclear, but both have been demonstrated to bind to components of the actin cytoskeleton . A distinct family of Ras-exchange factors is regulated not only by calcium, but also by membrane DAG that is generated along with Ins(1,4,5)P3 during activation of PLC. These are termed CalDAG-GEFs (calcium and DAG-regulated GEFs) . One of these, CalDAG-GEF1, functions as an exchange factor for Rap. Interestingly, PLCβ-dependent T-cell migration responses to chemokines was dependent on increased intracellular calcium , although other studies have reported previously that calcium signalling does not play a critical role during lymphocyte migration [43,71,139]. Moreover, some studies have reported that T-cell migration in vitro is resistant to broad-spectrum PKC inhibitors . This is somewhat discrepant with evidence for the involvement of conventional/novel PKC (that are dependent on calcium and/or diaclyglycerol) isoforms in migratory responses, as PKCβI and PKCδ each associate with distinct areas within the microtubules in the uropod during LFA-1-mediated locomotion of activated T-cells. Moreover, in T-cell models where migration to CCR4 stimulation can occur independently of PI3K, the use of pharmacological tools has indicated that PKCδ is required for chemotactic responses to CCR4 ligands [89,140]. This may not fit well the notion that PKCδ is a substrate for the so-called master kinase PDK1 (3′-phosphoinositide-dependent kinase 1) , but it is worth noting that PKCδ is often a functional enzyme in the absence of phosphorylation in the activation loop . Moreover, PDK1 binds PKC via a substrate-docking site domain in PDK1 known as the PDK1-interacting fragment domain, and PDK1–substrate interactions mediated by this domain do not need the PDK1 PH domain . Hence PDK1 phosphorylation at the PKC activation loop almost certainly happens without any necessity for activation of PI3K.
The mechanism by which PKC isoforms regulate cell motility/migration in T-lymphocytes is unclear, although it is likely to be via effects on actin reorganization/polymerization as well as changes in integrin affinity [92,143]. Indeed, the atypical ζ isoform of PKC (another substrate for PDK1) is a component of the Par polarity complex mentioned above that helps establish or stabilize the anterior–posterior axis [80,81]. In this regard, inhibition of PKCζ does not affect initial F-actin formation, but does affect subsequent polarization . However, several PKC isoform knockout mice exist, including those for the β, δ and ζ isoforms, but no major defect in T-cell migration in these mice has so far been reported, possibly due to redundancy in function between individual isoforms .
REGULATION OF T-CELL TRAFFICKING BY THE TCR AND CO-STIMULATORY MOLECULES
Although endowed with tissue-targeting homing properties that allow access to specific organs, such as the skin and gut, primed T-cells must patrol very large areas to locate antigen-rich nonlymphoid tissue to exert their function. So, in addition to acquisition of tissue-selective integrins and chemokine receptors, an additional level of specificity for T-cell trafficking into the tissue is provided by specific recognition of antigen displayed by the endothelium. Within SLOs, antigen-loaded dendritic cells present antigen–peptide–MHC complexes to the TCR on the surface of clonotypic T-cells. Co-stimulatory signals delivered to T-cells in conjunction with TCR engagement are well known to sustain T-cell division, differentiation and survival [145,146]. Negative co-stimulators (such as CTLA-4) counteract these effects, thus promoting homoeostatic mechanisms and tolerance induction. The best-characterized co-stimulatory interaction is mediated by the CD28–B7 molecule pairing and is antagonized by CTLA-4, which binds to the same ligands, namely the B7 family molecules CD80 and CD86. In the absence of co-stimulation, the T-cell will undergo apoptosis or enter a non-responsive or anergic state . We will now consider the role of both the TCR and CD28 family members in determining the destination of antigen-specific T-cells.
Cognate antigen recognition by the TCR influences T-cell trafficking
A direct participation of TCR-derived signals in the regulation of T-cell motility has been suggested by the observation that TCR-triggering enhances integrin activity  and immobilizes migrating T-cells . Moreover, the influence of antigen on the localization, accumulation and retention of specific T-cells in tissue has long been recognized from work performed in animal models of autoimmunity and infection [149,150]. More recently, antigen recognition of B7-deficient human and murine endothelial cells was shown to directly enhance T-cell transendothelial migration without inducing anergy (T-cell unreponsiveness) in vitro [151–153], leading to the hypothesis that memory T-cells are actively recruited to tissues when specific antigens are presented by the endothelium of infiltrating vessels. Several lines of evidence substantiated that antigen presentation by the endothelium contributes to the development and specificity of T-cell infiltrates in vivo. First, the expression of MHC class II molecules by microvascular endothelium in the central nervous system both precedes and is required for the formation of T-cell infiltrates in an EAE (experimental autoimmune encephalomyelitis) model . Secondly, homing of insulin-specific CD8+ T-cells to the islets of Langerhans during the onset of autoimmune diabetes in NOD (non-obese diabetic) mice in vivo was impaired in IFN-γ-deficient NOD mice due to impaired autoantigen presentation by MHC class I and class II molecules . Thirdly, islet-specific homing by insulin-specific CD8+ T-cells was abrogated in mice lacking MHC class I expression or in mice displaying impaired insulin peptide presentation by local endothelium due to deficient insulin secretion, highlighting the ability of endothelial cells to cross-present tissue antigens . Finally, up-regulation of antigen-presenting molecules by local vessels led to peritoneal recruitment of antigen-specific CD8+ lymphocytes in a male antigen-specific transplantation model .
Intravital microscopy revealed that antigen presentation by the endothelium selectively enhanced T-cell transendothelial migration without affecting rolling and adhesion . This is somewhat surprising, given the well-established effect of TCR signalling on the up-regulation of integrin clustering and adhesiveness [147,148]. It is worth noting, therefore, that several studies have shown that the migration of T-cells over the endothelial surface occurs in a highly organized environment with specialized endothelial projections enriched in ICAM-1 and VCAM-1 molecules that act as docking structures for leucocytes [157,158]. In addition, specialized zones with high affinity forms of LFA-1 have been defined in T-cells , and further signals provided by chemokines permit the completion of transendothelial migration and crossing of the endothelial barrier . The dynamic and sequential regulation of adhesive structures ensures that TCR engagement by the endothelium occurs in an environment that facilitates rapid migration, while preventing any loss of cellular attachment and no overall increase in arrest on to the endothelium.
Regulation of T-cell migration by CD28 family receptors
In addition to the TCR, co-stimulatory receptors can also regulate T-cell motility in vitro and their trafficking in vivo. Hence CD28 enhances integrin-mediated adhesion in vitro [159,160] and induces cytoskeleton rearrangements via the Rho family GTPase Rac1  and Cdc42 . Similarly, CTLA-4 enhances integrin-mediated T-cell adhesion via a Rap1-mediated pathway . A prominent feature of the CD28-deficient immune responses is the inefficient localization of primed T-cells to non-lymphoid antigenic sites [164–166]. For example, despite vigorous T-cell responses to myelin-derived antigens ex vivo, CD28-deficient mice exhibited strongly reduced delayed-type hypersensitivity responses and failed to develop EAE [164–166]. CD28 does not appear to mediate adhesion directly, but may favour primed T-cell migration to non-lymphoid tissue by inducing integrin mediated-adhesion  or chemokine receptor expression .
Surprisingly, the negative co-stimulator CTLA-4 has also been shown to enhance T-cell adhesion and migration of pre-activated T-cells on ICAM-1-coated plates . However, the physiological relevance of these in vitro observations remains unclear, as CTLA-4 appears to exert the opposite effects on T-cell motility in vivo by reducing conjugate formation with cognate dendritic cells and specific T-cell retention in antigen-rich lymph nodes . This may provide an additional mechanism by which CTLA-4 contains the expansion of specific T-cells by reducing their cumulative interactions with cognate dendritic cells. Furthermore, CTLA-4 down-regulates the recruitment of TEFF cells to target tissue mediated by antigen-induced signals .
The mechanism by which both CD28 and CTLA-4 can enhance integrin-mediated T-cell adhesion in vitro, yet exert opposite effects on T-cell motility in vivo, is unclear. It is important to highlight that although adhesion is an essential component of cell migration, correct cell polarization and de-adhesive events are equally required for the progression of T-cell motility over a cell surface or through the extracellular matrix, whereas hydrodynamic forces produced by blood flow through the vascular vessels can influence integrin–ligand bonds [31,122,171–173]. As we have seen when considering the role of PI3K in chemokine-stimulated cell migration, these events are replicated rarely in in vitro assays of transendothelial migration. It is also possible that non-overlapping or dominant signalling pathways induced by these two molecules may regulate de-adhesion processes or cell polarization differentially.
ROLE OF PI3K IN ANTIGEN DRIVEN T-LYMPHOCYTE LOCALIZATION
As well as being implicated in cell migration, PI3K has been a well-documented and robust signal elicited upon both TCR and CD28 ligation [2,174–176] and efforts have therefore focused on assessing its role in antigen-mediated cell trafficking. The p110δ isoform is believed to be the primary isoform coupled to TCR and CD28 [3,174,177]. TCR transgenic mice carrying OT-II (ovalbumin-specific T-cell receptor) and a mutation in the cytoplasmic tail of CD28 that abrogates class I PI3K recruitment without leading to defects in clonal expansion (CD28Y170F)  were generated to allow discrimination of conventional co-stimulation-driven clonal-expansion from their ability to infiltrate antigenic tissue (OT-II/CD28Y170F). OT-II and OT-II/CD28Y170F naïve T-cells proliferated equivalently following immunization with OVA323–339 (where OVA is ovalbumin) peptide. However, OT-II/CD28Y170F CD8+ memory T-cells failed to localize to target tissue upon antigen challenge. The PI3K binding motif in CD28 is also required for the binding of the adaptors Grb2 (growth-factor-receptor-bound protein 2) and Gads [Grb2-related adaptor downstream of Shc (Src homology and collagen homology)] and the involvement of PI3K is implied, but not proven. However, subsequent studies using T-cells from mice expressing a catalytically inactive p110δ isoform revealed an essential role for p110δ in TCR-dependent localization of both CD4+ and CD8+ T-cells in a male antigen-specific transplantation model . Interestingly, and in support of previous findings , there was no defect in the p110δ mutant mice of either normal constitutive trafficking or migratory response to non-specific chemokine agonists . Defects in TCR-induced T-cell proliferation and signalling have been reported in p110γ-deficient T-cells . Hence it is important to highlight recent work performed using the OT-II transgenic TCR model in mice lacking p110γ, which revealed no defect in TCR signalling or proliferation in response to antigen, yet their ability to traffic to peripheral inflammatory sites in vivo was severely impaired . This was interpreted as a consequence of the inability of p110γ−/− cells to migrate toward inflammatory chemokines that prevented migration to inflammatory sites. Certainly, signals provided by chemokines permit the full crossing of the endothelial barrier and the completion of antigen-dependent transendothelial migration , and it seems likely that there should be co-operation between TCR and chemokine-mediated signals in the regulation of T-cell migration. Indeed, there is evidence of direct cross-talk between TCR- and chemokine-receptor-mediated signalling. In this regard, ZAP-70 [ζ-chain (T-cell receptor)-associated protein kinase of 70 kDa], SLP-76 (Src homology 2 domain-containing leucocyte protein of 76 kDa) and Tec kinases (key elements in TCR signalling) have been implicated in CXCR4 signal transduction in human T-cells [181,182]. Moreover, CXCL12 (the ligand for CXCR4) stimulates a physical association between CXCR4 and the TCR and utilizes the ZAP-70-binding immunoreceptor tyrosine-based activation motifs of the TCR for signal transduction [183,184]. So it seems that p110δ and p110γ are likely to play complementary (but probably spatially and temporally distinct) roles in the migration of effector cells out of vessels and into tissues. The reader is referred to the animation that accompanies this review for further information (see http://www.BiochemJ.org/bj/418/0013/bj4180013add.htm).
PI3K CAN INFLUENCE EFFECTOR T-CELL MIGRATION AT POST-TRANSLATIONAL AND TRANSCRIPTIONAL LEVELS
Previous studies have focused on the role of PI3K in regulating the mechanical events that are required for polarization and provide forces that drive cell motility, namely the signalling that culminates in the reorganization of the actin cytoskeleton and microtubule network. However, it is becoming clear that an additional route by which PI3K can influence cell migration is via the regulation of transcription factors that regulate cell quiescence and expression of homing receptors on T-cells. In this regard, elegant genetic and pharmacological studies have revealed that p110δ plays an essential role in the events that lead to proteolytic shedding and reduced transcription of L-selectin, as well as reduced transcription of the chemokine receptor CCR7 and the S1P1 (S1P receptor 1) . These surface proteins play an essential role in the homing of CD8+ cells to secondary lymphoid tissues and prevent their egress to sites of peripheral inflammation. A critical event in this pathway is down-regulation of the transcription factor KLF-2 (Kruppel-like factor-2), which promotes L-selectin, CCR7 and S1P1 transcription [186,187]. Inhibition of individual components of the PI3K-dependent pathway prevented the loss of L-selectin and CCR7. In particular, those authors show that the mTOR (mammalian target of rapamycin) inhibitor rapamycin prevents the loss of L-selectin and CCR7 and further demonstrate that rapamycin-treated CD8+ cells preferentially home to lymphoid tissues rather than peripheral sites. These studies not only demonstrate a novel pathway by which PI3K co-ordinates CD8+ cell trafficking, but also provide a previously undiscovered potential mechanism of action of rapamycin as an immune-suppressant drug . Hence the ability of rapamycin to contain activated effector cells in secondary lymphoid organs could result in the destruction of antigen-primed dendritic cells and termination of immune response  and prevent immune destruction of target cells in the periphery. Another study revealed that the transcription factor FOXO1 (where FOX is forkhead box) exerts a steady-state control on L-selectin expression in resting T-cells and this is opposed by PI3K . Central to the regulation of FOXOs by PI3K is their phosphorylation by the PI3K effector protein kinase B/Akt. As a consequence, FOXO molecules are excluded from the nucleus and their transcriptional activities are switched off in the activated cell . It seems to be likely, therefore, that these transcriptional events regulated by PI3K will be an important feature of antigen-dependent trafficking of T-cells to tissue and this will be an important avenue for future research.
The majority of our knowledge concerning signalling mechanisms elicited during T-cell migratory responses is based on work performed on a small fraction of known chemoattractant receptors (predominantly the constitutively expressed CXCR4 and CCR7). It seems unlikely that signalling mechanisms for individual receptors involved in homoeostatic and inflammatory trafficking of cells should be identical. Although we are beginning to understand how a relatively small number of chemokine receptors signal in in vitro settings, the challenge now is to understand how these pathways are integrated, tailored and fine-tuned for each receptor and how they are modulated in inflammatory situations, where there is a vast abundance of chemotactic factors. The diverse milieu of chemokines, adhesion ligands and stromal cell architecture in different regions within lymphoid organs and peripheral tissues, as well as the varying state of activation of individual cells, will determine the expression of antigen, co-stimulatory and chemoattractant receptors that will probably shape the degree of PI3K involvement in T-lymphocyte migration and motility and the choice of isoform. Non-agonist chemokines are capable of associating with known chemokine agonists, resulting in a stronger cellular response, although the molecular basis for this phenomenon has not been determined . As a consequence, inflamed and other chemokine-rich tissues would create an environment that renders many leucocyte types more competent to respond to migratory cues. The influence of the local environment, such as partial pressure of oxygen, has also been demonstrated to influence cell migration , whereas various components of the PI3K pathway, in particular the lipid phosphatase PTEN, are sensitive to local redox conditions . In this regard, T-cell motility has been shown to vary between the sub-capsular and deep paracortical regions of the node . Some insight into how leucocytes translate different migratory cues into distinct biochemical messages has been provided by a recent study that demonstrated neutrophil migration toward fMLP is dependent on p38 MAPK (mitogen-activated protein kinase) and phospholipase A2 activation, whereas migration toward the chemokine CXCL2 is dependent on PI3K . Remarkably, PTEN-deficient neutrophils migrate normally to single gradients of the end-target bacterial chemoattractant fMLP, but not to intermediary chemokines. Those authors proposed a model in which fMLP receptor stimulation results in p38-dependent redistribution of PTEN from the uropod to the entire cell body, eliminating cell polarity established by chemokine-stimulated PI(3,4,5)P3 and redirecting the neutrophil toward the site of bacterial infection . The choice of biochemical signal elicited by individual receptors will also be shaped according to the mode of presentation of agonist (e.g. whether it is immobilized by glycosaminoglycan binding or forms homo- or hetero-oligomers), receptor coupling to additional G-proteins other than Gαi  and/or whether the receptor forms dimers . Certainly, chemokine receptor heterodimerization promotes signalling events distinct from those elicited by homodimer receptors, including the recruitment of G-proteins other than Gαi [194–196]. The impact of the now widely accepted hetero- and homo-dimerization of chemokine receptors on receptor pharmacology, signalling and cell biology is not fully appreciated, and the impact on cell trafficking remains a mystery.
Several therapeutic strategies have been explored to prevent leucocyte migration, including blockade of adhesion molecules, chemokine receptors and signalling events, such as those mediated by p110γ . However, the role of PI3K isoforms in T-cell migration is not ubiquitous, but it is probably determined by individual agonist–receptor interactions as well as the activation state of the cell (Figure 3). Inhibitors targeting p110γ are certainly effective in several models of chronic inflammatory disease, most probably due to inhibition of neutrophil as well as CD4+ T-cell migration [197–199]. However, these inhibitors were used at concentrations that will almost certainly inhibit other isoforms, so the contribution of additional isoforms cannot be discounted. The differing dependence of individual chemokine receptors and antigen receptors on PI3K isoforms at different stages of activation makes it difficult to design a ‘one fits all’ drug to inhibit the inflammatory recruitment of cells. Both genetic and pharmacological strategies have revealed that p110γ can make a contribution to the migration of naïve cells to lymph nodes, whereas both p110γ and p110δ contribute to the migration of effector cells to sites of inflammation. Our recently acquired appreciation of the role of p110δ in regulating primed T-cell migration to antigenic sites provides an additional dimension to its potential as a pharmacological target in the control of T-cell-mediated pathologies, including autoimmunity and transplantation. Hence selective targeting of p110δ may avoid undesired T-cell-dependent inflammation by preventing antigen-dependent T-cell migration and subsequent cell–cell interactions without inducing overt immune suppression.
PI3K influences T-lymphocyte migration and motility at multiple levels
CC chemokine ligand
CC chemokine receptor
CXC chemokine ligand
CXC chemokine receptor
cytotoxic T-lymphocyte antigen-4
dedicator of cytokinesis
experimental autoimmune encephalomyelitis
calcium and DAG-regulated GEF
growth-factor-receptor-bound protein 2
intercellular adhesion molecule
IL-2- inducible T-cell kinase
junctional adhesion molecule
lymphocyte function-associated antigen 1
mucosal addressin-cell adhesion molecule type 1
mammalian diaphanous-related formin 1
ovalbumin-specific T-cell receptor
3′-phosphoinositide-dependent kinase 1
domain, pleckstrin homology domain
protein kinase C
peripheral lymph nodes
phosphatase and tensin homologue deleted on chromosome 10
S1P receptor 1
SH2-containing inositol phosphatase
secondary lymphoid organ
Src homology 2 domain-containing leucocyte protein of 76 kDa
T-cell antigen receptor, TCM cell, central memory T-cell
cell, effector T-cell
cell, effector memory T-cell
T-cell lymphoma invasion and metastasis 1
tumour necrosis factor α
vascular cell adhesion molecule-1
ζ-chain (T-cell receptor)-associated protein kinase of 70 kDa
The work of the laboratory of F. M. M.-B. is supported by the British Heart Foundation [grant numbers PG/05/136/19997, PG/07/090/23697]. S. G. W. is the recipient of a Royal Society Industrial Fellowship.