The blood brain barrier The blood brain barrier (BBB) is a selectively permeable cellular boundary between the brain and the peripheral circulation. The principal component of the BBB is the capillary or micro‐ vessel endothelial cell (Figure 1) . The endothelial cells in the brain capillaries differ from those in the peripheral vasculature in several key features: 1. Presence of tight junctions (TJ) that limit the paracellular passage of macromolecules. 2. Restricted rate of fluid-phase endocytosis that limits the transcellular passage of macro‐ molecules  3. Presence of specific transporter and carrier molecules  4. Lack of fenestrations  5. Increased mitochondrial content  Thus, the endothelial cells of the BBB are less “leaky” than those of the peripheral vessels. However, it has been shown that if the endothelial cells of the brain capillaries are removed from their natural environment and allowed to vascularize the peripheral tissue, they become more leaky . In contrast, the endothelial cells from the periphery form tight junctions when allowed to vascularize the brain parenchyma. Morphologically, the tight junctions of the BBBNext-> TRP Channels in Neuronal and Glial Signal Transduction
resemble the tight junctions between epithelial cells rather than those between peripheral vascular endothelial cells . The unique tight junctions of the BBB are responsible for producing very high transendothelial electrical resistance (TEER) of 1500 – 2000 Ωcm2 . Though the microvessel endothelial cells play a primary role in the formation of the BBB, several other cells are equally important in maintaining the integrity of the BBB. These cells, namely, the astrocytes, pericytes, neurons and other glial cells are said to form a “neurovascular unit” . Integrity of the BBB is of utmost importance in maintaining the homeostasis of the brain microenvironment. Disruption of the BBB is seen in various states of inflammation (multiple sclerosis), neoplasia, infections (meningitis, encephalitis), trauma and Alzheimer disease . It would be highly desirable to develop therapeutic strategies to reverse this disruption and tighten the BBB. At the same time, a transient opening of the BBB would be advantageous for delivery of drugs into the brain in conditions like epilepsy or Parkinson disease . Functions of the BBB The BBB is responsible for maintaining the appropriate ionic composition of the interstitial fluid of the brain that is required for optimum functioning of the neurons. To achieve this, the BBB functions as a transport barrier by facilitating the uptake of the required nutrients, while preventing the uptake of, or actively effluxing certain other molecules or toxic by-products of metabolism  The BBB also functions as a metabolic barrier by virtue of possessing intracel‐ lular and extracellular enzymes. For example, extracellular enzymes such as peptidases and nucleotidases break down peptides and ATP, respectively. Intracellular enzymes like cyto‐ chrome P450 (CYP450), primarily CYP1A and CYP2B degrade noxious substances and prevent their entry into the brain parenchyma . Role of astrocytes in the BBB It is now known that the astrocytes play a key role in the conditioning and development of the brain microvessel endothelial cells (BMEC). Astrocytes are one of the glial cells of the central nervous system (CNS) that play several important roles in the structure and function of the CNS. They are intimately associated with the BMEC such that their foot processes ensheath 99% of the external surface of the BMEC . Astrocytes have been shown to alter the properties of cocultured brain endothelial cells in the following ways [11,12]. 1. Increase in barrier-related marker enzyme activities, such as that of γ-glutamyl transpep‐ tidase (GGT) and alkaline phosphatase. 2. Enhanced expression of a glucose transporter. 3. Elevation of trans-endothelial electrical resistance (TEER). 4. Tightening of the BBB as seen by decreased paracellular permeability of sucrose. 5. Increase in tight junction number, length and complexity. It has also been shown that BMEC monolayers are less leaky if grown in the presence of astrocyte-conditioned medium (ACM) [1,11]. The precise molecular nature of the astrocyte-derived factors that is responsible for the tightness of the BBB have yet to be unequivocally elucidated. However, several factors have been postulated to play a role including glial cell-derived neurotrophic factor (GDNF), transforming growth factor-beta (TGF-β), and src-suppressed C-kinase substrate (SSeCKS) that leads to increased angiopoietin-1 secretion. The BMEC themselves are known to secrete factors that help in the maintenance of astrocyte health. One such putative factor is the leukemia-inhibitory factor (LIF), a cytokine known to be involved in astrocyte differentiation . Role of pericytes in the BBB The pericytes are specialized cells of mesenchymal lineage that have multiple organ-specific roles. For example, they are present in the kidney as mesangial cells, in the liver as perisinu‐ soidal stellate cells and in the bone as osteoblasts [13,14]. The pericytes in the central nervous system are closely associated with the BMEC and play an important role in the maintenance of the BBB. Their functions include . 1. Cerebrovascular autoregulation and blood flow distribution 2. Differentiation of the BBB 3. Formation and maintenance of the tight junctions of the BBB. 4. Initiation of the extrinsic (tissue factor) pathway of blood coagulation following cerebro‐ vascular injury 5. Brain angiogenesis via secretion of angiopoietin-1 6. Phagocytic and scavenging (macrophage-like) functions 7. Production of immunoregulatory cytokines like IL-1β, IL-6 and GM-CSF 8. Regulation of leukocyte transmigration, antigen presentation and T-cell activation. 2. Molecular components of the tight junctions The tight junctions consist of both membrane proteins as well as cytoplasmic proteins  (Figure 2). The integral membrane proteins are Claudins, Occludin and Junctional adhesion molecules (JAM). There are also several cytoplasmic accessory proteins that form a plaque and function as adapter proteins to link the membrane proteins to the actin cytoskeleton of the cell [16,17]. These include Zonula occludens proteins (ZO-1, ZO-2, ZO-3), Cingulin, AF-6, 7H6 antigen and Symplekin. These tight junctional complexes are not static structures but rather very dynamic entities that can “bend without breaking”, thereby maintaining structural integrity . Claudins The claudins are a large family of transmembrane phosphoproteins . Twenty-four members have been characterized so far, claudins 1-24 [18,19]. Of these, claudins 1, 3, 5 and 12 have been shown to form the tight junctions of the BBB [9,17,20,21]. Claudin-5 appears to be specific to the tight junctions of the endothelial cells and is called the “endothelial claudin” . Each claudin molecule has 4 transmembrane domains. The claudin on one cell binds homotypically to the claudin on the adjacent cell to form the seal of the tight junction. The claudins, along with occludin and the JAMs, form the tight Junctional strands that keep the cells together and prevent paracellular flux of macromole‐ cules from the apical to the basolateral side of polarized cells like BMEC . The cytoplas‐ mic carboxy terminal of the claudins binds to the cytoplasmic ZO proteins . Claudin-1 is an integral component of the tight junctions and its loss is associated with certain pathologic conditions like tumours, strokes and inflammatory diseases . Occludin Occludin is a 65-kDa transmembrane phosphoprotein and is distinct from the claudins. However, its subcellular localization parallels that of claudins and, like the claudins, it has four transmembrane domains.
Neurovascular Unit of the blood brain barrier consists of the endothelial cells (pink) surrounded by base‐ ment membrane (gray), pericytes (yellow) and astrocyte foot processes. The tight junctions (black lines) formed be‐ tween two endothelial cells restrict the paracellular diffusion of compounds.
The expression of occludin is higher in the adult BMEC compared to the peripheral endothelial cells. However, it is not expressed in the fetal or newborn human brain. Occludin plays an important structural, as well as a functional, role in the regulation of BBB permeability. As is the case with several other tight junction-associated proteins, phosphorylation or dephosphorylation of serine, threonine or tyro‐ sine residues on the occludin molecule is crucial for its proper functioning [17,18,22,23]. For example, phosphorylation of occludin at serine and tyrosine residues correlates with tight junction assembly or tightening . Occludin and the claudins interact intricately on the BMEC membrane. Together, they form channels that tightly regulate the paracellular flow of ions and other hydrophilic molecules. Thus, they are both essential in the formation, maintenance and regulation of the BBB [16,18]. Junctional Adhesion Molecules (JAM) These molecules play an important role in the regulation of tight junction permeability in endothelial and epithelial cells . These glycoproteins are members of the immunoglobulin superfamily of proteins. Three different JAMs have been characterized in humans, JAM-1, JAM-2 and JAM-3, also referred to as JAM-A, JAM-B and JAM-C, respectively. Besides endothelial and epithelial cells, these molecules are also found on the surface of erythrocytes, leukocytes and platelets and are thought to contribute to various processes like leukocyte migration, platelet activation, angiogenesis and binding of reovirus . The JAMs have short cytoplasmic tails that interact with cytoplasmic accessory proteins like ZO-1 and may require activation by phosphorylation, mediated by certain atypical protein kinases. Cytoplasmic accessory proteins Several cytoplasmic proteins appear to be essential components of the tight junctions. Among them, the zonula occludens proteins (ZO-1, ZO-2, ZO-3) play an important role. These 3 proteins have a molecular mass of 220, 160 and 130 kDa, respectively. They belong to a family of proteins called MAGUK (membrane-associated guanylate kinase-like protein) and form the submembranous plaque of the tight junction [2,15]. They are structurally complex proteins with several domains that make direct contact with claudins, occludin and JAM on one side and the actin cytoskeleton on the other . Cingulin is a double-stranded myosin-like protein that serves as scaffolding and links the TJ accessory proteins with the cytoskeleton . Actin, the cytoskeletal protein, plays a central role in the maintenance of the TJ. Actin-degrading macromolecules, such as cytochalasin-D, phalloidin and certain cytokines lead to disruption of the actin cytoskeleton and hence, of the tight junctions . The tight junctional proteins can be modulated by several intracellular processes that involve calcium-signaling, phosphorylation, G-proteins, proteases and by TNF-α [4,8]. The tight junctional complexes also help localize the proteins and lipids of the apical and basolateral cell membranes in their respective compartments and prevent free mixing of these cell membrane macromolecules between the two domains. Thus, the BMEC owe much of their polarity to the TJ complexes [2,26]. Regulation of BBB permeability Various factors play a role in regulating the permeability of the BBB as follows : 1. Post-translational modifications of the TJ proteins. For example, phosphorylation and dephosphorylation mediated by protein kinases and phosphatases, respectively. 2. Alteration of the actin cytoskeleton. 3. Proteolytic degradation of certain TJ components like occludin, mediated by metallopro‐ teinases. 4. In vitro models to study the BBB In vitro models of the BBB have proven very effective to study the transport of endogenous macromolecules like fatty acids across the BMEC. They have also been used extensively in pharmaceutical research to study the passage of therapeutic molecules across the BMEC [5-7]. Several studies have shown that the BMEC lose many of their special properties when removed from their natural environment and show “dedifferentiation” behaviour. Thus, one potential limitation of in vitro BBB models is that the BMEC may not behave as site-specific specialized endothelial cells in vitro, but rather as common peripheral endothelial cells . In spite of this shortcoming, several successful in vitro models of the BBB have been described . Many of these have used human, bovine, and porcine or rat endothelial cells: 1. Alone [5,6,28-30], or 2. in combination with astrocyte conditioned medium supplemented with agents that elevate intracellular cAMP , or 3. Co-culture of endothelial cells on one side of a filter, with astrocytes on the other . 5. FA transport across the BBB and effects of FA on BBB permeability Fatty acids (FA) are key components of membranes and exhibit many biological functions in a variety of tissues, including the key energy source for mitochondrial β-oxidation [32,33]. Cells acquire fatty acids through de novo synthesis, hydrolysis of triglycerides (TG) or uptake from exogenous sources . Minimal amount of FA are derived from TG hydrolysis and most cells are dependent upon fatty acid uptake from the peripheral blood [32,34]. FA from the diet are absorbed by enterocytes in the small intestine and packaged into chylomicrons as TG. The liver also produces very low density lipoprotein (VLDL), a rich source of endogenously generated TG. Circulating chylomicrons and VLDL particles are hydrolyzed by lipoprotein lipase in the capillary lumen of tissues and the released FA from these lipoproteins may be taken up by tissues in the body . FA that enter into cells are then esterified and stored as TG or transported to the mitochondria for β-oxidation. The importance of FA for the devel‐ oping and adult brain has been recently reviewed . FA transport from blood into paren‐ chymal neurons is much more difficult than other cells since the tight junctions of the BBB severely restrict passage into the brain. FA must first move via transcellular transport across both the luminal (apical) and abluminal (basolateral) membranes of the endothelial cells and then across the plasma membrane of the neural cells [36-38]. The mechanism of FA transport into the brain remains controversial. Several studies support the notion that FA can move across membranes by diffusion [39,40]. Alternatively, others studies indicate that FA may enter into cells via specific protein-mediated transport [32,41,42]. In the diffusion model, once bound to the outer membrane leaflet, they quickly reach ionization equilibrium and the non-ionized form of fatty acids move across the membrane more rapidly than the ionized form . The main problem with the FA diffusion model has always been whether diffusion is rapid enough to supply cells, which have a high long-chain FA metabolic requirement with sufficient amount of FA for β-oxidation . In the protein-mediated transport model selective transport of FA occurs via specific protein transporters found on the cell membrane [33,41,45-47]. The mechanism of FA transport into the brain and the involvement of FA protein transporters has been reviewed . We recently showed that the transport of various FA across confluent layers of HBMEC was, in part, mediated by fatty acid transport proteins (FATPs) [5,6]. Knock down of FATP-1 and CD36 resulted in reduced FA transport. In addition, transport appeared to be dependent upon fatty acyl chain length and degree of unsaturation. The role of FA, such as arachidonic acid (AA), on BBB permeability is well documented and controversial. Studies have indicated that a rapid influx of AA into the brain occurs upon plasma infusion with AA [48,49]. In addition, a permeability-enhancing and neurotoxic effect of AA has been observed [50-52]. AA is a precursor for the formation of various bioactive molecules including prostaglandins, such as PGE2, and leukotrienes. Several studies have indicated that the increase in BBB permeability is correlated with the formation of PGE2 [29,30, 53-56]. The prostaglandin EP2 receptor was shown to be responsible for mediating the neuroinflammatory and neurodegenerative effects of PGE2 in a mouse model of status epilepticus . The permeability increase caused by AA in pial microvessels of rats was effectively blocked by a combination of indomethacin (COX inhibitor) and nordihyroguaria‐ retic acid (LOX inhibitor) but not singly by either agent . In that same study, AA-mediated permeability increase was blocked by superoxide dismutase and catalase. These authors concluded that free radicals generated by either COX or LOX pathways were responsible for the permeability response to AA. In a mouse model of diabetic retinopathy 12-HETE and 15-HETE, products of the lipoxygenase pathway, were shown to be responsible for increasing the permeability of retinal endothelial cell barrier via an NADPH oxidase-dependent mechanism . Interestingly, AA inhibited the cytokine-induced up-regulation of several genes involved in endothelial cell inflammation . However, other studies have suggested that AA metabolites, such as PGE2, have a protective role in the microvessels of the CNS and that PGE2 prevents permeability increases. For example, the permeability increase caused by bradykinin was prevented or attenuated by exogenously added PGE2 and iloprost, a prostacyclin analog . In that study, COX-inhibitor drugs potentiated the permeability increases caused by bradykinin, thus suggesting an inhibitory role of PGE2 in increasing endothelial cell permeability. In addition, PGE2, acting via EP4 receptors, inhibited the increase in BBB permeability in a mouse model of experimental autoimmune encephalomyelitis . Moreover, PGE2, acting via EP2 receptors, has neuropro‐ tective properties and limits ischemic damage in mice stroke models . It has been postulated in these studies [61,62] that engagement of EP2 and EP4 receptors by PGE2 leads to an increase in cAMP levels. This cAMP accumulation has been shown to potentiate cadherin-mediated cell-cell contact and enhance endothelial barrier function. Thus, PGE2 may promote BBB integrity via direct action on endothelial cells . Several studies have demonstrated that microvessel endothelial cells from various organs have the capacity to produce a range of eicosanoids, notably, PGE2, PGI2 and PGF2α. In most of these studies the endothelial cells were stimulated with the calcium ionophore A23187 in addition to exogenously added AA [64-66]. However, in one study endothelial cells exposed to plasma from preeclamptic women showed increased production of prostaglandins . In addition,bovine brain microvessel endothelial cells (BBMEC) exposed to TNF-αreleased large amounts of PGE2 over a 12-hour period . Previous work has shown that docosahexanoic acid (DHA) is converted to its vasodilator metabolite, 17S-HDoHE in endothelial cells . DHA is a precursor in the formation of several bioactive molecules in human blood cells and in glial cells . However, in those experiments, the cells were exposed to stimulants like zymosan A or the calcium ionophore, A23187 to facilitate the release of DHA metabolites. These metabolites have been shown to have several biological effects like inhibition of inflammation and platelet aggregation, mediation of vasodilation, anti-arrhythmic effects and lowering of triglyceride levels .
Schematic representation of proteins that are involved in the formation of the tight junction and adherens junctions in brain microvessel endothelial cells.