Second messenger systems

A variety of intracellular messengers have been implicated in regulation of intercellular junctions. In some instances, a link between extracellular permeability regulators and second messenger systems has been established.

G’Proteins

G-proteins have been implicated in junctional regulation, although the classes which are involved have still to be defined. Thrombin, bradykinin and histamine receptors are all coupled to G-proteins (Lum and Malik, 1994). These pathways can activate PLCp, leading to phosphoinositide turnover. Pertussis toxin, which inactivates certain G-protein subunits, increases the permeability of bovine pulmonary endothelial cells (Patterson et al., 1995)

Phospholipase C

Phospholipase C (PLC) can affect junctions, via interaction with other second messenger systems. Hydrolysis of both PIP and PI by PLC generates diacylglycerol, an activator of protein kinase C. PLC-mediated hydrolysis of PIPj generates IP3, in

addition to diacylgycerol. IP3 increases permeability of pulmonary endothelial cell

layers, perhaps via the IP3*mediated release of Ca^^ from intracellular stores (Patton et

al., 1991). Thus, PLC could affect intercellular junctions by activating both the PKC and Ca^^ second messenger systems (see below).

Histamine, thrombin and H^O^ promote both a transient rise in intracellular Ca^^ and a concomitant increase in endothelial permeability (Lum et al., 1989; Ehringer et

aL, 1996). Polymorphonuclear neutrophils (PMNs) migrating across human umbihcal vein endothelial cell monolayers cause similar Ca^^ and permeability increases. Both the increase in permeability and PMN migration is blocked if intracellular Ca^^ is clamped at a constant low level (Huang et a l, 1993). In T84 epithelial cell monolayers, use of Ca^^ ionophores to raise intracellular Ca^^ leads to an increase in the permeability of the monolayer (Tai et a l, 1996).

Protein kinase C

Another important intracellular messenger is protein kinase C (PKC), and in some cases the effects of Ca^^ on paracellular permeability can be attributed to the activation of Ca^^-dependent PKC isoforms. The increase in endothelial permeability in response to thrombin, bradykinin and HjOj can be blocked, or attenuated by inhibition of PKC (Johnson et al., 1989; Lynch et at., 1990; Lum and Malik, 1994). Unlike thrombin and H^O^, the bradykinin-induced increase in permeability does not require Ca^"" (Ehringer et at., 1996). It is possible that the effect of bradykinin is mediated via phospholipase D-mediated generation of diacylglycerol, which leads to activation of PKC (Lum and Malik, 1994). Treatment of many endothelial cells with diacylglycerols or phorbol esters to activate PKC increases paracellular permeability (Lynch et al.,

1990), although there are some endothelia in which these reagents confer resistance to permeability modulators (Yamada et al., 1990). Indeed, activation of PKC acts to promote the increased electrical resistance that develops across bovine brain microvessel endothelial cells co-cultured with C6 glioma cells (Raub, 1996).

PKC also has diverse effects on cell-cell adhesion in epithelial cells. Activation of PKC by phorbol ester or diacyl glycerol has been shown to reduce transepithelial resistance in a number of cell lines, although the effect is again dependent on cell type (Ojakian, 1981; Ellis et al., 1992; Soler et al., 1993; Stenson et al., 1993). The basis for the different effects of PKC activation is not known, although the identification of at least eleven different isoforms of PKC provide a possible explanation for diverse effects (for review see Kiley et al., 1995; Nishizuka, 1995).

cAMP

cAMP is an intracellular messenger that generally decreases the permeability of tight junctions. Addition of cAMP increases the resistance of Necturus gall bladder epithelium, whilst simultaneously increasing the number of tight junction strands (Duffey et al., 1981). cAMP also increases the resistance of tight junctions in brain endothelial cells (Rubin et at., 1991). Hypoxia decreases cAMP levels (and increases cell permeability) whereas prostaglandin El and E2 increase cAMP (and tend to enhance the barrier function of junctions) (Dejana etal., 1995). Treatment of cells with cAMP can block the permeability increase caused by ATP, HjOj, neutrophils, and thrombin (He and Curry, 1993; Suttorp et al., 1993; Siflinger-Bimboim et al., 1993; Patterson et al., 1994). Again there are exceptions; instances where cAMP has been associated with increased vascular permeability (Hempel et al., 1996). But generally, the cAMP, presumably acting via PKA, prevents disruption of junctions. cGMP can have similar effects to cAMP, blocking the permeability effects of thrombin and HjOj in certain endothelia (Lofton et al., 1990; Suttorp et al., 1996).

Interestingly, intracellular messengers such as PKC, inositol phosphates and Ca^^, which tend to promote disruption of established tight junctions, also played a role in de novo development of tight junctions. Development of resistance in MDCK cells following switch from low to high Ca^^ medium (the ‘Ca^"^ switch’ experiment) is blocked by clamping intracellular Ca^^ at constant levels by loading cell with the Ca^^ chelator BAPTA (Stuart et al., 1994). Activators of PKC induce formation of tight junctions between MDCK cells even in low Ca^^ medium; ie in the absence of cadherin- mediated cell-cell adhesion (Baida et al., 1993). The establishment of resistance and recruitment of ZO-1 to tight junctions following the Ca^^ switch can be prevented by inhibitors of PKC (Nigam et al., 1991; Stuart and Nigam, 1995). PLC may also be involved in establishing tight junctions following the Ca^^ switch in MDCK cells, since inhibition of this enzyme blocks the formation of tight junctions. This is likely to be due to an activation of PKC and/or Ca^^-mediated signal transduction pathways. There is also some evidence that G-proteins negatively regulate formation of tight junctions.

Addition of A IF3, which activates G-proteins, blocks development of resistance across MDCK epithelial cell monolayers following the Ca^^-switch, and pertussis toxin facilitated the development of resistance. In contrast, cAMP, which tend to enhance the barrier function of established junctions is actually an inhibitor of de novo formation of junctions; addition of cAMP prevents development of resistance across MDCK monolyers following the Ca^^ -switch (Baida et al., 1991). It seems that similar intracellular pathways can be involved in both the de novo formation of tight junctions, and the disruption of established junctions.

In conclusion, there are multiple second messenger pathways affecting the junctional integrity of epithelia and endothelia. These pathways interact and cross talk with each other in ways that have yet to be properly defined. The effects of specific agents will also vary depending on cellular context. Little is known of how these pathways increase or decrease paracellular permeability. Possible targets are the cytoskeleton and/or components of the intercellular junctions.