Ca²⁺ signaling in airway epithelial cells facilitates leukocyte recruitment and transepithelial migration

Review on consequences of TLR2 signaling the activation of Ca2+-dependent proteases (calpains) and link to junctional proteins to facilitate PMN migration into the airway lumen.

Abstract

In airway cells, TLR2 stimulation by bacterial products activates Ca²⁺ fluxes that signal leukocyte recruitment to the lung and facilitates transepithelial migration into the airway lumen. TLR2 is apically displayed on airway cells, where it senses bacterial stimuli. Biochemical and genetic approaches demonstrate that TLR2 ligands stimulate release of Ca²⁺ from intracellular stores by activating TLR2 phosphorylation by c-Src and recruiting PI3K and PLCγ to affect Ca²⁺ release through IP3Rs. This Ca²⁺ release plays a pivotal role in signaling TLR2-dependent NF-κB activation and chemokine expression to recruit PMNs to the lung. In addition, TLR2-initiated Ca²⁺ release activates Ca²⁺-dependent proteases, calpains, which cleave the transmembrane proteins occludin and E-cadherin to promote PMN transmigration. This review highlights recent findings that demonstrate a central role for Ca²⁺ signaling in airway epithelial cells to induce proinflammatory gene transcription and to initiate junctional changes that accommodate transmigration of recruited PMNs.

Introduction

The airways are exposed continuously to microorganisms; however, healthy individuals maintain a pathogen-free environment in their lower respiratory tract. This is a testament to the efficient host defense system at the mucosal surface of the lung. Airway epithelial cells are positioned strategically to play a key role in host response by providing a physical and an immunological barrier to inhaled pathogens [1, 2]. The physical barrier mediated by epithelial cell-to-cell junctions prevents pathogens from invading, and the immunological barrier is achieved through epithelial cell detection of bacteria and rapid signaling to recruit phagocytes and clear the infection.

A central component in regulating the immunological and physical barriers of the airway epithelium is Ca²⁺. The participation of Ca²⁺ as a second messenger is vital to numerous physiological processes [3,4,5]. Cytosolic Ca²⁺ fluxes are involved in mediating responses to extracellular stimuli, such as the activation of G-protein-coupled receptors [3], TCRs [4, 6, 7], and BCRs [8]. The common respiratory bacterial pathogens, Pseudomonas aeruginosa and Staphylococcus aureus, activate 100 nM Ca²⁺ fluxes immediately upon contact with airway epithelial cells, and this Ca²⁺ signal initiates the activation of proinflammatory signaling events [9].

In this overview, we describe a central role for intracellular Ca²⁺ fluxes in controlling the physical and immunological barriers of the respiratory epithelium during bacterial infection. In airway cells, Ca²⁺ fluxes mediate the expression of proinflammatory cytokines and chemokines necessary to recruit leukocytes to the lung and also initiate modifications to the epithelial junctions to facilitate leukocyte transmigration into the airway lumen.

HOST—PATHOGEN INTERACTIONS IN THE AIRWAY

Airway epithelial cells contribute to the immunological barrier by detecting the presence of pathogens and secreting chemokines and cytokines (IL-8, IL-6, GM-CSF) to recruit and activate phagocytic cells to clear the invading organism [2]. Airway epithelial cells are protected by mucin and cilia and infrequently come in contact with intact bacteria. Instead, they detect bacterial components or PAMPs through PRRs that detect molecules broadly shared by pathogens [2]. Recognition of extracellular or endosomal PAMPs is mediated predominantly by the TLR, a family of type-I transmembrane proteins, which detects PAMPs through extracellular leucine-rich repeat motifs and mediates signaling through an intracellular TIR domain [10]. Each given tissue or cell expresses multiple TLRs [10]; in human airway epithelial cells, TLRs 1–10 are expressed [11]. Activation of signaling through TLRs results in recruitment of cytoplasmic adaptor molecules such as MyD88, TIR domain-containing adaptor protein (also called Mal), Trif, and Trif-related adaptor molecule. The adaptors activate downstream molecules including protein kinases (IRAK1, IRAK4, TRAF family member-associated NF-κB activator-binding kinase 1, and IKK) that amplify the signal and activate proinflammatory transcription factors, which induce the expression of genes involved in the inflammatory response. In airway cells, activation of NF-κB, a major proinflammatory transcription factor, rapidly signals the expression of IL-8, a chemokine that recruits PMNs [12], which are the first immune cells recruited in response to infection [13, 14]. Interestingly, IL-8 was shown to recruit PMNs from the bloodstream to the basolateral surface of airway epithelial cells; however, it was not involved in mediating PMN transmigration further into the airway lumen [15]. In this regard, the secretion of an arachidonic acid metabolite, hepoxilin A3, by airway epithelial cells has been proposed to direct PMN migration across airway epithelial cells [15]. PMNs are critical in eradicating respiratory pathogens; however, excessive numbers of PMNs in the airway can lead to tissue damage and compromised lung function [2, 16].

In addition to signaling the presence of bacteria to immune cells, epithelial cells form a physical barrier through their cell-to-cell junctions, which prevents pathogen invasion. This barrier is established by the apical junctional complex composed of the tight and adherens junctions. As the epithelial tight junctions mainly function to create a barrier to maintain sterility in the lung, they are more substantial and are 10 times less leaky than endothelial junctions [17]. In the respiratory tract, PMN migration across endothelial cells has been studied in some detail [18, 19]. However, unlike crossing the endothelium, where movement is in the apical-to-basal direction, PMNs migrate in a basal-to-apical direction as they cross the epithelium. This is particularly important in understanding the molecules that may be involved in adhesion and migration. For example, ICAM-1 is responsible for allowing PMNs to adhere to endothelial cells and initiate the steps in transmigration [20]. ICAM-1 is not expressed at the basolateral surface of epithelial cells and is expressed rather abundantly at the apical surface [20]. Therefore, it is unlikely that ICAM-1 plays a role in PMN transepithelial migration but rather, may be involved in retaining PMNs at specific locations in the airway lumen. PMN transepithelial migration occurs independent of a number of molecules, CDlla/CD18, selectins, CD31, and ICAM-1, which are required for transendothelial migration [21].

Much of what is known about neutrophil transepithelial migration comes from in vitro studies using intestinal epithelial monolayers [21]. PMN infiltration into the intestinal mucosa is a central feature of active inflammatory bowel diseases, such as ulcerative colitis and Crohn’s disease [22]. Direct targeting of junctional proteins is one mechanism to facilitate PMN migration. Biopsies of colonic mucosa from patients with ulcerative colitis have dramatic down-regulation of the tight junction protein occludin in regions of actively transmigrating PMNs [23]. Paracellular permeability across epithelial and endothelial cells is also enhanced by disassembly of the junctions upon contraction of the perijunctional actin-myosin ring induced by myosin light chain phosphorylation, which by MLCK or Rho kinase, has been demonstrated in cells exposed to enteropathogenic Escherichia coli [24] or stimulated by basolateral contact with PMNs [21] to facilitate transmigration of bacteria and PMNs.

PMNs must migrate across a complex network of tight and adherens junction proteins to clear the infection in the airway lumen [18, 19, 23]. However, disruption of the junctions in the presence of pathogens in the airway lumen could facilitate bacterial invasion and disseminated infection. Exactly how PMNs are able to migrate between respiratory epithelial cells without breaching the barrier provided by tight junctions is not well understood. We reasoned that signaling through receptors that detect bacteria may induce chemokine expression to recruit PMNs and changes to the epithelial barrier to facilitate PMN paracellular migration.

TLR2-MEDIATED Ca²⁺ FLUXES SIGNAL PROINFLAMMATORY RESPONSES

There is substantial literature demonstrating that various bacterial products induce Ca²⁺ fluxes and proinflammatory gene expression in mucosal cells [25]. Several different mechanisms have been implicated, including direct effects of bacterial toxins [26] or activation of purinergic signaling by bacterial adhesins that bind asialoGM1 [27]. The PRR TLR2 initiates the rapid release of intracellular Ca²⁺ in airway cells [28]. TLR2 is well positioned for initiating signaling, as these receptors are mobilized to the surface of airway cells to rapidly recognize PAMPs in the airway lumen [29, 30]. TLR2 recognizes a variety of microbial components including lipoproteins/lipopeptides, lipoteicoic acid, lipoarabinomannan, and zymosans [29, 31]. Broad ligand specificity of TLR2 is attributed to its ability to dimerize with TLR1, TLR6, or asialoGM1, a glycolipid receptor [29, 32]. At the apical surface of airway cells, TLR2 is associated with necessary downstream signaling molecules including MyD88, IRAK, TNFR-associated factor 6, and c-Src in caveolin-1-containing lipid rafts called caveolae [30, 33]. Efficient downstream signaling can be achieved when receptors and their signaling molecules are in close proximity to each other within these signaling domains.

The canonical Ca²⁺ release pathway involves PLCγ hydrolysis of phosphatidylinositol 4,5 bisphosphate to IP3 and DAG. IP3 binds to IP3Rs expressed on the ER and signals the release of Ca²⁺ from the ER, and DAG regulates the activation of members of the PKC family [34].

TLR2 phosphorylation by c-Src

The mechanism of TLR2-induced Ca²⁺ fluxes involves TLR2 phosphorylation by c-Src [28], which signals recruitment and activation of PI3K and PLCγ to affect Ca²⁺ release through IP3Rs [28] (Fig. 1). Phosphorylation of tyrosine 616 and 761 on the cytoplasmic tail of TLR2 was observed initially in cells treated with S. aureus [35]. The involvement of Src family kinases in the phosphorylation of TLR2 is consistent with several published studies demonstrating the requirement for Src activity in cellular responses to bacterial ligands [36, 37], for cyclooxygenase-2 activation by Helicobacter pylori [38], and for MUC-2 activation [36] and IL-8 expression in response to P. aeruginosa ligands [30]. Biochemical inhibitors, PP1 or PP2 for Src family kinases, are often used to demonstrate a role for Src in TLR2 phosphorylation. A specific role for c-Src and not other Src family kinases was demonstrated in more recent work using c-Src siRNA to inhibit TLR2-dependent signaling [39, 40]. c-Src is the likely TLR2 kinase, as it was also shown to coimmunoprecipitate with TLR2 in airway cells following bacterial stimulation [30]. c-Src and its family members are ordinarily maintained in equilibrium between inactive and primed states by a balance of negative regulatory kinase C-Src tyrosine kinase and its counteracting tyrosine phosphatase(s), both of which act on a regulatory C-terminal tyrosine [41]. We speculate that ligand-binding promotes conformational changes to TLR2 that promote phosphorylation by the available active c-Src.

Figure 1.

TLR2-mediated Ca²⁺ fluxes signal proinflammatory responses in airway epithelial cells. TLR2 ligands stimulate release of Ca²⁺ by signaling the phosphorylation of TLR2 by c-Src, recruiting PI3K and activating PLCγ. TLR2 or thapsigargin-mediated Ca²⁺ fluxes signal NF-κB activation and IL-8 expression. Silencing TLR2 by siRNA, blocking TLR2 by overexpression of a TLR2 tyrosine mutant, or using biochemical inhibitors of Src (PP1, PP2), PI3K (LY294002), or PLCγ (U73122) blocks intracellular Ca²⁺ release, NF-κB activation, and IL-8 expression. Chelating intracellular Ca²⁺ using BAPTA/AM also blocks NF-κB activation and IL-8 expression.

TLR2 signals PI3K and PLCγ activation

In addition to or in concert with the known TLR pathway involving MyD88, TLR2 directs the activation of class 1A PI3K, which contributes to NF-κB signaling [35, 42]. In response to S. aureus or P. aeruginosa, PI3K is recruited to phosphorylated tyrosines 616 and 761 on the cytoplasmic tail of TLR2 [28, 35]. PI3K catalyzes the phosphorylation of phosphotidylinositides in the plasma membrane, which then become docking sites for PH domain-containing proteins including PLCγ and Akt [43]. The importance of PI3K and PLCγ in immune signaling is well established [44]. Ca²⁺-dependent responses to E. coli and Salmonella typhimurium also implicate PI3K or PLCγ, as might be expected [45, 46]. PI3Ks are involved in mediating signals generated by bacterial recognition of various TLRs besides TLR2 and are involved in pro- and anti-inflammatory pathways [47,48,49]. As the link between PI3K and specific TLRs is suggested to be tissue-specific, our studies have focused on TLR2 expression in airway cells. However, murine peritoneal macrophages similarly generate Ca²⁺ fluxes following stimulation with TLR2 ligands [28]. Therefore, the recruitment of PI3K and PLCγ to TLR2 to induce intracellular Ca²⁺ release may be a generally conserved signaling mechanism.

TLR2 signals activation of NF-κB and expression of IL-8

It is well established that Ca²⁺ transients regulate gene transcription and expression [5]. The generation of Ca²⁺ fluxes in airway cells by intact bacteria and flagella stimulates transcription of NF-κB-dependent genes, MUC-2 and IL-8 [30, 50]. Increases in cytosolic Ca²⁺ induced by thapsigargin, a Sarco/ER calcium ATPase pump inhibitor, are sufficient to activate NF-κB and induce IL-8 expression in airway cells [28]. NF-κB transcription factors are dimers of two out of five subunits—p65 (RelA), c-Rel, RelB, p50 (NF-κB1), and p52. In the absence of stimuli, most NF-κB dimers remain in the cytosol bound to IκBs, which are specific inhibitory proteins. Proinflammatory stimulation downstream of TLR signaling can activate NF-κB through IKK-dependent phosphorylation and degradation of IκB [51]. The exact Ca²⁺-dependent component of TLR2-mediated NF-κB activation and IL-8 expression is still unknown. However, conventional PKCs (PKCα, PKCβ, and PKCγ) and calmodulin-dependent kinases II are likely candidates, as they are activated by Ca²⁺ and have been shown to be involved in NF-κB activation [52, 53]. Recent studies have revealed that NF-κB activity is additionally regulated by Ser 276 phosphorylation of the p65 subunit by a variety of kinases [35, 54,55,56]. Most notably, PI3K-mediated activation of Akt, a PH domain containing kinase, was shown to phosphorylate the p65 subunit and optimize NF-κB activation in monocytes stimulated with S. aureus [35].

IL-8 is a CXC chemokine, structurally classified based on the NH₂-proximal cysteines, which are separated by an amino acid (X) [57,58,59]. Endothelial cells as well as airway epithelial cells rapidly induce IL-8 expression after bacterial stimulation [2, 60, 61]. IL-8 is expressed in response to microbial components IL-1β and TNF-α [62]. The anti-inflammatory cytokine IL-10 is a potent inhibitor of IL-8 synthesis [63]. The regulation of IL-8 expression by NF-κB is well documented [62, 64]. Nucleotides –1 to –133 within the 5′-flanking region of the IL-8 gene are essential and sufficient for transcriptional regulation of the gene and contain a NF-κB element that is required for activation in all cell types studied [12, 62, 64]. The core IL-8 promoter also contains AP-1- and C/EBP-binding sites [62]. The latter two sites are dispensable for transcriptional activation in some cells but contribute to activation in others. Thus, unlike the NF-κB site, the AP-1 and C/EBP sites are not essential for induction but are required for maximal gene expression [62]. Once secreted, IL-8 binds to two PMN receptors, CXCR1 and CXCR2, which are 7-transmembrane G-protein-coupled receptors that signal PMN chemotaxis and firm adhesion to the endothelium [59, 61]. Other CXCL chemokines, such as growth-related gene product-α and -β (CXCL1 and -2), epithelial cell-derived and neutrophil-activating 78 aa peptide (CXCL5), and granulocyte chemotactic protein-2 (CXCL6), are controlled by NF-κB activation and are chemotactic for PMNs, but their induction has not yet been tested in TLR2-stimulated airway epithelial cells.

Rodents do not express IL-8 and instead, produce CINCs. KC (CINC-1) and MIP-2 (CINC-3) are critical to inflammatory responses in mice and rats [65]. KC and MIP-2 are involved in pulmonary PMN responses during bacterial pneumonia, ozone, and silica dust inhalation and immune-complex deposition [66,67,68,69,70,71].

TLR2-mediated Ca²⁺ fluxes are communicated from cell to cell

Ca²⁺ fluxes coordinate signaling in conductive tissues of the muscular and nervous system [72, 73]. In airway epithelial cells, Ca²⁺ fluxes travel from cell to cell and can also serve to amplify Ca²⁺-activated proinflammatory signals. In human airway cells, gap junction channels were found to provide a conduit for the movement of Ca²⁺ from cell to cell. Gap junctions are comprised of Cx monomers, which oligomerize and associate with adjacent oligomers to form a lateral pore between cells. Primary human airway epithelial cells express nine Cx isoforms: Cx26, Cx30, Cx30.3, Cx31, Cx31.1, Cx32, Cx37, Cx40, and Cx43, which exhibit unique conductive and regulatory properties [74, 75]. In response to TLR2 stimulation, gap junctions function to transiently amplify proinflammatory signaling by communicating Ca²⁺ fluxes from stimulated to adjacent, nonstimulated cells, thus increasing epithelial IL-8 production [76]. Cx43 heterozygous mice displayed reduced PMN influx in response to respiratory LPS exposure [77], further demonstrating the importance of this mechanism of communication in inflammation. However, as unregulated signaling would result in excessive inflammation, the subsequent closing of Cx43 channels is initiated by c-Src-mediated phosphorylation [76]. Similarly, airway epithelial cells treated with TNF-α signaled c-Src-dependent closure of Cx43 gap junction channels [78]. A defect in this closing of Cx43 was observed in cystic fibrosis airway epithelial cells, which have exaggerated inflammatory responses [78].

Ca²⁺ FLUXES ALTER THE APICAL JUNCTIONAL COMPLEX TO FACILITATE PMN MIGRATION

As a major function of TLR2-dependent Ca²⁺ signaling in airway cells is stimulating chemokine expression and the resultant PMN recruitment, we postulated that the same signaling event would modify the cell junctions to facilitate migration of PMNs across the epithelial barrier. Following bacterial exposure, PMNs accumulate rapidly in the airway lumen [13, 14], suggesting that changes in the permeability characteristics of the paracellular junctions are an immediate consequence of the epithelial proinflammatory signaling cascade. PMNs have been implicated in the pathogenesis of many inflammatory lung diseases, including cystic fibrosis, acute respiratory distress syndrome, chronic obstructive pulmonary disease, and asthma. To access the airway lumen, PMNs must pass through distinct tissue compartments. PMNs exit the bloodstream by migrating across the endothelial cells. Once in the interstitial space, PMNs travel through the extracellular matrix and interact with the basolateral surface of the epithelium. Finally, they migrate between the epithelial cells to reach the airway lumen [13, 17]. Endothelial Ca²⁺ fluxes promote transendothelial migration of PMNs by opening their intercellular junctions [79]. Ca²⁺ flux is not required for PMN adhesion but is required for PMN migration across HUVEC monolayers [79]. In endothelial cells, extracellular Ca²⁺ entry has also been implicated in modifying cell-to-cell junctions. Ca²⁺ release from TRPCs was shown to alter permeability across endothelial junctions by signaling cytoskeletal reorganization through the activation of Ca²⁺-dependent MLCK [80, 81]. Ca²⁺entry through TRPC1 was shown to promote rearrangement of the actin cytoskeleton and thus increase endothelial permeability [80]. In addition, endothelial cells from TRPC4 knockout mice demonstrated a significantly less dramatic decrease in transendothelial resistance compared with wild-type cells in response to thrombin [82]. Whether TRPCs are activated in response to TLR signaling is unknown. However, in airway epithelial cells, inhibiting Ca²⁺ entry through L-type Ca²⁺ channels or chelating extracellular Ca²⁺ did not affect TLR2 signaling [9, 28], suggesting that intracellular Ca²⁺ release is responsible for TLR2-dependent responses.

In airway epithelial cells, TLR2 stimulation transiently raised levels of cytosolic Ca²⁺ [83] and increased the accessibility of biotin at the cell-to-cell junctions [84]. Consistent with the increase in biotin at the cell-to-cell junctions, the expected “chicken wire” distribution of transmembrane junction proteins, occludin and E-cadherin, was altered substantially following 4 h exposure of airway cells to P. aeruginosa or the TLR2 agonist P3C [84]. Yet, despite loss of occludin and E-cadherin at the cell borders, there was no concomitant decrease in the transepithelial resistance measured across the monolayers over this time period nor was there an increase in permeability to 10,000 MW fluorescent dextran or to bacteria across the paracellular space, thereby indicating that the barrier function of the monolayer remained intact [84]. Therefore, we propose that TLR2-dependent intracellular Ca²⁺ fluxes cause a subtle change in the cell-to-cell junctions to facilitate PMN migration without disrupting the barrier function of the epithelium. Recent work by Chin et al. [85] demonstrates more dramatic junctional changes after direct contact of the PMN with the basolateral surface of intestinal epithelial cells. Serine proteases released by PMNs activated epithelial protease-activated receptors to signal myosin light chain phosphorylation, decrease transepithelial resistance, and increase epithelial permeability to facilitate PMN transepithelial migration [85]. Taken together, TLR2-mediated signaling recruits PMNs to the site of infection and initiates changes in the junctions that can be altered further by direct contact of the epithelial cells with the PMNs.

PROTEASES TARGET MEMBERS OF THE APICAL JUNCTIONAL COMPLEX

The movement of fluid, ions, macromolecules, and inflammatory cells across airway epithelium depends on the integrity of its apical junctional complex composed of the tight and adherens junctions. Previous studies suggest involvement of occludin, a component of the tight junction, and E-cadherin, a component of the adherens junction, in immune cell diapedesis across a monolayer of cells [23, 86, 87]. The importance of proteases, especially the MMP, in inflammatory processes is well established [88]. MMP7 targets several epithelial components [89], including E-cadherin, and contributes to the shedding of its ectodomain and endocytosis [90]. MMP9 facilitates translocation of PMNs from endovascular spaces causing MMP9^−/− mice to have a defect in PMN transepithelial migration in response to infection [91]. Occludin proteolysis is also reported to be a component of PMN-dependent inflammation [92]. However, the observation that occludin proteolysis remains unaltered in the MMP9^−/− mice [91] indicated that additional protease(s) must target this tight junction protein. The failure of a general MMP inhibitor to block occludin or E-cadherin cleavage in airway cells, as we demonstrate [84], also suggested the involvement of a different protease. As TLR2 signaling initiates Ca²⁺ fluxes, the Ca²⁺-dependent calpain proteases were likely candidates to target epithelial junction proteins. TLR2 stimulation of airway epithelial cells signals a rapid increase in calpain activation consistent with the rapid Ca²⁺ fluxes observed [28, 84]. This response was blocked in cells treated with a calpain inhibitor calpeptin or cells expressing TLR2 siRNA oligonucleotides [84] (see Fig. 3).

Figure 2.

Calpains cleave occludin and E-cadherin. (A) A schematic representation of the 80-kD large subunit of the classical calpains is shown. The N-terminal Domain I undergoes intermolecular autolysis upon calpain activation. Ca²⁺-binding sites and catalytic activity are contained within Domain II. Domain III comprises a C2-like domain that contains sites for phospholipid binding. Five consecutive EF-hand motifs make up Domain IV and contribute to Ca²⁺ binding of the large subunits and dimerization with the small regulatory subunit, CSS1/Capn4. (B) Calpains are positively regulated by Ca²⁺-binding, Domain I autolysis, phosphorylation, and phospholipid binding and inhibited by calpastatin and autolysis. The N-terminus (N) of occludin and the C-terminus (C) of E-cadherin are targets for calpain cleavage.

Figure 3.

TLR2-mediated Ca²⁺ fluxes activate calpains, which cleave junctional proteins and facilitate PMN transmigration. Bacterial detection by TLR2 mediates a Ca²⁺ release that signals IL-8 expression to recruit PMNs and also initiates modifications to the epithelial junctions to facilitate PMN transmigration. TLR2-mediated Ca²⁺ fluxes activate calpains that target the transmembrane junctional proteins, occludin and E-cadherin. Calpain activation is inhibited in TLR2 siRNA-expressing cells. Occludin and E-cadherin cleavage is blocked in airway cells expressing a TLR2 tyrosine mutant or calpain siRNA oligonucleotides. Occludin and E-cadherin cleavage and PMN migration across airway cells are blocked by calpain inhibitors.

Calpains were found throughout the cytoplasm of unstimulated airway epithelial cells, whereas the transmembrane junctional proteins, occludin and E-cadherin, were concentrated at the plasma membrane [84]. By 1 h after bacterial or P3C exposure, calpains assumed a peripheral distribution colocalizing with membrane-associated occludin and E-cadherin [84]. This membrane localization of calpain allows it to be in close proximity to its junctional substrates and is also consistent with the role of membrane phospholipid-binding in regulating calpain activation [93,94,95].

Calpains

Calpains are Ca²⁺-dependent cysteine proteases that are known to target junctional proteins [96, 97]. They are linked to numerous cellular processes including motility, apoptosis, and inflammation [98, 99]. Strategies to target calpain activity in inflammatory diseases such as arthritis have been successful in mice [100]. The calpain family of proteases contains 16 genes [98]. Of these, 14 encode for a large 80-kD subunit that contains a catalytic domain; the other two encode smaller 28 kD regulatory proteins that associate with the larger subunit to form a heterodimeric protease [98]. Although most calpains are expressed in specific tissues, the typical or conventional calpains, μ-calpain (calpain 1) and m-calpain (calpain 2), along with their regulatory subunit CSS1 (calpain 4), are ubiquitously expressed and abundant in the airway [101]. The involvement of conventional calpains in regulating the cytoskeleton of the lung was reported over a decade ago in response to phorbol esters [102].

The crystal structure of the conventional calpain molecules reveals four distinct domains [103,104,105]. Domain I contains a 19-residue N-terminal domain that is cleaved intermolecularly upon activation. Domain II comprises the catalytic active site and also the Ca²⁺-binding sites that act as the Ca²⁺ switch to align the catalytic triad Cys 105, His 262, and Asn286 [105, 106]. Domain III is a regulatory domain, which contains phosphorylation sites and a C2 domain first identified in PKC as a region that binds phospholipids in a Ca²⁺-dependent manner [93]. Domain IV contains five EF-hand Ca²⁺-binding domains, important in forming the heterodimer with the regulatory subunit CSS1/Capn4 [98] (Fig. 2A).

μ-Calpain (calpain 1) and m-calpain (calpain 2) were named for the micromolar and millimolar concentrations of Ca²⁺ required to activate them in vitro, respectively [98]. However, such high levels of Ca²⁺ do not exist in living cells, and additional regulatory mechanisms to eliminate or lower the Ca²⁺ requirement have been identified. Phosphorylation at several sites activates calpains in the absence of increased Ca²⁺ [107]. Autolysis of the N-terminal fragment in Domain I and membrane phospholipid binding to Domain III were shown to lower the requirement for Ca²⁺ [108]. As excessive protease activity could be detrimental to cells, calpain activity is negatively regulated by rapid autolysis or binding to calpastatin, an endogenous inhibitor that associates with activated Ca²⁺-bound calpain [109, 110] (Fig. 2B).

In general, calpains cleave proteins at a limited number of sites and produce large polypeptide fragments rather than small peptides or amino acids [98]. This would suggest that calpain cleavage regulates the function of its targets as opposed to degrading them. Calpain 1 and calpain 2 appear to have similar, if not identical, substrate specificity [98]. There is no putative amino acid sequence that defines a calpain cleavage site, but rather, subsite specificity is recognized by conformation of the polypeptide chain [98]. Calpains targeted and cleaved occludin and E-cadherin directly but did not target all transmembrane junctional proteins, as claudin-1 and junctional adherence molecule 1 were not cleaved by calpains [84]. The involvement of calpain-dependent cleavage in vivo was verified by identifying occludin and E-cadherin cleavage products in whole lung lysates of infected wild-type mice but not in uninfected controls and substantially decreased cleavage products in the calpeptin-treated animals [84].

Occludin

The tight junction is the most apical component of the junctional complex and functions as the major paracellular barrier and “fence” separating apical from basolateral domains. These functions of tight junctions are critical for epithelial and endothelial cells to establish distinct tissue compartments and maintain homeostasis [111,112,113]. Occludin, the first tight junction protein identified, is a tetraspan membrane protein with two extracellular loops: a short intracellular turn and N- and C-terminal cytoplasmic domains. The extracellular loops are thought to associate with corresponding occludin loops on adjacent cells to regulate paracellular permeability and cell adhesion [114]. Overexpression experiments using full-length and mutated occludin in MDCK cells or Xenopus cells, as well as a study using synthetic peptides corresponding to the extracellular loops of occludin, suggest a role of occludin in the barrier and fence function of tight junctions [115,116,117]. However, occludin null mice showed no abnormalities in the structure of epithelial tight junctions, although they did display inflammation and hyperplasia of the gastric mucosa [118, 119]. Other tight junction proteins, most notably, members of the claudin family, have been implicated in maintaining the barrier properties of the junctions [120]. Although occludin does not appear to be essential in maintaining the tight junctions, it is thought to have a critical regulatory function through its interaction with proteins that form the junctional plaque [114, 121]. Occludin is linked to the actin cytoskeleton at the tight junctions through its C-terminal interaction with ZO-1, a major scaffolding molecule of the tight and adherens junctions [122]. Dissociation of ZO-1 and occludin disassembles the tight junction complex [123, 124].

Proteolytic cleavage of occludin disrupts barrier properties of the epithelial junctions. Occludin is a 60-kD protein, which is cleaved to products that range from 20 kD to 55 kD. During apoptosis, caspase 3 specifically cleaves occludin C-terminal to Asp320, generating a 31-kD fragment [125]. Several studies demonstrate proteolytic cleavage of occludin in epithelial cells exposed to bacteria (e.g., Vibrio cholerae, H. pylori, and Burkholderia cepacia), dust mite allergen (Der p 1) and ATP [114, 126, 127]. A role for calpains in occludin cleavage was demonstrated in cervical epithelial cells, where a biochemical inhibitor of calpains, N-acetyl-leu-leu-Nle-CHO, blocked ATP or diacylglyceride-induced occludin cleavage [126].

In airway epithelial cells, calpain cleavage of occludin occurs on the N-terminal tail of occludin to generate a 45-kD cleavage fragment [84] (Fig. 2B). This is consistent with previous studies, which have shown that expressing occludin mutants with modified N-terminal cytoplasmic domains up-regulated PMN migration, whereas deletion of the C-terminal cytoplasmic domain did not have an effect [86]. In addition, expression of an occludin N-terminal truncation mutant decreases barrier function in murine epithelial cells [128], consistent with our mapping data. TLR2-dependent occludin cleavage was blocked in cells expressing TLR2 siRNA and cells treated with the intracellular Ca²⁺ chelator, BAPTA/AM, or the calpain inhibitor, calpeptin [84]. To further confirm the role of calpains in occludin cleavage, calpain 1 and 2 expression was silenced by siRNA. The occludin cleavage product was detected in cells expressing scrambled oligos in response to P3C stimulation but not in cells expressing calpain 1 and 2 siRNA [84]. Knockdown of calpain 1 or calpain 2 individually was not sufficient to block occludin cleavage in response to P3C stimulation [84].

E-cadherin

Adherens junctions link membrane and cytoskeletal components at discrete contact sites and are required for basic cell–cell adhesion. Adhesive binding is mediated by cadherins, a family of type I single transmembrane-spanning glycoproteins that dimerize and form extracellular Ca²⁺-dependent interactions with identical molecules on the surfaces of adjacent cells [129]. Mutant vascular endothelial-cadherin lacking the extracellular domain results in impaired PMN migration in response to chemotactic stimuli [130]. The cadherin cytoplasmic domain controls structural and signaling activities required for adhesion through its association with three distinct members of the catenin family: β-catenin, α-catenin, and p120ctn [113, 129, 131]. These catenins work together to protect cadherins from proteolytic degradation and link the cadherin cytoplasmic domain to actin and myosin filaments [113, 129, 132]. Interactions between cadherins and the small GTPases that regulate actin polymerization affect paracellular permeability in endothelial cells [133].

E-cadherin, the best-characterized cadherin, is expressed specifically in epithelial cells [113, 129]. E-cadherin null mouse embryos do not survive past the blastocyte stage, demonstrating the importance of E-cadherin in early development [134,135,136]. Keratinocyte-specific deletion of E-cadherin in mice results in hyperproliferation of epidermal basal cells and impaired terminal differentiation, consistent with the observed decrease in E-cadherin in many human cancers [135]. In response to various stimuli, E-cadherin undergoes endocytosis as part of the dynamic process of membrane homeostasis [129, 137]. E-cadherin trafficking has been studied intensively and plays a central role in cellular growth and development, maintenance of epithelial polarity, and Wnt signaling [129]. E-cadherin is bound indirectly to the cytoskeleton through a linkage of its intracellular C-terminus to β-catenin, which in turn, is linked to α-catenin, binding to the actin cytoskeleton [129]. E-cadherin is involved in the Wnt signaling pathway through its association with β-catenin, a key component of this pathway [138].

E-cadherin cleavage can also signal the dissociation of β-catenin and E-cadherin and lead to accumulation of β-catenin in the cytoplasm to promote Wnt signaling and subsequent transcription of target genes [139]. E-cadherin is a target of proteases, such as MMP7 [140], TNF-α-converting enzyme (ADAM 17) [141], and ADAM 10 [142], which mediates the shedding of the extracellular domain of E-cadherin in damaged and apoptotic cells.

Airway epithelial cells treated with P3C or P. aeruginosa induce calpain cleavage of the intracellular domain of E-cadherin to generate a 100-kD E-cadherin cleavage product, which is inhibited in the presence of calpeptin [84] (Fig. 2B). Treatment of cells with thapsigargin to elevate cytosolic Ca²⁺ also induced colocalization of calpains with E-cadherin and generated the 100-kD E-cadherin cleavage product [84]. In prostate and mammary epithelial cells, calpain activation efficiently generated the same 100-kD cleavage fragment of E-cadherin, also found in abundance in metastatic prostate cancer [143]. The calpain cleavage site on E-cadherin was mapped to residues 782–787 in the C-terminal domain, directly upstream of β-catenin- and γ-catenin-binding domains [144]. These data suggest that calpains are involved in disassociation of the E-cadherin catenins complex, which signals loss of E-cadherin at the junctions.

CALPAIN ACTIVATION FACILITATES PMN TRANSMIGRATION

A major function of TLR2 signaling is stimulating chemokine expression and signaling PMN recruitment. The Ca²⁺ fluxes that initiate TLR2-NF-κB-IL-8 signaling also stimulate calpain activity [84] (Fig. 3). In response to P. aeruginosa or a synthetic TLR2 agonist, PMN transepithelial migration was inhibited by the calpain inhibitor calpeptin, which did not inhibit the migration of PMNs directly across a porous transwell in response to bacteria or a chemokine gradient [84]. In resting PMNs, calpains are constitutively active and block the ability of PMNs to migrate, and in fact, calpain inhibition was shown to promote PMN movement by initiating MAPK and Rac GTPase activation [145, 146]. Thus, targeting epithelial, and not PMN, calpain activity is responsible for inhibiting PMN transmigration.

Although TLR2-mediated Ca²⁺ fluxes and subsequent calpain activation are rapid and observed within seconds or minutes [28, 84], cleavage of junctional proteins and PMN transmigration is detected after 1 or 4 h stimulation [84]. This discrepancy in kinetics may be a result of the cells’ ability to tightly control the levels of cytosolic Ca²⁺ by uptake into internal stores, extracellular release through efflux pumps, or chelation by Ca²⁺-buffering molecules [147, 148]. Additionally, activated calpain proteases are autolysed rapidly or inhibited by calpastatin [108] and may require further protein synthesis to achieve enough active protease to detect downstream responses. Furthermore, limitations in the detection of cleavage fragments may also contribute to the discrepancy in kinetics, as the initial generation of these fragments may also be degraded rapidly.

The biological importance of TLR2-induced calpain activity in vivo was tested in neonatal and adult murine models of airway infection. Compared with untreated controls, calpeptin-treated mice had 37% fewer PMNs recruited in the whole lung and 90% fewer PMNs recruited into the airway lumen, suggesting a more pronounced effect on the epithelial junctions [84]. In contrast to in vivo models of intestinal inflammation, which use organisms or chemical agents that induce significant epithelial injury [21], intranasal P. aeruginosa treatment in our murine model of infection is cleared readily with minimal damage to the lungs. The colon, unlike the airway, is colonized with commensal organisms and is hyporesponsive to TLR2 and TLR4 stimulation and requires epithelial damage to facilitate PMN recruitment [52]. In contrast, the normally sterile airway is hyperresponsive to pathogenic insult, and epithelial integrity is maintained, preserving lung function, and specific junctional changes facilitate PMN recruitment [84]. Finally, the rapid junctional changes and subsequent PMN recruitment into the airway in response to bacteria make it a favorable model to study PMN transepithelial migration.

CONCLUDING REMARKS

The participation of Ca²⁺ as a second messenger is vital to numerous physiological processes including the epithelial response to bacteria [3,4,5]. Ligation of TLR2 stimulates Ca²⁺ fluxes important in NF-κB activation as well as calpain activity.

TLR2-mediated Ca²⁺ fluxes signal changes in epithelial junctions to accommodate PMN egress to the airway luman. There is likely TLR2-independent signaling involved in facilitating PMN transmigration as well. TLR5, which activates Ca²⁺ fluxes in airway cells in response to flagella [9], may also contribute to PMN transmigration.

TLR2 activation in the airway epithelium coordinates the afferent and efferent limbs of the initial inflammatory response by signaling intracellular Ca²⁺ release. Not only do the airway epithelial cells produce chemokines to direct PMN recruitment, but they also use the same signaling cascade to modulate the cell junctions to accommodate PMN egress without breaching the epithelial barrier. Neither a decrease in transepithelial resistance nor an increase in bacterial invasion across the epithelium was observed in response to this Ca²⁺-dependent signaling event. Therefore, this pathway may provide a useful pharmacological target in pulmonary infection to selectively limit PMN recruitment into the lung, without entirely compromising host defenses to bacterial infection.