Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

There are several unique physiologic properties of the pulmonary vasculature that are both necessary for normal lung function and impaired in certain disease states. All of these properties ultimately contribute to facilitating gas exchange between the alveoli and the pulmonary capillaries. At the most basic level, in order for gas exchange to occur, the lung requires both alveolar ventilation and capillary perfusion. In the healthy lung, matching of ventilation and perfusion (V/Q) is essential to optimizing gas exchange, yet pathologic conditions can have considerable effects on the distribution of either ventilation or perfusion, creating large inequalities, termed V/Q mismatch. To ensure optimal gas exchange, the lung vasculature has evolved a mechanism to correct for V/Q mismatch, termed hypoxic pulmonary vasoconstriction (HPV). Specifically, the small pulmonary arteries constrict in response to alveolar hypoxia, which serves to shunt blood away from areas of poor ventilation and redirect flow toward areas receiving adequate ventilation (Von Euler and Liljestrand 1946; Weir et al. 2005). However, persistent or uncontrolled vasoconstriction can have devastating consequences, leading to smooth muscle cell proliferation and other pathologic changes that are characteristic of PH (Mandegar et al. 2004).

Additional functions of the pulmonary vasculature are lung fluid balance and barrier integrity. The thin and delicate blood-gas barrier of the pulmonary capillaries is necessary for efficient diffusion of gases; however, this property also predisposes to barrier disruption and fluid leakage (Effros and Parker 2009; West 2013a). Furthermore, the immense surface area of the pulmonary vasculature creates a situation in which even small disruptions in barrier integrity can cause significant leakage of fluid into the alveoli (Dudek and Garcia 2001). The severity of these changes cannot be overemphasized, as the leakage of vascular contents into the alveoli can significantly impair gas exchange, leading to severe oxygenation defects and multisystem organ failure characteristic of acute respiratory distress syndrome (ARDS) (Matthay et al. 2012; Ware and Matthay 2000). Lastly, the endothelium plays a role in the regulation of the body’s inflammatory response and hemostatic mechanisms (Schouten et al. 2008; Seeley et al. 2012). This is particularly important in the pulmonary vasculature because the proximity of the lung to the outside environment makes it a common entry point for infectious microorganisms and inorganic particulate matter. In this chapter, we will discuss each of these physiologic mechanisms with an emphasis on how each is critical for normal lung function and how impairments in these processes contribute to diverse pathologic conditions including PH, ARDS, sepsis , and disseminated intravascular coagulation (DIC ). As each of these processes is immensely complex, the goal of this chapter is to provide an overview of each while describing mechanistic details and acknowledging controversies in the field as much as possible.

Ventilation-Perfusion Matching

In order for effective gas exchange to occur, the lung requires both alveolar ventilation and capillary perfusion. Ideally, both ventilation and perfusion would be evenly distributed across the entire lung to optimize gas exchange. In the normal lung, the biophysical properties of the air and blood create a situation in which there is a slight physiologic mismatch between ventilation and perfusion. However, in certain disease states, profound alterations in the distribution of either ventilation or perfusion cause significant V/Q mismatch and resulting hypoxemia. Here, we will review the properties of the lung and vasculature that govern the distribution of ventilation and perfusion and the mechanism underlying HPV, which aids in compensation for V/Q mismatch.

Zones of the Lung

Due to the fact that blood is significantly denser than air, the effect of gravity on blood flow to the lung is greater than its effect on airflow. Thus, although both ventilation and perfusion are greatest at the dependent portion of the lung, the magnitude of the blood flow gradient across the upright lung is greater than that of airflow (West et al. 1964). Therefore, the differing magnitudes of blood flow and airflow and their corresponding pressures in the capillaries and alveoli, respectively, create a situation in which the ratio of ventilation to perfusion differs throughout the healthy upright lung (West et al. 1964). This finding is used to define three basic zones of the lung based on the relative alveolar, arterial, and venous pressures (Fig. 1).

Fig. 1
figure 1

Zones of the lung. Schematic diagram indicating the three zones of the lung proposed by West et al. (1964). Zone 1 indicates the area of the lung in which alveolar pressure overcomes the circulatory pressure in both the arterial and venous systems. Zone 2 indicates the area of the lung in which arterial pressure in the pulmonary circulation overcomes alveolar pressure which is still maintained above the pulmonary venous pressure. Zone 3 illustrates the area of the lung in which pressure in the pulmonary circulation overcomes the pressure in the alveoli, and blood flow is determined by the pressure differences between the arterial and venous systems. P A indicates alveolar pressure, P a arterial pressure, P v venous pressure

Full size image

To understand the physiologic significance of these zones, it is necessary to understand the implications that the addition of alveolar pressure has on the vascular system. In the systemic circulation, the pressure differential between the arterial and venous ends of the circuit drives blood flow. However, in the pulmonary circulation, the need to accommodate the changing alveolar pressure, as well as pressure differentials caused by the unequal distribution of air and blood, creates a system in which three dynamic pressure components influence blood flow (Gil 2011). In the upper portion of the lung, where the alveolar pressure (PA) is higher than the pulmonary arterial pressure (Pa) and pulmonary venous pressure (Pv) (Zone 1), the high alveolar pressure compresses the capillaries, creating a situation in which blood flow through the capillaries is significantly limited. Progressing down the lung, Pa overcomes PA, and both remain higher than Pv (Zone 2). Here, slight imbalances in these pressures create a situation in which some capillaries are perfused (Pa > PA) and others are collapsed (PA < Pa). Lastly, approaching the base of the lung, the pressure differential is such that PA is lower than both Pa and Pv (Zone 3). This situation is most similar to the systemic circuit, in which blood flow is driven solely by the difference between arterial and venous pressures. At this level, all of the capillary beds are perfused, and the high intravascular pressure causes small undulations in the alveoli (West et al. 1964). In summary, although there are slight variations in blood flow and airflow throughout the lungs, in the normal lung, ventilation and perfusion are sufficient to achieve the volume of gas exchange needed to adequately oxygenate the blood.

Hypoxic Pulmonary Vasoconstriction

In certain disease states, substantial V/Q mismatch occurs, which the pulmonary vasculature attempts to compensate for via HPV (historically also referred to as the Von Euler-Liljestrand reflex). This physiologic property is dependent on the small pulmonary arteries, which respond to alveolar hypoxia with vasoconstriction. This response serves to divert blood away from the hypoxic alveoli so as not to waste valuable perfusion on the alveoli that are not being adequately ventilated (Von Euler and Liljestrand 1946). It is necessary to point out that this response is in direct contrast to that of the systemic vessels, which respond to hypoxia with vasodilation in order to bring oxygen-rich blood to the supplied tissues. HPV has been demonstrated in single pulmonary artery smooth muscle cells (PASMCs) as well as isolated pulmonary arteries denuded of endothelium, indicating that PASMCs are responsible for this process (Murray et al. 1990; Yuan et al. 1990). The fundamental mechanism of this response relies on the ability of a redox-based oxygen sensor to detect alveolar hypoxia and a diffusible redox mediator to couple this signal to an effector, which is primarily thought to be through Kv channel inhibition. Inhibition of Kv channels serves to block K+ efflux, which causes PASMC depolarization, opening of voltage-gated Ca2+ channels (VDCCs), Ca2+ influx, and ultimately the actin-myosin cross bridging and contraction which are responsible for vasoconstriction (Fig. 2; Moudgil et al. 2005). Although the mechanisms of PASMC depolarization and the downstream contractile apparatus activation have been well characterized, there is still significant debate regarding the exact molecular mechanisms of oxygen sensing in the pulmonary vasculature. The mechanisms of HPV remain under intense investigation for their significance in normal human physiology, as well as their role in disease states. Thus, although many details of this complex physiologic mechanism are still in question, we will point out some key components of this process.

Fig. 2
figure 2

Vasoreactivity in the pulmonary vasculature. Schematic diagram illustrating the mechanism by which downregulation of potassium efflux leads to increased cytosolic [Ca2+]. Ca2+-Calmodulin interaction then leads to upregulation of myosin light chain kinase and phosphorylation of myosin light chain. This stimulates the ATPase activity of myosin to provide the energy necessary for actin-myosin cross bridging and cycling, which leads to PASMC contraction and vasoconstriction. Alternatively, upregulation of myosin light chain phosphatase decreases the amount of phosphorylated myosin light chain to promote PASMC relaxation and vasodilation. K + indicates potassium, Ca 2+ calcium, SR sarcoplasmic reticulum, ER endoplasmic reticulum, cyt cytosolic, CaM calmodulin, ATP adenosine triphosphate, MLCK myosin light chain kinase, MLCP myosin light chain phosphatase, MLC myosin light chain, P phosphate

Full size image

Oxygen-Sensing Mechanisms

The degree of vasoconstriction induced by hypoxia is directly proportional to both the degree and duration of the hypoxic exposure suggesting the presence of a highly sensitive oxygen sensor (Archer et al. 1993, 1995; Weir et al. 2005). Additionally, HPV requires a mechanism to relay the signal of low partial pressure of alveolar oxygen (PAO2) to an effector protein that ultimately acts to alter the PASMC membrane voltage. The finding that inhibitors of oxidative adenosine triphosphate (ATP) production cause vasoconstriction and a loss of response to hypoxia in isolated, perfused lungs strongly suggests that the mitochondria may be the source of oxygen sensor required for HPV (Rounds and McMurtry 1981). Building on these findings, the Redox Theory was proposed to explain the oxygen-sensing mechanisms that mediate HPV (Archer et al. 1986). According to this theory, hypoxia alters production of a diffusible redox mediator which inhibits K+ efflux through Kv channels, ultimately leading to PASMC membrane depolarization. The oxygen-sensitive components are thought to be reactive oxygen species (ROS) generated by electron transport chain (ETC) components, rather than ATP itself, because only inhibitors of select ETC components were found to mimic the effects of HPV (Archer et al. 1993, 1999; Michelakis et al. 2002b). Mitochondrial superoxide dismutase (SOD) converts ROS to hydrogen peroxide (H2O2), a diffusible molecule that is also a powerful oxidant that can alter the redox state of the Kv channel (Moudgil et al. 2005). Thus, under normoxic conditions, tonic ROS generation leads to H2O2 production, which oxidizes Kv channels, leaving them in an open conformation; however, hypoxia induces a decrease in ROS generation, which decreases H2O2 production. This leads to Kv channel closure, which impairs K+ egress and ultimately causes membrane depolarization (Michelakis et al. 2002a; Moudgil et al. 2005; Reeve et al. 1995). Components of this hypothesis are widely accepted; however, the identities of both the oxygen-sensing component and the redox mediator are still being debated. Additionally, it is important to note that the effects of hypoxia on ROS generation remain highly controversial with different groups reporting both increases and decreases in ROS generation in response to hypoxia (Archer et al. 1989, 1993; Liu et al. 2003; Paky et al. 1993; Waypa et al. 2001). For a full discussion of this literature, please refer to comprehensive reviews on the subject (Moudgil et al. 2005; Sylvester et al. 2012; Weir et al. 2011).

Potassium Channel Inhibition and PASMC Depolarization

The increase in Ca2+ that causes the actin-myosin contractile force is initiated by Ca2+ influx through VDCCs. Thus, HPV relies on an effector component (Kv channel) that is able to couple the signal of low PAO2 to PASMC depolarization. In order to understand how Kv channel inhibition causes PASMC depolarization, it is necessary to review the factors that determine membrane voltage. The Nernst equation (1) quantifies the electrical potential across the cell membrane that is created by the electrochemical gradient of an individual ion.

$$ {\mathrm{E}}_{\mathrm{m}}=\frac{\mathrm{RT}}{\mathrm{zF}}\times \ln \left(\frac{{\left[\mathrm{X}\right]}_{\mathrm{out}}}{{\left[\mathrm{X}\right]}_{\mathrm{in}}}\right) $$
(1)

Here, Em is the membrane potential, R is the universal gas constant, T is absolute temperature, z is the valence of the ion, F is Faraday’s constant, and Xout and Xin are the ion concentrations outside and inside the cell respectively. This equation gives the electrochemical force across the cell membrane for an individual ion. However, the membrane voltage of the whole cell is determined by the electrochemical gradients of all cellular ions, particularly Na+, K+, and Cl− due to their abundance and their relative permeabilities (p) according to the Goldman-Hodgkin-Katz equation (2) (Wang et al. 2011):

$$ {\mathrm{V}}_{\mathrm{m}}=\frac{\mathrm{RT}}{\mathrm{F}} \ln \left[\frac{\mathrm{pK}\left[\mathrm{K}\right]\mathrm{o}+\mathrm{p}\mathrm{N}\mathrm{a}{\left[\mathrm{N}\mathrm{a}\right]}_{\mathrm{o}}+\mathrm{p}\mathrm{C}\mathrm{l}\left[\mathrm{C}\mathrm{l}\right]\mathrm{i}}{\mathrm{pK}\left[\mathrm{K}\right]\mathrm{i}+\mathrm{p}\mathrm{N}\mathrm{a}{\left[\mathrm{N}\mathrm{a}\right]}_{\mathrm{i}}+\mathrm{p}\mathrm{C}\mathrm{l}\left[\mathrm{C}\mathrm{l}\right]\mathrm{o}}\right] $$
(2)

Based on the electrochemical gradients present in PASMCs, their resting membrane potential is −60 mV (Nelson and Quayle 1995). Notably, the high intracellular K+ concentration in PASMCs generates a strong outward electrochemical force, which is responsible for its tonic efflux from the cell. Therefore, blocking the outward flowing K+ current causes membrane depolarization by increasing the intracellular K+ concentration. In the context of the HPV response, it has been demonstrated that within seconds of exposure to hypoxia, K+ channels in PASMCs are blocked, which allows for onset of vasoconstriction within 1 min, a maximum response at 15 min, and rapid reversal of vasoconstriction upon restoration of normal oxygen tension (Bindslev et al. 1985; Yuan et al. 1995).

The K+ current in PASMCs is determined by multiple types of K+ channels, each with their own unique properties. In PASMCs, four families of K+ channels are expressed: voltage-gated (Kv), inward rectifier (Kir), calcium-sensitive (KCa), and two-pore (K2P) (Firth and Yuan 2011; Moudgil et al. 2005, 2006). A key observation that aided in determining the molecular identity of the K+ channels involved in the HPV response was that tetraethylammonium chloride and 4-aminopyridine, which inhibit Kv channels, are able to induce vasoconstriction in perfused rat lungs; however, glibenclamide, which inhibits ATP-sensitive channels, does not cause this effect (Hasunuma et al. 1991). This suggests that Kv channels are responsible for inducing membrane depolarization and activating VDCCs. Particular Kv channel subtypes that have been strongly associated with HPV include Kv1.5 and Kv2.1. Experiments that utilized blocking antibodies to Kv1.5 and Kv2.1 as well as Kv1.5 knockout (KO) mice demonstrate impaired vasoreactivity in response to hypoxia, confirming that these subtypes play an important role in the HPV response (Archer et al. 1998, 2001). It has been established that depolarization of PASMCs due to inhibition of Kv currents ultimately causes the influx of Ca2+ through VDCCs (McMurtry 1985; McMurtry et al. 1976; Post et al. 1992; Tolins et al. 1986; Yuan et al. 1993).

The key role of Kv channels in the molecular mechanism of HPV suggests that changes in their expression may contribute to the pathophysiology of PH, a condition in which uncontrolled vasoconstriction leads to activation of numerous pathogenic mechanisms responsible for PH development and progression. In PH, there is a metabolic shift from oxidative phosphorylation to glycolysis resulting in decreased ROS production, vasoconstriction, and hypoxia-inducible factor (HIF) 1α activation under normoxic conditions (Archer et al. 2008). It has been demonstrated that HIF1α activation causes decreased expression of Kv1.5, which contributes to normoxic vasoconstriction, strongly implicating HIFs in the development of PH (Shimoda et al. 2001). This has also been supported by the fact that HIF1α and HIF2α heterozygous KO mice display an attenuated response to chronic hypoxia as measured by right ventricular hypertrophy and pulmonary artery pressure (Brusselmans et al. 2003; Yu et al. 1999).

Actinomyosin Contraction and Calcium Sensitization

Vasoconstriction ultimately results from generation of actinomyosin contractile forces in PASMCs that are coupled and propagated both longitudinally and circumferentially through the vessel; the magnitude of which relies both on levels of free cytosolic Ca2+ concentration ([Ca2+]cyt) and Ca2+ sensitization of several protein components of the contractile apparatus (Mandegar et al. 2004). The direct effect of [Ca2+]cyt occurs via formation of a Ca2+/calmodulin (Ca2+/CaM) complex, which activates myosin light chain kinase (MLCK), an enzyme that phosphorylates myosin light chain (MLC), an essential component of the contractile apparatus. Phosphorylated MLC (pMLC) stimulates the ATPase activity of myosin, which provides the energy necessary for actin-myosin cross-bridge cycling (Somlyo and Somlyo 1994). However, it is well established that the sensitivity of the contractile apparatus can be altered by multiple cellular components. This process, termed Ca2+ sensitization, is characterized by an increase in contractile force generation in the absence of an increase in [Ca2+]cyt (Morgan and Morgan 1984). Due to the fact that the actin-myosin contractile force is ultimately triggered by pMLC, enzymes that alter the level of pMLC, such as myosin light chain kinase (MLCK) and myosin light chain phosphatases (MLC phosphatases), serve to modulate the sensitivity of the contractile apparatus (Mandegar et al. 2004; Somlyo and Somlyo 2003). Specifically, Rho/Rho kinase activation increases Ca2+ sensitization via inhibition of MLC phosphatase, causing an increase in cellular levels of pMLC and augmenting actin-myosin contractility (Gong et al. 1996; Sauzeau et al. 2000). Recently, multiple additional signaling pathways have been implicated in Ca2+ sensitization of PASMCs through interactions with MLC phosphatase, including protein kinase C, receptor tyrosine kinases, and p38 mitogen-activated protein kinases (MAPKs) (Mandegar et al. 2004).

Barrier Integrity and Lung Fluid Balance

In order to attain the volumes of gas exchange needed to adequately oxygenate the blood, the body has evolved an extremely thin blood-gas barrier with an immense surface area (West 2013a, b). Although evolutionarily advantageous, these properties pose challenges to the barrier function of the endothelial monolayer of the pulmonary vasculature. Specifically, the delicate blood-gas barrier makes the lung vasculature prone to disruption, and the immense surface area makes it exquisitely sensitive to changes in vascular permeability (Effros and Parker 2009). This is critical because barrier disruption leads to the leakage of protein-rich fluid and inflammatory cells into the alveoli, ultimately causing oxygenation defects that can lead to multisystem organ failure characteristic of ARDS (Ware and Matthay 2000). Thus, endothelial barrier function and maintenance of normal lung fluid balance is an essential function of the pulmonary vasculature.

Biophysical Determinants of Fluid Flux Across the Vasculature

To understand the barrier function of the pulmonary vasculature, it is necessary to analyze the forces responsible for fluid balance across the endothelium. The Starling equation (3) explains how the biophysical properties of the vasculature determine the magnitude of fluid flux across the endothelium (Mehta and Malik 2006; Michel and Curry 1999):

$$ {\mathrm{J}}_{\mathrm{v}}={\mathrm{L}}_{\mathrm{p}}\mathrm{S}\ \left[\left({\mathrm{P}}_{\mathrm{c}}-{\mathrm{P}}_{\mathrm{i}}\right)-\upsigma \left({\uppi}_{\mathrm{c}}-{\uppi}_{\mathrm{i}}\right)\right] $$
(3)

Here, Jv is the volume of fluid flux from the endothelium to the interstitium; Lp is the hydraulic conductivity; S is the capillary surface area; Pc,i and πc,i are the hydrostatic and oncotic pressures in the capillary and interstitium, respectively; and σ is the osmotic reflection coefficient. From this equation, it is clear that fluid flux across the vasculature depends on the gradients of hydrostatic and oncotic pressures, as well as the permeability and surface area of the vessel wall itself. Thus, the low hydrostatic pressure in the pulmonary circulation (Pc) protects against vascular leak. Additionally, the oncotic pressure gradient, which is mainly attributable to albumin due to its abundance in plasma, plays a role in fluid balance across the endothelium and will be discussed with the transcellular transport pathway. Lastly, the term LpS, which can also be expressed as the filtration coefficient (Kf,c), describes the permeability of the barrier itself. This variable is altered by many inflammatory mediators that disturb endothelial monolayer integrity and will be discussed further with the paracellular transport pathway.

As we have alluded to above, there are two distinct pathways, paracellular and transcellular, that account for the transport of fluid and small molecules across the pulmonary endothelium (Fig. 3). Evidence for this view is based on studies of the relative permeability of molecules varying in size. Specifically, the permeability of the endothelium to substances with a molecular radius less than 3 nm is proportional to size, whereas the permeability of substances with a radius larger than 3 nm is size independent (Mehta and Malik 2006; Pappenheimer et al. 1951; Renkin et al. 1974; Siflinger-Birnboim et al. 1987, 1988). Thus, transport of solutes larger than 3 nm in radius has been attributed to the transcellular pathway, an active process by which vesicles form at the plasma membrane of the luminal surface, move through the cell, and ultimately release their contents at the basal aspect of the cell (Minshall and Malik 2006; van Nieuw Amerongen et al. 2011). In contrast, the transport of solutes less than 3 nm in radius has been mainly attributed to the paracellular pathway, by which molecules travel down their concentration gradient between adjacent endothelial cells (ECs). Although both of these processes are involved in fluid balance, the transcellular pathway is thought to be responsible for basal fluid balance across the vasculature, with minimal movement of solutes through tight paracellular gaps. However, in response to inflammatory mediators, large paracellular gaps form between ECs, which provide a conduit for pathologic leakage of fluid, protein, and inflammatory cells (Chiang et al. 2011).

Fig. 3
figure 3

Transport across the endothelium. Schematic figure which illustrates the two primary modes of transport across the intact endothelium. The transcellular pathway is an active process by which fluid and proteins cross the endothelium in membrane-bound vesicles, primarily mediated through receptor binding. The paracellular pathway transports fluids and proteins across the endothelium between cells and may be affected by numerous agents that disrupt the connections between intact endothelial cells

Full size image

The Transcellular Pathway

Movement of fluid via the transcellular pathway , termed transcytosis, is an active process by which fluid and proteins cross the endothelium in membrane-bound vesicles. Since albumin is the most abundant protein in plasma, its transport via the transcellular pathway is critical to maintaining the oncotic pressure gradient, making it a key determinant of fluid flux across the endothelium (Landis and Hortenstine 1950; Weisberg 1978). Given its role in fluid balance under physiologic conditions, it comes as no surprise that this process is both highly regulated and energy dependent (Frank et al. 2003). In this process, albumin binding to its receptor causes the formation of small invaginations in the plasma membrane, termed caveolae, which then pinch off the membrane, are transported across the cell, and then are ultimately released at the basal portion of the cell (Fig. 4).

Fig. 4
figure 4

Proposed structure and role of caveolae in the endothelium. Caveolae structure in which membrane invaginations are produced by interaction of caveolin-1 with the plasma membrane. This structure has been proposed to play an important role in cellular signaling processes including ion channel activity and G-protein-coupled receptor activation among others (a). Caveolae also play a role in transcellular transport in Src activation leading to phosphorylation of dynamin, which is responsible for contractile forces which allow caveolae to pinch off the plasma membrane into intracellular vesicles (b). Ca 2+ indicates calcium, G G-protein, Src sarcoma nonreceptor tyrosine kinase, Cav-1 caveolin-1

Full size image

The Role of Caveolin-1

The importance of caveolae , and their main component caveolin-1 (Cav-1), is underscored by the fact that caveolae account for 20 % of the volume of ECs (Predescu and Palade 1993). Cav-1 is a 21–22 kDa protein that is part of the caveolin family of proteins, which includes caveolins 1, 2, and 3; however, only Cav-1 and Cav-2 are abundant in ECs (Couet et al. 2001; Komarova and Malik 2010). Studies of Cav-1 KO mice have established that Cav-1 is necessary for the formation of caveolae and that Cav-2 expression does not compensate for its absence (Drab et al. 2001; Razani et al. 2001). Thus, research regarding the mechanisms of the transcellular pathway has centered on understanding the function of this complex protein. Structurally, Cav-1 is composed of two cytoplasmic domains and a central hydrophobic domain that forms a hairpin structure allowing it to tether to the plasma membrane (Machleidt et al. 2000). The cytoplasmic domains are involved in oligomerization, whereas the scaffold domain allows it to interact with several proteins that are necessary for the formation of caveolae (Mehta and Malik 2006). Additionally, there are a few key residues in Cav-1 that are subject to tyrosine phosphorylation and palmitoylation, which serve to regulate its activity and localization, respectively (Dietzen et al. 1995; Li et al. 1996).

The process of transcytosis is initiated by albumin binding to its transmembrane glycoprotein receptor, gp60, which induces receptor aggregation and interaction with Cav-1 (Minshall et al. 2000). Cav-1 monomers then assemble into oligomers to form flask-shaped invaginations of the plasma membrane (Sargiacomo et al. 1995; Schlegel and Lisanti 2000). The scaffold domain of Cav-1 then recruits additional proteins that are necessary for caveolae formation including Src family kinases and G proteins (Komarova and Malik 2010). Upon formation of this signaling complex, Cav-1 triggers Gαi activation, which serves to activate c-Src, which sequentially phosphorylates Cav-1 (Tyr14) and dynamin (Tyr231, Tyr597). Phosphorylation of dynamin induces its recruitment and oligomerization at caveolae and increases its intrinsic GTPase activity, which is responsible for the contractile force generation that allows caveolae to pinch off from the plasma membrane (Fig. 4; Komarova and Malik 2010). Notably, although pulmonary microvascular ECs from Cav-1 KO mice were unable to form caveolae, they had increased permeability to albumin via the paracellular pathway, suggesting interaction between these two pathways (Razani et al. 2001; Schubert et al. 2002).

The Paracellular Pathway

The paracellular pathway was conceptualized by Majno and Palade, who demonstrated through electron microscopy studies that inflammatory agents cause cellular rounding and the development of large gaps between ECs. These gaps allow increased solute movement from the vasculature into the interstitium (Majno and Palade 1961; Majno et al. 1961). The fact that inflammatory agents alter endothelial cell shape strongly implicates the endothelial cytoskeleton in the regulation of barrier permeability. Subsequent work has demonstrated that paracellular gap formation involves activation of actinomyosin contractile forces that cause cell rounding and disruption of the cell-cell and cell-matrix contacts responsible for the structure and selectivity of the endothelial barrier (Chiang et al. 2011).

Regulation of the Actinomyosin Contractile Apparatus

F-actin microfilaments are cytoskeletal elements critical to maintaining the shape and structure of ECs. They are formed by polymerization of G-actin monomers and are subject to highly dynamic regulation by a variety of actin binding proteins , performing diverse functions which include polymerization/depolymerization, branching, capping, and severing (Dudek and Garcia 2001). In resting ECs, actin microfilaments form a cortical band at the periphery of the cell which promotes monolayer integrity via cell-cell linkage and cell-matrix contacts. However, edemagenic agonists such as thrombin , histamine , and vascular endothelial growth factor (VEGF) cause a loss of this cortical actin band and formation of cytoplasmic actin stress fibers . Stress fibers span the cell and generate tensile forces to cause cellular rounding and decreased monolayer integrity that ultimately leads to fluid leakage across the endothelium via the paracellular pathway (Goeckeler and Wysolmerski 1995). Thus, regulation of the actin-myosin contractile apparatus in ECs mediates a balance between contractile forces , which cause cell rounding, paracellular gap formation, and tethering forces , which promote the formation of actin-based protrusions that aid in the restoration of monolayer integrity (Dudek and Garcia 2001; Garcia et al. 1995).

The driving force behind these changes in ECs is activation of nonmuscle MLCK (nmMLCK ), a Ca2+/CaM kinase-dependent enzyme that phosphorylates MLC on Ser-19 (monophosphorylation) or Ser-19/Thr-18 (diphosphorylation) which ultimately drives actin-myosin contraction (Garcia et al. 1995). Although nmMLCK and smooth muscle MLCK are both transcribed from the MYLK gene , nmMLCK contains a unique N-terminal domain with several residues that allow for its regulation via posttranslational modifications, as well as an SH2 domain that mediates additional protein-protein interactions (Birukov et al. 2001; Garcia et al. 1999). Several edemagenic agents operate as part of signaling pathways that lead to increased nmMLCK activation at the center of the cell, leading to stress fiber formation and decreased endothelial barrier function (Becker et al. 2001; Garcia and Schaphorst 1995). Furthermore, pharmacologic inhibition of nmMLCK (PIK and ML-7) attenuates the increased EC permeability induced by edemagenic agonists, and nmMLCK KO mice are protected from vascular leak after exposure to these agents (Mirzapoiazova et al. 2011). Thus, nmMLCK is a central target of current work aimed at developing novel therapeutic strategies to attenuate vascular leak in complex illnesses such as sepsis and ARDS. Furthermore, the MYLK gene contains several single nucleotide polymorphisms (SNPs) associated with sepsis , acute lung injury (ALI), and severe asthma, further implicating nmMLCK as critical mediator of endothelial barrier function (Christie et al. 2008; Flores et al. 2007; Gao et al. 2006, 2007). In contrast, nmMLCK has also been identified as a component of the barrier protective response to stimuli including the bioactive sphingolipid sphingosine 1-phosphate (S1P). This protective effect is caused by peripheral nmMLCK activation, which causes the formation of a cortical actin ring and actin-based cellular protrusions called lamellipodia which aid in restoration of the endothelial barrier (Dudek et al. 2004; Zhao et al. 2009). Continuing work is aimed at understanding the spatiotemporal regulation of nmMLCK in order to develop therapies that can enhance endothelial barrier function.

Endothelial Cell-Cell and Cell-Matrix Junctions

The endothelial cell-cell junctions include adherens junctions (AJs), tight junctions (TJs) , and gap junctions (GJs) ; however, AJs are considered the main junctional complex between ECs (Fig. 5; Mehta and Malik 2006). AJs mediate Ca2+-dependent adhesions between adjacent ECs and are composed principally of the transmembrane protein VE-cadherin (Dejana et al. 2008). The extracellular domain of VE-cadherin mediates homotypic binding with adjacent ECs, whereas the intracellular domain of VE-cadherin binds primarily to catenin proteins , including β-catenin, γ-catenin, and p120-catenin, which serve as a linkage to the actin cytoskeleton (Dejana and Giampietro 2012). The importance of these junctions has been established by experiments demonstrating increased vascular leak with both VE-cadherin-blocking antibody infusion in mice and Ca2+ chelation in cells (Corada et al. 1999; Gao et al. 2000). The regulation of AJ integrity is thought to occur mainly by phosphorylation of VE-cadherin or its binding partners, β-catenin, plakoglobin, and p120-catenin, which causes internalization of AJ proteins and ultimately a loss of vascular integrity (Dejana et al. 2008; Ozawa and Kemler 1998; Vogel and Malik 2012). Furthermore, phosphorylation of AJ proteins has been demonstrated after edemagenic agonists such as histamine , suggesting that this method of regulation alters junctional integrity (Andriopoulou et al. 1999; Guo et al. 2008; Shasby et al. 2002).

Fig. 5
figure 5

Endothelial cell-cell junctions. Schematic diagram of the three types of endothelial cell-cell junctions. Tight junctions consist of three primary members, the occludins, claudins, and JAM proteins. Gap junctions are created by connexins in order to provide cellular cross talk. Adherens junctions are the most prominent connection between endothelial cells and are formed by VE-cadherins which connect on the exterior surface of endothelial cells and are anchored to the actin cytoskeleton within cells. JAM indicates junctional adhesion molecule, VE-cadherin vascular endothelial-cadherin

Full size image

TJs form a more restrictive barrier than AJs; however, in ECs they comprise only 20 % of the cell-cell junctions (Wojciak-Stothard et al. 2005). They are composed predominantly of claudins, occludins, and junctional adhesion molecules (JAMs). In a similar arrangement to AJs, both occludins and claudins contain extracellular domains that mediate binding to adjacent cells and cytoplasmic domains that bind to proteins that link them to the actin cytoskeleton. Although the regulation of TJ function following inflammatory agonists has not been well characterized in pulmonary ECs, it has been demonstrated that barrier-altering agonists induce phosphorylation of TJ components in a pathway involving RhoA-ROCK (Fischer et al. 2002; Hirase et al. 2001; Wojciak-Stothard et al. 2005). Lastly, gap junctions are also present between ECs; however, the function of these complexes is to metabolically and electrically couple adjacent cells, rather than to provide a barrier to fluid leakage; thus, we will not discuss them further here.

Cell-matrix junctions are also critical to maintaining EC shape and barrier function. These connections, termed focal adhesions (FAs ), are principally composed of integrins , which exist as a heterodimer of α- and β-subunits that are each transmembrane glycoproteins. The external portion of integrins binds to the RGD domain of extracellular matrix (ECM) proteins including fibronectin, fibrinogen, vitronectin, and collagen with disruption of these interactions leading to cellular rounding and an increase in vascular permeability (Curtis et al. 1995; Qiao et al. 1995; Wu et al. 2001). The intracellular domain connects to actin binding proteins including vinculin, α-actinin, paxillin, talin, zyxin, tensin, and filamin (Geiger et al. 2001). These intracellular and extracellular contacts allow integrins to coordinate bidirectional signaling between the ECM and the actin cytoskeleton. Similar to AJs and TJs, the regulation of focal adhesions involves Rho family GTPases and phosphorylation of proteins that are part of these complexes (Mehta and Malik 2006).

Contributions of the Pulmonary Vasculature to Inflammation and Hemostasis

The endothelium has been demonstrated to plays a role in the regulation of the inflammatory response as well as multiple processes involved in hemostasis (Fig. 6). This is particularly important in the pulmonary vasculature because the proximity of the lung to the outside environment makes it a common entry point for infectious microorganisms and inorganic particulate matter. Additionally, the intimacy of the endothelium to the blood volume and leukocyte population enables the body to target the inflammatory response to the injured area, thus ensuring a localized response. This is a critical function in which localized inflammation and coagulation are components of normal physiology, yet, when extensively stimulated or uncontrolled, they can contribute to devastating disease states including sepsis and DIC (Schouten et al. 2008). Overall, the unperturbed endothelium provides an anti-inflammatory and anticoagulant conduit for blood, which serves to promote blood fluidity and prevents intravascular coagulation. However, when ECs sense the presence of pathogens, they undergo localized activation and play a highly active role in containing the infection and minimizing collateral tissue damage (Félétou and Vanhoutte 2006). This response includes the production of cytokines to promote inflammatory cell recruitment and the expression of cell surface adhesion molecules to coordinate leukocyte extravasation. EC activation also promotes a procoagulant state via changes in gene expression and the exposure of molecules that promote platelet adhesion (von Willebrand Factor , vWF) and coagulation (tissue factor , TF) (Moore et al. 1987; Pober and Sessa 2007; Schouten et al. 2008).

Fig. 6
figure 6

Endothelial response to inflammatory stimuli. Schematic diagram showing the role of the endothelium in production and propagation of inflammatory mediators and inflammatory cell migration. Inflammatory stimuli induce the production of numerous cytokines and chemokines in the activated endothelium which promote inflammatory cell recruitment and activation. Leukocyte recruitment to the endothelium is initiated by selectins on the endothelial surface binding to glycoprotein receptors on the leukocyte. Integrins, such as CD11/CD18, on the surface of leukocytes bind cellular adhesion molecules, such as ICAM-1, to tightly adhere leukocytes to the endothelial surface. PE-CAM at the area of cell-cell junctions then promotes transmigration of the inflammatory cells across the surface to the site of injury. IL indicates interleukin, IFN-γ interferon gamma, TNF-α tumor necrosis factor alpha, LT leukotriene, PE-CAM platelet endothelial cell adhesion molecule, ICAM intracellular adhesion molecule, CD cluster of differentiation

Full size image

Endothelial Activation in Response to Infection

In order to effectively respond to infections, the human body needs to recognize the inciting pathogen and activate the immune system. Pathogen-associated molecular patterns (PAMPs) are molecules that are common to multiple pathogens and can be recognized by the immune and endothelial cells. The most well-characterized PAMP is lipopolysaccharide (LPS), a component of the cell wall of gram-negative bacteria. LPS is recognized by binding to toll-like receptor 4 (TLR4), which is expressed on EC plasma membrane, and initiates a NFκB-dependent pathway leading to changes in gene expression of inflammatory cytokines and cellular adhesion molecules (CAMs). These changes activate endothelial cells, enabling them to recruit inflammatory cells to the site of injury. Endothelial activation can also be caused by cytokines and mediators produced by other cell types during inflammatory processes, including tumor necrosis factor (TNF)α , interleukin (IL)-1 , IL-4, interferon (IFN)γ, thrombin, histamine, bradykinin, leukotriene C4, and H2O2 (Albelda et al. 1994). The activated endothelium initiates production of pro-inflammatory cytokines and chemokines, including TNFα and IL-1, stimulating inflammatory cell recruitment and activation (Mantovani et al. 1997; Strieter and Kunkel 1994). However, TLR-4 signaling in the active endothelium also induces the expression of anti-inflammatory cytokines including IL-10 and IL1Rα, which serve to contain the inflammatory response and limit damage to surrounding tissue (Calvano et al. 2005).

Additionally, in order to fight infection, leukocytes need to be recruited to the site of infection and allowed to extravasate into the tissues, where they can phagocytize and kill harmful pathogens. This process involves both ECs and the leukocytes themselves, which together coordinate a highly regulated sequence of events, including rolling, adhesion, and extravasation. Each step of this process is dependent on a different set of cellular adhesion molecules (CAMs). The interaction between selectins and their glycoprotein ligands mediates the rolling step of neutrophil recruitment, which serves to slow circulating neutrophils and enable their subsequent adhesion to the endothelium. L-selectin is constitutively expressed on leukocytes; however, expression of E-selectin and P-selectin is induced by EC activating signals, such as those produced by stimulation with LPS (Albelda et al. 1994). Neutrophil adhesion is mediated by cluster differentiation (CD); the CD11/CD18 integrins are constitutively expressed on neutrophils with intercellular adhesion molecule (I-CAM)-1 and vascular cellular adhesion molecule (V-CAM)-1 , which are expressed on ECs in response to TNFα and IL-1 (Albelda et al. 1994). In the inactive state, integrins are kept in a conformation which prevents leukocyte adhesion to the endothelium; however, in activated ECs, these molecules adopt an active conformation that allows leukocytes to adhere to the endothelium. Subsequently, in order for the leukocytes to cross the endothelial barrier, ECs express platelet endothelial cell adhesion molecule (PE-CAM)-1 at the areas of cell-cell junctions which promotes transmigration (Vaporciyan et al. 1993). In summary, the response of the endothelium to indications of infection includes upregulation of the expression of CAMs and pro-inflammatory cytokines, which serve to facilitate the inflammatory response.

Role of the Pulmonary Vasculature in Hemostasis

The hemostatic mechanisms of the body serve to maintain a balance between blood fluidity and vascular integrity. In the absence of damage, the endothelium produces factors that oppose coagulation and platelet activation while simultaneously promoting fibrinolysis (van Hinsbergh 2012). The lung vasculature has a high thrombotic potential, due to its relatively high expression of both the platelet adhesion molecule von Willebrand Factor (vWF) and coagulation cascade initiator tissue factor (TF) (Mackman et al. 1993; Yamamoto et al. 1998). Thus, effective anticoagulant mechanisms are necessary to prevent inappropriate coagulation cascade activation and thrombosis. However, when the vasculature is damaged, the body needs to initiate a fast and localized response in order to prevent blood loss and repair the damaged endothelium. This involves activation of both primary hemostasis, which generates the initial platelet plug, and the coagulation cascade, which consolidates the initial platelet plug by fibrin generation (Crawley et al. 2011).

Anticoagulant Properties of the Resting Vasculature

The resting endothelium provides an anticoagulant surface by expressing surface proteins that inactivate coagulation components and by secreting anticoagulant molecules (Bombeli et al. 1997; Schouten et al. 2008). The three main anticoagulant mechanisms, tissue factor pathway inhibitor (TFPI), activated protein C (APC ), and antithrombin , each rely on the endothelium (Crawley et al. 2011). Upon secretion from ECs, TFPI inhibits TF-mediated coagulation by binding and inhibiting factor (F) Xa (Huang et al. 1993; Jesty et al. 1994). The complex of TFPI-FXa also further blocks coagulation by inhibiting TF-FVIIa (Broze et al. 1988). Additionally, the endothelium expresses the glycoproteins thrombomodulin and endothelial protein C receptor (EPCR ), which contribute to the protein C pathway for anticoagulation (Visovatti et al. 2011; Weiler and Isermann 2003). When thrombin binds to thrombomodulin, a direct inhibition of thrombin function occurs. The complex of thrombin-thrombomodulin also activates protein C which is bound to EPCR on the endothelial surface, forming APC (Crawley et al. 2011). APC exerts its anticoagulant effects via inhibition of FVa and FVIIIa (Esmon 1987, 1992). Lastly, the endothelium inhibits coagulation via production of antithrombin, a serine protease inhibitor that serves to directly oppose thrombin, FIXa, FXa, FXIa, and FXIIa (Quinsey et al. 2004).

In addition to inhibition of the coagulation cascade, the resting endothelium contributes to blood fluidity by secreting molecules that block platelet activation (nitric oxide , NO) and aggregation (prostacyclin , PGI2) while promoting fibrinolysis (tissue plasminogen activator, tPA) (Visovatti et al. 2011). Under resting conditions, endothelial nitric oxide synthase (eNOS , NOS3) generates NO from L-arginine (Marsden et al. 1992). In addition to the profound vasodilatory effects of NO, this molecule also blocks gpIIb/IIIa binding to fibrinogen, an interaction that induces platelet activation (Mellion et al. 1981). Additionally, NO suppresses the surface expression of CAMs and the production of pro-inflammatory cytokines, which keeps the endothelium in a quiescent state (De Caterina et al. 1995). PGI2 is produced through the arachidonic acid (AA) pathway from prostaglandin H2 by prostacyclin synthase. It plays a role in blocking platelet activation, secretion, and aggregation (Visovatti et al. 2011). Lastly, tissue plasminogen activator (tPA) is a serine protease inhibitor of coagulation that binds the TF-FVIIa-FXa complex to limit the amount of FXa available to activate thrombin. Together these various molecules produced by the endothelium serve to maintain fluid blood flow in the unperturbed state by blocking platelet activation and coagulation while promoting fibrinolysis.

Procoagulant Functions of the Activated Endothelium

In the context of inflammation, primary hemostasis and the coagulation cascade are initiated as part of an adaptive response that serves to prevent blood loss through damaged vasculature. Specifically, Weibel-Palade bodies containing vWF are released from the ECs due to an increase in intracellular Ca2+ that occurs in response to many inflammatory agents including thrombin and histamine (Pober and Sessa 2007). The vWF binds to subendothelial collagen present on the denuded endothelial basement membrane and promotes platelet adhesion to the damaged vasculature, thus initiating the process of primary hemostasis. Additionally, in response to LPS and in numerous disease states, ECs are shifted toward a procoagulant phenotype via decreased expression of thrombomodulin , tissue-type plasminogen activator, and heparin, as well as increased expression of TF (Moore et al. 1987; Seeley et al. 2012). TF on the basal surface of ECs becomes exposed when damage occurs, which quickly initiates the extrinsic pathway of the coagulation cascade. This activity is also heightened in the activated endothelium by mediators including IL-1 and TNFα, which induce the synthesis of additional TF (Bevilacqua et al. 1986). Thus, in contrast to the resting endothelium, the activated endothelium serves to repair the damage via activation of both primary hemostasis and initiation of the coagulation cascade.

Summary

In summary, the physiologic properties of the pulmonary vasculature play an active role in optimization of gas exchange. Although these processes are necessary for lung function, in certain disease states they can be a source of significant complications. For example, the physiologic property of hypoxic pulmonary vasoconstriction allows the pulmonary vasculature to shunt blood away from unventilated alveoli, thus improving gas exchange; however, if prolonged, this process leads to vascular remodeling and ultimately pulmonary hypertension. Additionally, in order to facilitate gas exchange, the pulmonary vasculature has evolved an extremely thin endothelial monolayer. However, although this is beneficial in the physiologic state, in inflammatory states damage to this monolayer can cause life threatening pulmonary edema. Lastly, the pulmonary vasculature plays a key role in modulating inflammation and hemostasis. Although this is critical due to the proximity of the lung to the outside environment, these processes can become deregulated in disease states such as sepsis and DIC. Future work in the area aims to further characterize the mechanisms underlying each these processes and prevent the devastating consequences that can occur in disease states.