OUP user menu

Molecular regulators of leucocyte chemotaxis during inflammation

Connie H.Y. Wong , Bryan Heit , Paul Kubes
DOI: http://dx.doi.org/10.1093/cvr/cvq040 183-191 First published online: 2 February 2010


A fundamental feature of any immune response is the movement of leucocytes from one site in the body to another to provide effector functions. Therefore, elucidating the molecular mechanisms underlying the migration of leucocytes from the blood to tissues is critical to our understanding of immune function during inflammation. The classic steps of leucocyte trafficking involve leucocyte tethering and rolling on vessel walls of the vasculature, followed by firm adhesion to the endothelium. Recent evidence suggests that upon adhering, leucocytes crawl within the vessels before transmigrating across vessel walls and crawling into targeted tissues. The directed nature of the crawling events is orchestrated by a complex array of soluble factors and molecular regulators in combination with the local intravascular and extracellular environment. In fact, this process is known as chemotaxis and orientates cell movement in relation to the ligand gradient. Several signalling pathways have been proposed to be involved in this gradient-sensing and amplification process, but the best studied, discussed in detail here, is the phosphatidylinositol 3-kinase pathway. Substantial progress has been made in understanding how cells roll and adhere in blood vessels; however, how cells crawl in blood vessels, emigrate, and then crawl in tissues has received much less attention. Therefore, the focus of this review is to provide recent insights into molecular mechanisms and cellular processes that mediate leucocyte crawling in blood vessels and tissues during the inflammatory response.

  • Neutrophils
  • Chemotaxis
  • Directed migration

1. Introduction

A successful immune response depends on the capacity of leucocytes to travel from one location in the body to another promptly.1 This requires that the cell negotiates high shear rates in blood vessels, tethers via selectins to the endothelium, rolls via selectins, and adheres via integrins upon encountering activating molecules such as chemokines expressed on endothelium. The cell than emigrates out of the vasculature via an array of adhesion molecules. A new essential step in the classic leucocyte recruitment cascade was recently revealed. Specifically, intraluminal crawling from a site of firm adhesion to the junctional extravasation site has led to a revised and expanded version of the original leucocyte recruitment cascade.2,3 In addition, in complex multicellular organisms, leucocytes often have to move significant distances and therefore encounter multiple microenvironments and various chemoattractants to reach the target site of infection or inflammation. The directional migration is a process known as chemotaxis. In fact, chemotaxis is a fundamental phenomenon exhibited by a wide variety of cell types in the context of angiogenesis, inflammation, wound healing, and many other complex physiological processes. Remarkably, many of the molecular mechanisms involved in controlling leucocyte chemotaxis arose billions of years ago in the simple eukaryotic organism Dictyostelium discoideum. The movements of leucocytes are thought to be controlled by internal and external signals which activate complex signal transduction cascades to mediate chemotactic responses, resulting in highly dynamic and localized remodelling of the cytoskeleton that ultimately leads to directed migration. This review aims to summarize some of the molecular mechanisms underlying chemotaxis in various environments in vivo including the vasculature and the extravascular space.

2. Directed intraluminal crawling

Immune cells undergo a series of sequential steps during extravasation from blood into tissues, which includes tethering, rolling, adhesion, crawling, and transmigration.3 It is likely that crawling in blood vessels and in tissues requires directionality. A recent in vitro study first reported the additional and essential crawling step in this multi-step paradigm. This group observed human monocytes crawled from a site of firm adhesion to a junctional site, a process termed locomotion.4 At this site, the monocytes then either emigrated immediately or followed the junction until they found a site of emigration. Blocking various adhesion molecules on monocytes or endothelial cells prevented the monocytes from reaching junctions. For example, blocking integrins LFA-1 (CD11a/CD18; αLβ2 integrin) and Mac-1 (CD11b/CD18, αMβ2 integrin) on monocytes and integrin ligands ICAM-1 and ICAM-2 on endothelial cells prevented monocyte crawling. These monocytes were able to adhere and polarize normally but could no longer crawl along the surface of the endothelium. Interestingly, when crawling was inhibited, the subsequent emigration was also prevented, suggesting directed crawling is a crucial step in the monocyte recruitment cascade.

More recently, intraluminal crawling was demonstrated in vivo.2,57 Neutrophils crawled perpendicularly to or even against the flow of blood within inflamed post-capillary venules.2,7 Since neutrophils that adhered crawled before emigrating, it was hypothesized that neutrophils did not adhere at optimal sites of emigration and thus crawling was a mechanism used to find these optimal sites. It was possible to test the importance of crawling for emigration in vivo, in part because a specific molecule appeared to be important for neutrophil crawling. Whereas adhesion was entirely dependent upon LFA-1, crawling required Mac-1. In fact, only a few of the LFA-1-deficient neutrophils adhered in inflamed venules, but those that did crawled efficiently.2 In contrast, Mac-1-deficient neutrophils adhered well but failed to crawl in response to exogenously applied chemokines, including macrophage inflammatory protein-2 or endogenously produced chemokines induced by applying tumour necrosis factor-α (TNF-α).2 Additionally, Mac-1-deficient neutrophils were delayed in their ability to emigrate.

Despite intravascular monocyte crawling in vitro was absolutely essential for emigration,4 impaired neutrophil crawling in vivo led to similar levels of emigration, although the process took longer.2 There are a number of possibilities of why the in vitro and in vivo studies do not completely agree. The difference is possibly related to the macrovascular endothelium used in vitro when compared with the microvascular endothelium in vivo. Alternatively, in vivo shear forces were absent in the in vitro studies. It is noteworthy that shear forces were absolutely necessary to allow for the emigration of certain cell types including eosinophils8 and lymphocytes.9 Although we are not aware of a study that examined whether monocytes emigrated better under shear conditions, one could imagine if they were, then preventing them from crawling without shear could prevent them from emigrating. Finally, the in vivo study revealed a dramatic increase in neutrophil emigration through the endothelial cells (transcellular migration) rather than at junctions (paracellular) when crawling was inhibited by blocking Mac-1. If the Mac-1-deficient neutrophil adhered in the middle of an endothelial cell (a common occurrence), it migrated right through the endothelium. Therefore, it is possible that monocytes are unable to emigrate transcellularly the way neutrophils can.

There is some evidence that the intraluminal crawling observed during an inflammatory response is directional. For example, placing chemokines on one side of a blood vessel or injuring tissue at one side of a vessel resulted in cells adhering along the side of the vessel proximal to the inflammation. However, neutrophils also adhered on the opposing side of the blood vessel but these cells then crawled perpendicularly towards the opposing vessel wall and then emigrated. Cells were generally not seen crawling away from the wall closest to the chemokine source (M. Philipsson, personal communication) or injury (Menezes and Kubes, unpublished results). If there is a gradient of a molecule in the vasculature under flow conditions, it would need to be a gradient immobilized to the substratum (termed haptotaxis) to avoid being washed away. A recent paper has challenged this view and reported no chemotaxis in blood vessels. Visualization and quantitative analysis of directional migration of neutrophils in their natural in vivo environment post-chemoattractant or bacterial delivery via perivenular microinjection demonstrated this point.10 Presumably, the microinjection of a chemoattractant or bacteria into the interstitial tissue would establish a chemotactic gradient and mimic the local release of mediators during inflammation. Indeed, this approach induced a significant increase in leucocyte adhesion, transmigration, and motility of transmigrated leucocytes; however, there was no preferential adhesion or emigration on the ipsilateral aspect of the venule. In fact, the straight-line velocity and directionality of leucocytes were several times higher on the vessel side ipsilateral to chemoattractant microinjection compared with those on the contralateral side.10 Taken together, these findings demonstrated that microinjection of inflammatory mediators or bacteria induced non-selective intravascular adhesion and transmigration, suggesting no chemotactic gradient inside the blood vessels but directional interstitial migration of leucocytes towards the locally administered stimulus. Although a gradient was clearly retained outside the vessels, the gradient could have rapidly been destroyed inside blood vessels via blood flow.

Mechanotaxis or responses to forces exerted by blood flow can also affect crawling. For example, neutrophils under non-flow conditions will crawl in all directions on a coverslip. Once shear is applied, the cells begin to move in a coordinated fashion perpendicular to the applied shear7 (Figure 1). Mechanoreceptors including perhaps the β2-integrins present on neutrophils may instruct the cells to move perpendicularly to flow. One could speculate that upon adhesion in the vessel, perhaps in the middle of an endothelial cell, crawling perpendicularly would increase the probability of reaching the nearest junction. Alternatively, crawling in blood vessels is optimal perpendicular to flow.

Figure 1

Neutrophil crawling. (A) Neutrophils demonstrate random migration under non-flow conditions in vitro. (B) Once shear is applied, neutrophils crawl in a coordinated fashion perpendicular to the applied shear. (C) Intraluminal crawling requires an exquisite communication between LFA-1 and Mac-1. LFA-1 mediates firm adhesion of neutrophil to the endothelium and would perhaps disengage only after Mac-1 binds its ligand and mediates the subsequent crawling mechanisms. Vav-1 acts downstream of adhesion and is essential for optimal Mac-1-dependent crawling and emigration of neutrophils under physiological stress.

In addition to the intraluminal crawling observed during an inflammatory response, there is now good evidence to suggest that there are populations of cells that continuously patrol organ microvasculature under non-inflamed conditions. For example, there is a population of monocytes localized to the vasculature of skin that continuously patrols the skin, mesentery, and brain microcirculation, and upon infection, these are the very first cells to enter the tissue, perhaps enhancing subsequent neutrophil recruitment. A second population of intravascular cells are the CD1d-restricted iNKT cells, which crawl randomly or ‘patrol’ within liver sinusoids at a speed of 10–20 μm/min and stop crawling upon TCR activation, revealing a novel form of intravascular immune surveillance.11 Intravenous injection of a CD1d ligand resulted in the arrest of iNKT cells presumably on CD1d-expressing endothelial cells, Kupffer cells, or parenchymal cells that extend pseudopods through the fenestrae of the endothelium. However, when pathogens were infused intravascularly and captured by Kupffer cells, the iNKT cells formed large clusters around these Kupffer cells suggestive of chemotactic behaviour within the vasculature ( Lee and Kubes, unpublished results).

3. In vivo vs. in vitro migration

The hallmarks of crawling include polarization, cytoskeletal rearrangement, dynamic pseudopod protrusion and retraction, flexible oscillatory shape changes, and rapid low-affinity crawling.12 Much of the in vitro work has focused on how chemotaxing cells polarize—that is, form a leading edge (pseudopod) facing the source of a chemoattractant, and a trailing tail (uropod). This process is central to chemotaxis, as vastly different events occur in the front vs. back of the cell. The pseudopod of the chemotaxing cell is highly dynamic, with actin polymerization forming a fine network of actin fibrils which act to push out the leading edge.13 At the same time, adhesive structures, consisting of activated integrins, actin fibres, and linking proteins such as α-actinin and talin, adhere to the pseudopod to the substratum.14,15 Vastly different processes are active in the rear of the cell. Initially, the loose network of actin formed along the pseudopod becomes structured into large bundles, producing stress fibres which the cell uses to retract the uropod.16,17 The pathways mediating this bundling are not well characterized, although actin-binding proteins such as heat shock protein 27 and leucocyte-specific protein-1 are likely involved.18,19 The contractile force which pulls the uropod along is provided by myosin type II, which contracts the cytoskeleton, and thus retracts the uropod towards the pseudopod.20 At the very rear of the cell, these actin bundles are disassembled, reverting to actin monomers which can then diffuse to the pseudopod, where they can be used for additional actin polymerization. Simultaneously, the adhesive structures are also broken down and the constituent molecules are either recycled or degraded.20

Whether these same molecules are important for crawling in vivo remains largely unexplored. It is feasible that the above polarization occurs in blood vessels but additional molecules to resist shear may also be activated. Most experiments in vivo in knockout systems are done as a black box approach by simply evaluating whether more or less cells enter tissues. However, in these studies, a particular mutation could be impacting rolling, adhesion, crawling, or emigration, and so these studies provided limited insights into intravascular crawling per se. In the presence of shear, it is conceivable that additional or different molecules are invoked for crawling than in the absence of shear. Indeed, it has been known that applying shear to neutrophils causes them to respond by upregulating additional adhesion molecules. A clear example is the fact that under static conditions, neutrophils crawl randomly, but once shear is applied, neutrophils prefer a perpendicular direction.7 The precise intracellular signalling pathways that might be important to impact crawling/chemotaxis leading to emigration are only being identified recently.

Vav-1 is a haematopoietic-specific Rho/Rac guanine nucleotide exchange factor and is one of the few, if not the only, intracellular signalling molecule known to contribute to crawling within the vasculature. Vav-1 has been demonstrated to play a central role in mediating the activity and localization of β2 integrins.2123 As such, Vav-1 was hypothesized to play a central role in mediating both chemotaxis and intravascular crawling, as both of these processes are highly dependent on the activity of the β2 integrins.2,24 Several groups have reported various degrees of impact on crawling in neutrophils deficient in Vav-1 in vitro.21,25,26 A recent in vivo study demonstrated that Vav-1 acts downstream of adhesion and is essential for optimal Mac-1-dependent crawling and emigration of neutrophils under physiological stress7 (Figure 1C). Under static in vitro conditions, wild-type and Vav-1-deficient neutrophils crawl randomly in all directions. When physiological shear was applied, the majority of wild-type neutrophils displayed perpendicular amoeboid crawling, but Vav-1-deficient neutrophils appeared stretched with a narrow prolonged uropod and almost always moved in the direction of flow. This phenotypic difference in crawling behaviour was observed in both in vivo and in vitro settings.7 Interestingly, Vav-1-deficient neutrophils had impaired emigration out of inflamed microvasculature and it was likely due to the failure of the uropod to detach from the endothelium in these cells. Indeed, crawling under flow conditions requires molecular pathways of migration as well as signalling mechanisms which result in the attachment of the neutrophils onto endothelium that would prevent detachment due to shear but still allow for smooth forward motion. Therefore, it is feasible that mechanoreceptors (perhaps integrins) are activated during flow conditions and this in turn induces the activation of signalling molecules such as Vav1, which plays a central role in mediating directed neutrophil crawling in blood vessels.

The vast majority of studies on directional migration were performed in vitro, but it is important to keep in mind that the mechanisms underlying migration differ between in vivo and in vitro settings.1,27,28 For example, the complex architecture of the interstitial tissue as well as the phenotypic and functional changes of leucocytes that result from their interactions with endothelial cells and basement membrane during adhesion and diapedesis are disregarded in in vitro studies.28 Investigators studying neutrophil chemotaxis in vitro harvest neutrophils from blood and then perform chemotactic assays. In contrast, a neutrophil that chemotaxes through tissue has previously attached to the endothelium via selectins and then integrins and both sets of adhesion molecules responsible for these steps have been shown to be associated with outside-in signalling.3 Next, the cell must emigrate, a step that appears to be important for the upregulation of various adhesion molecules including the matrix protein ligands (β1 integrins), not normally found on circulating neutrophils.29,30 Therefore, the chemotaxis of circulating neutrophils may not reflect the chemotaxis in tissues but might be relevant to crawling within the vasculature.

Additionally, the complicated interplay of chemotactic cues in vivo is dramatically different from their capacity to induce chemotaxis in vitro.1,27,31 The diffusion properties of a chemoattractant may differ in vitro vs. what might occur in tissues. In fact even in vivo, injection of a chemoattractant into tissues or superfusing tissues with chemoattractants is somewhat artificial.3234 In an attempt to overcome this, an agarose gel containing chemokine was placed onto interstitial tissue adjacent to a post-capillary venule in the murine cremaster muscle to allow for the slow release of chemokines.35 However, this may still be lacking all of the events of a true inflammatory response which needs to take into account the distribution of chemoattractant inside tissues and not topically where it can diffuse much faster.

In vivo imaging plays an increasingly important and powerful role in the analysis of spatio-temporal dynamics of directed leucocyte migration. However, many of the intravital microscopic studies to date have not delineated what leucocyte subset is being studied due to the lack of lineage-specific markers for in vivo studies. With the advent of green fluorescence protein (GFP)-leucocyte subset-specific mice, there are now mice available with fluorescent monocytes, neutrophils, various lymphocyte subsets, and dendritic cells. Using genetically modified animals, one group of investigators were able to demonstrate differential migratory behaviours between neutrophils and monocytes.10 Monocytes were distinguished from neutrophils via their expression of GFP in the Cx3CR1gfp/gfp mice.36 Monocytes began their target-oriented interstitial migration later and slower than neutrophils, following perivenular microinjection of monocytes chemotactic protein-1, as also reported in in vitro studies.37,38 Intuitively, the high motility of neutrophils would enable them to accumulate more rapidly at the site of inflammation, and it is also feasible that neutrophils are required to produce chemotactic factors for monocyte migration under certain inflammatory conditions39 or perhaps to clear the way for monocyte entry.

4. Phosphatidylinositol 3-kinase-dependent and -independent sensing of chemoattractants

The ability of cells to sense external chemical cues and respond by directionally migrating towards them is a fundamental process essential for many biological responses in the human body, including the invasion of leucocytes to sites of inflammation.3 Chemotaxis orients cell movement in relation to the ligand gradient. This gradient signal is received through specific receptors on the surface of the cell and is subsequently transduced through intracellular signalling machinery to generate a coherent response. This process has been studied for many years in a variety of model systems.40,41 Chemotaxing cells have an amazing ability to detect small changes in the concentration of a chemoattractant.42 This high-sensitivity detection requires some form of gradient amplification within the cell. Several signalling pathways have been proposed to be involved in this gradient-amplification process, but the best studied being the phosphatidylinositol 3-kinase (PI3K) pathway.4345 Sensing of chemotaxing cues is achieved by the activation of G protein-coupled receptors (GPCRs). It has been proposed that upon detecting a chemotactic stimulus, cells will activate PI3K in such a way that PI3K is active along the region of the cell facing the chemoattractant, resulting in the phosphorylation of phosphatidylinositol (4,5) bis-phosphate (PI(4,5)P2) to produce an accumulation of phosphatidylinositol triphosphate (PIP3) along the pseudopod of the cell.4650 This is followed by the accumulation of proteins containing PIP3-binding domains, thus recruiting the proteins required to form the pseudopod of the migrating cell.49,5153 At the same time, enzymes which mediate the breakdown of PIP3 are active on the sides and the uropod, thus limiting PI3K activity to the front of the cell.46,54,55

Numerous in vitro studies have reported deficiencies in neutrophil chemotaxis in the presence of non-selective PI3K inhibitors or in the absence of PI3Kγ.44,56,57 Moreover, PI3Kδ was shown to have a significant role in human neutrophil migration towards formyl–methionyl–leucyl–phenylalanine (fMLP).58 In lipopolysaccharide (LPS)-primed cells, this chemoattractant-stimulated migration became completely insensitive to PI3K inhibitors,59 suggesting the roles described above for PI3Ks are context dependent, potentially explaining a large number of contradictory reports of the effects of PI3K inhibitors on the migration of human neutrophils in vitro. Nevertheless, several in vivo studies have resolved multiple roles for PI3Ks (particularly PI3Kγ and δ) in neutrophil migration.60,61 PI3Kγ, in the neutrophils, and not endothelium, was important for the emigration of neutrophils from the circulation in response to brief stimulation with chemokines, and after longer stimulation, emigration was dependent on PI3Kδ.60 Furthermore, the inhibition of PI3K activity resulted in fewer cells exiting the circulation and therefore fewer cells being in a position to migrate to a site of inflammation.

The concept that PI3K signalling becomes sharply polarized to the pseudopod during chemotaxis has received substantial support, but the molecular mechanisms sustaining this response have not been fully clarified. There is also controversy regarding the role of PI3K in the regulation of movement, direction, or both during chemotaxis-mediated neutrophil migration.62,63 Additionally, there are conflicting data questioning the absolute requirement of PI3K activity in neutrophil chemotaxis towards the bacterial peptide fMLP.58,59,64 Moreover, the primary functions of the PIP3 signal also remain poorly understood. Despite lacking precise detailed molecular mechanisms, it is feasible that PIP3 functions to regulate the distribution of key regulators of the location and the extent of actin polymerization.51,65,66 Clearly, the exact role of PI3K during chemotaxis remains to be systematically elucidated, and is a matter of some controversy.

Perhaps surprisingly, the GPCRs that bind to chemoattractants remain uniformly distributed across the cell surface.67 Also, the G protein subunits do not become concentrated at the pseudopod, instead localizing in a shallow anterior-to-posterior gradient that approximates receptor occupancy.68 Instead, PIP3 needs to accumulate along the pseudopod, where it can then mediate a variety of processes including actin polymerization and integrin activation.69 At the same time, PIP3 needs to be excluded from the remainder of the cell; otherwise, a chemotaxing cell would be unable to form a single leading edge. This restriction of PIP3 to the pseudopod appears to be mediated by a protein called phosphatase and tensin homolog deleted on chromosome 10 (PTEN).46 PTEN catalyses the reverse reaction of PI3K (dephosphorylates PIP3 to form PI(4,5)P2), thus directly antagonizing PI3K activity.70,71 During chemotaxis, PTEN localized to the uropod and to the sides of the cell, therefore neutralizing PI3K activity in these regions, resulting in the restriction of PIP3 accumulation to the pseudopod.46,72 Recent studies have examined the effects on PIP3 levels and chemotaxis in neutrophils lacking PTEN.54,73 In one study, these cells displayed significantly higher Akt phosphorylation, increased PIP3 levels, enhanced chemoattractant-induced membrane ruffling, and actin polymerization but surprisingly this resulted in only a small defect in chemotaxis.73 In contrast, another study observed comparable PIP3 levels and no impact on chemotaxis.54 Instead, the Src homology 2 domain-containing inositol-5-phosphatase 1 (SHIP1) was shown to regulate PIP3 levels and that cells lacking SHIP1 had severe polarity and motility defects. These cells were unable to assemble actin properly, but could still migrate, though very slowly. A third group reported that the absence of PTEN in neutrophils results in a dramatic increase in PI3K activity and a propensity for random migration in response to endogenous chemokines. In contrast, responses to bacterial products such as fMLP had little impact on neutrophil chemotaxis. On the basis of these findings, a potential two-step polarization process was proposed. When exposed to chemoattractant, neutrophils can become partially polarized, independent of PI3K (via an unknown mechanism), whereby the neutrophils are already capable of directional movements at intermediate speed. Thereafter, PI3Kγ/PTEN and SHIP1 cooperatively confine PIP3 to the leading edge and full polarization, which are required for full motility. Alternatively, different classes of chemoattractants use different signalling pathways.

Immune cells are required to travel considerable distances through many microenvironments and encounter a combination of chemoattractants and perhaps even chemorepellents that modify their directional trafficking in vivo. Whereas a single chemoattractant is frequently used in vitro, multiple chemoattractants might be encountered in vivo. During a host response to bacterial infection, neutrophils must first respond to intermediary cues from endogenous chemokines expressed by endothelial cells in the vasculature and by sentinel cells such as macrophages and mast cells in the interstitium. These signals lead the neutrophils into the general vicinity of the bacteria. Ultimately, the neutrophils must disengage from the intermediary chemoattractants and migrate unidirectionally towards the bacteria-derived end-target chemoattractants.74 Indeed, Butcher and colleagues75 have described a hierarchy of chemoattractants with bacterial products (fMLP) and chemoattractants produced in the vicinity of bacteria (C5a) dominating over endogenous chemokines produced along the trek towards the bacteria. From this work, it seemed reasonable to propose that there are multiple chemotactic signalling pathways. Indeed, in follow-up to the work by Butcher and colleagues, others have demonstrated that (i) there are different signalling pathways for different chemoattractants and (ii) there is an intracellular signalling hierarchy in response to different chemoattractants.64,76

In an under-agarose assay, neutrophils predominantly migrated towards end-target chemoattractants (e.g. fMLP and C5a) via p38 MAPK, whereas intermediary chemoattractant (e.g. IL-8 and LTB4)-induced migration was PI3K dependent. Moreover, when faced with competing gradients of end-target and intermediary chemoattractants, Akt activation was significantly reduced within neutrophils, and the cells migrated preferentially towards end-target chemoattractants even at one thousandths that of intermediary chemoattractants, further suggesting an intracellular signalling hierarchy.64 This could be particularly important in the presence of multiple sources of chemoattractants that could distract neutrophils during their final approach to end-target pathogens, therefore neutrophils must have the ability to integrate and prioritize chemotactic cues so that they can selectively form a unidirectional path to reach the site of infection. However, the situation may be complicated by the fact that there may be an early role for PI3K in response to fMLP. Exposure of neutrophils to fMLP induced an immediate polarization and directional migration towards this chemoattractant within 2–3 min. PI3K-inhibited neutrophils also polarized and migrated in a directional manner, but this process was delayed by ∼15 min, demonstrating that PI3K accelerates the initial response to fMLP, whereas an alternative pathway replaces PI3K over time.77 Indeed, p38 MAPK-inhibited cells or cells lacking MK2, a protein kinase downstream of p38 MAPK, failed to polarize in response to fMLP. Therefore, although PI3K can enhance early responses to the bacterial chemoattractant fMLP, it is not sufficient or required for migration towards this chemoattractant. Whether this paradigm holds true in other leucocyte is unclear, and recent studies using mouse T cells have shown that although PI3Kγ is not essential for constitutive migration of these cells, it is important for the migration of these cells to sites of inflammation.78 Collectively, the role that PI3K plays in directional migration is context dependent and suggests that cells have evolved a series of interdependent signalling pathways that mediate this essential cellular function to provide the needed flexibility to respond under different physiological conditions.

Importantly, so that both PI3K and p38 are not competing against each other, it appears that the latter inhibits the former perhaps via relocalization of PTEN. Indeed, when neutrophils are activated via fMLP, PTEN is placed at the sides and rear of the cells. Now, addition of a chemokine at sites other than the pseudopod will be ignored since PIP3 can no longer form at other sites. In contrast, when a chemokine such as IL-8 induces a pseudopod of PIP3 and PTEN along the uropod and sides of the cells, fMLP is PIP3 and PTEN independent and so the addition of fMLP at the tail will add PTEN to the pseudopod, preventing further migration towards IL-8 and allowing a new leading edge to be formed towards fMLP, independent of PIP3 and PTEN. Neutrophils deficient in MK2 and neutrophils treated with a p38 MAPK inhibitor exhibited a disruption of PTEN localization at the rear of the cell, with some PTEN localizing to the pseudopod.79 Under these conditions, the neutrophil can no longer discriminate between different classes of chemoattractants and migrates towards whichever chemoattractant it sees first.80

Currently, there are at least three additional potentially inter-related alternative, parallel pathways in addition to the PI3K and p38 pathways that have been identified. There are at least 15 distinct groups of PLA2, clustered into four major groups, the cytosolic (cPLA2), the calcium-independent (iPLA2), the secreted (sPLA2), and a group that includes platelet-activating factor acetylhydrolase and related PLA2s. There is insufficient space to summarize the physiological role of all these lipases, so we will point out that two in particular, cPLA2α and iPLA2β have been shown to use different lipid mediators (arachidonic acid and LPS, respectively) to induce migration of monocytes towards monocyte chemoattractant-1.81

The other two pathways are less well defined particularly in neutrophils. Recent work by Van Haastert and colleagues revealed that cGMP and soluble guanylate cyclase function as important intracellular signalling molecules for chemotaxis, with cGMP being key at pseudopod and soluble guanylate cyclise-mediating uropod retraction.82 Finally, DOCK2/phosphatidic acid (PA) was reported as a new pathway for chemotaxis.83 PA localizes DOCK2 to the pseudopod to initiate chemotaxis.83 PIP3 also localizes DOCK2 to the pseudopod very early, whereas PA does it more slowly but independent of PIP3.51,83 We consistently see that fMLP-induced chemotaxis has a very early PIP3 component that is quickly replaced by a second PIP3-independent pathway and perhaps may involve PA.

5. Integrins and chemotaxis

In order to chemotax, migrating cells must adhere to the underlying substratum; the cells then gain the traction required to push their way through the tissue in which they are migrating. In human neutrophils, this adhesion comes predominantly from integrins. Neutrophil chemotaxis is largely dependent on the β2 family of integrins, as evidenced by patients with leucocyte adhesion deficiency type I (LAD-I). LAD-I patients have a genetic deficiency in CD18, the β2 integrin subunit, and their neutrophils are unable to be recruited from the vasculature or chemotax. Neutrophils express primarily two β2 integrins—LFA-1 (αLβ2) and Mac-1 (αMβ2). Some evidence has suggested that LFA-1 and Mac-1 play different roles during recruitment. Indeed, a study comparing neutrophil recruitment to TNF-α in LFA-1- and Mac-1-knockout mice showed that these two integrins play opposing roles. Neutrophil extravasation in response to TNF-α in LFA-1-knockout mice was dramatically decreased, whereas neutrophil extravasation in Mac-1-knockout mice was markedly increased, suggesting that Mac-1 may act as a ‘brake’ during this process.84 In contrast to the study described above, Mac-1 was found to be the predominant integrin involved in chemotaxis through synovial and dermal fibroblast barriers,85 hinting that the role of integrins may vary depending on the environment in which the cell is chemotaxing, or the chemoattractant to which the cell is responding. As already summarized for neutrophils in the vasculature, LFA-1 was necessary for adhesion, whereas Mac-1 was needed for subsequent crawling. Although there appears to be no role for LFA-1 in neutrophil crawling at least in muscle microvasculature, whether this is also the case in places like the brain microvasculature where shear forces are high and the neutrophil must hang on more avidly than in muscle microvasculature remains unknown. Moreover, in the case of monocyte crawling in skin, LFA-1 was critical.36

Only recently have details of the inside-out signalling pathways emerged. Inside-out signalling is defined as those events that induce conformational changes in the integrin leading to increased ligand-binding affinity (integrin activation) and clustering of integrins in the membrane, which together result in avidity modulation allowing cell attachment.86 The Rap GTPases have been implicated as major regulators of the inside-out pathway.87 The activation of Rap1 leads to the intracellular spatial redistribution of integrins to stimulate clustering which contributes to increased integrin avidity (Figure 2). Triggered by activated Rap1, RAPL and Mst1 associate with LFA-1 and rapidly relocate to the leading edge, resulting in the clustering of LFA-1 at immunological synapses.88,89 The constitutively active mutants of Rap1 increase the affinity and avidity of LFA-1 at the lymphocyte membrane.90 Furthermore, impaired activation of Rap1 has been documented in patients with leucocyte adhesion deficiency III (LAD-III).91 Although the leucocytes in these patients express normal levels of β integrins, their inside-out signalling is impaired and results in impaired leucocyte arrest and migration on vascular endothelium.92,93 Several guanine-exchange factors (GEFs) that activate Rap GTPases have been identified recently and one of these is activated by direct binding of calcium and diacylglycerols (DAG) called CalDAG-GEF1 (also known as RasGRP2).94 Rasgrp2−/− mice have impaired integrin-mediated adhesion in neutrophils,94 suggesting that CalDAG-GEF1 as a critical regulator of inside-out integrin activation in neutrophils. Another route of Rap1 activation is through protein kinase C (PKC), and similar to CalDAG-GEF1, is also responsive to calcium and DAG.95 Interestingly, PKC signalling does not always lead to Rap1 activation in all classes of integrins. Whereas LFA-1 activation is dependent on CalDAG-GEF1 and Rap1, α4β1 integrin was reported to be activated via a Rap1-independent PKC-mediated pathway.96

Figure 2

Integrin activation enhances avidity. (A) The β2 integrins of LFA-1 are normally in an inactive or non-ligand-binding state in which the integrin ectodomains are held in a bent or folded conformation when the leucocytes are in circulation. (B) Upon activation through stimulation with chemokines, the ‘inside-out’ signalling is induced such that the binding of intracellular Ca2+ and DAG to guanine-exchange factor CalDAG-GEF1 triggers the activation of Rap1. In turn, a conformational change of the cytoplasmic and transmembrane domains of the integrin relays to the ectodomain, increasing its ability to bind ICAM-1. (C) The activation of Rap1 leads to the rapid redistribution of integrins to the leading edge, resulting in the clustering of LFA-1 at immunological synapses and contributes to the increased avidity of the integrin.

Unlike the β2 integrins, the α4-integrin is expressed at low levels on human neutrophils, although its expression is increased during chronic inflammatory diseases97,98 and during acute systemic inflammation.99 Although α4-integrin expression has been demonstrated on neutrophils, its role in mediating the chemotaxis of neutrophils remains unexplored. The release of neutrophil progenitors from the bone marrow depends on the degradation of α4-integrins substratum (VCAM-1), suggesting that α4-integrin may act as a ‘brake’ on neutrophil chemotaxis.100 Similarly, α4-integrin keeps migrating B-cells from leaving the spleen marginal zone.101 In contrast, α4-integrin plays a central role in the recruitment of T-cells and eosinophils into sites of inflammation.102 Indeed, an anti-α4-integrin antibody has shown great promise in the treatment of a variety of autoimmune disorders.103,104 Although the mechanism of action is not well understood, it is believed that this antibody's therapeutic effect is due to its ability to inhibit T-cell infiltration into target tissues.105 α4-Integrin also appears to be involved in T-cell chemotaxis, with T-cell adhesion to fibronectin being partially dependent on the α4-integrin.106

Integrins have the potential to have a major regulatory role during chemotaxis, especially given their dual roles as both cell adhesion molecules and as signal transducing receptors. However, the function and regulation of integrins during neutrophil chemotaxis remain poorly understood. Controversies remain whether there are differences in integrin usage by different chemoattractants, or differences in integrin usage in different migratory environments. Given the central role these molecules play in neutrophil recruitment, it is necessary to further investigate the functions of these proteins in order to achieve a full understanding of the chemotactic process. Perhaps imaging the spatial localization and conformational changes of integrins over time upon various stimuli could help to address this.

6. Conclusion

Directed cell migration is essential in many important processes including lymphoid development, immuno-surveillance, and immune responses. Various studies have identified multiple signalling pathways that can mediate chemotaxis, often with great discrepancies in terms of the role and importance of these pathways. In addition, there are different forms of cross-talk between these pathways, ranging from inhibition, to serial activation, to parallel amplification of chemoattractant signals. Indeed, alternative molecular mechanisms of gradient sensing have tentatively been identified and these processes appear to be independent of the signalling pathways traditionally thought to mediate chemotaxis. Complicating matters further is the fact that the signalling pathways used for chemotaxis can vary depending on the environment in which the cell is chemotaxing. The identity of a chemoattractant, the steepness of the chemoattractant gradient, the makeup of the substratum upon which the cell is crawling, and even temporal changes in the chemoattractant gradient can affect how a cell migrates and which signalling pathways are required for chemotaxis. Understanding the dynamics and insights into these molecular regulators and their pathways would allow us to better understand, and perhaps intercede in, neutrophil-mediated inflammatory responses.

Conflict of interest: none declared.


This work was supported by Canadian Institutes for Health Research operating grant and grant group. C.H.Y.W is a fellow from Canadian Institutes for Health Research (CAG/CIHR/Janssen-Ortho). B.H. is a fellow from the Heart and Stroke Foundation of Canada. P.K. is an Alberta Heritage Foundation for Medical Research Scientist and the Snyder Chair in Critical Care Medicine.


View Abstract