Cold Spring Harb Perspect Biol-2015-Pober

Inflammation and the Blood MicrovascularSystem

Jordan S. Pober1 and William C. Sessa2

1Department of Immunobiology, Yale University School of Medicine, New Haven,Connecticut 06520-8089

2Department of Pharmacology, Yale University School of Medicine, New Haven,Connecticut 06520-8089

Correspondence: [email protected]

Acute and chronic inflammation is associated with changes in microvascular form andfunction. At rest, endothelial cells maintain a nonthrombogenic, nonreactive surface at theinterface between blood and tissue. However, on activation by proinflammatory mediators,the endothelium becomes a major participant in the generation of the inflammatory re-sponse. These functions of endothelium are modified by the other cell populations of themicrovessel wall, namely pericytes, and smooth muscle cells. This article reviews recentadvances in understanding the roles played by microvessels in inflammation.

Inflammation, a major component of the in-nate immune response, is typically a localprocess historically characterized by cardinalfeatures, such as rubor (redness) and calor(warmth), both caused by increased bloodflow to the inflamed site, and tumor (swelling),caused by extravasation of fluid, plasma pro-teins, and leukocytes (changes attributable toactions of the local microvasculature). In thisarticle, we will review how blood microvesselsand the cells from which they are formed re-spond to innate inflammatory signals, leadingto inflammation. We will not discuss recentlyreviewed topics, such as the lymphatic circula-tion (Alitalo 2011) or role of blood vessels inadaptive immunity (Pober and Tellides 2012).We begin with an overview of the blood micro-vascular system.

ORGANIZATION OF THE BLOODMICROVASCULAR SYSTEM

The blood vascular system consists of two (sys-temic and pulmonary) closed loops organizedinto distinct segments. Large elastic arteriesarising from the heart give rise to smaller mus-cular arteries that convey blood to specific or-gans. Within the organs, arteries further arbor-ize, ending as arterioles having diameters in tensof micrometers. Arterioles, the most proximalsegments of the microvasculature, are internallylined by a monolayer of endothelial cells (ECs),connected to each other by intermixed tight andadherens junctions, that is invested by circum-ferentially arranged layers (lamellae) of vascularsmooth muscle cells (SMCs). ECs are separatedfrom the SMC layers by both a condensed layer

Editor: Ruslan M. Medzhitov

Additional Perspectives on Innate Immunity and Inflammation available at www.cshperspectives.org

Copyright# 2015 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a016345Cite this article as Cold Spring Harb Perspect Biol 2015;7:a016345

1

on May 27, 2015 - Published by Cold Spring Harbor Laboratory Press http://cshperspectives.cshlp.org/Downloaded from

of extracellular matrix enriched in type IV col-lagen and laminin, known as basement mem-brane (BM), and deep to the BM, by a thickconnective tissue band enriched in elastin fiberscalled the internal elastic lamina (IEL). The IELmarks the boundary between the arterial intimaand the media. The intima between the BM andIEL is formed by a small (in arterioles) zone ofloose connective tissue that may contain bothSMCs and leukocytes. Focal discontinuities inboth BM and the IEL allow ECs, which are elec-trically coupled to each other through gap junc-tions, to form gap junctions with underlyingSMCs (Behringer and Segal 2012). A less well-defined outer elastic lamina marks the outerboundary of the media, separating it from theparenchyma of the surrounding organ.

At their distal end, arterioles arborize fur-ther to give rise to capillaries. This portion ofthe microvasculature varies widely in structurefrom tissue to tissue but has the general featureof being formed by a series of single ECs, each ofwhich curves around to form a lumenized tubewith diameters of,10 mm. Capillaries are verynumerous and comprise the major surface forexchange of gases, fluid, and nutrients betweenblood and tissue. Heterogeneity of capillary ECsaffects the degree to which they limit or permitsuch exchanges (Aird 2007a,b). Proteins maypass either through the junctions between cap-illary ECs (paracellular transit) or through theECs themselves (transcellular transit via vesicles,channels, or fenestrae). Highly impermeantcapillaries, such as those in the central nervoussystem (CNS), form many more tight junctionsbetween adjacent ECs than other capillaries,limiting paracellular passage of proteins, andlack fenestrae and channels, forcing all exchang-es through a limited vesicular transport system.At the other extreme, sinusoidal capillaries inliver, spleen, or bone marrow have open gapsbetween adjacent ECs. Some capillary ECs formfenestrae (holes) that allow transcellular passageof proteins. Fenestrae may be filled in by orga-nized protein structures called diaphragms,partially limiting protein transit. Capillariesare invested and supported by a single discon-tinuous layer of contractile cells known as peri-cytes (PCs) (Dore-Duffy and Cleary 2011). PCs

are arrayed longitudinally along the capillaryand eachPCmaycontactmultiple ECs.The ratioof PCs to ECs varies among different tissues,being highest at 1:1 in the CNS where perme-ability is the lowest. Unlike SMCs of arteriolesthat form their own layers of extracellular ma-trix typically enriched for type I collagen andelastin, PCs reside within the collagen IV- andlaminin-rich BMof the ECs. PCs lack tight junc-tions but may limit extravasation of plasmaproteins indirectly through influence on ECjunctions and/or BM composition (Goddardand Iruela-Arispe 2013).

At their termini, capillaries converge anddrain their contents into larger caliber vesselsknown as venules. Postcapillary venules furtherconverge into larger collecting vessels, ultimate-ly connecting the microvasculature to veins thatdrain blood from the organ and return it to theheart. The EC lining of larger venules and veinsis invested by layers of vascular SMCs, althoughthese are typically less well organized into dis-tinct lamellae and are fewer in number thanthose found in arterioles and the IEL is less wellformed. In contrast, postcapillary venules, likecapillaries, are primarily invested by PCs ratherthan SMCs. The ECs of the postcapillary ve-nules are connected to each other by adherensjunctions and lack tight junctions (Goddardand Iruela-Arispe 2013), making them moreintrinsically leaky than the continuous capillar-ies that empty into them (Rous and Smith1931). The capillaries themselves may be leakierat their venular end than they are at the arteri-olar origin. In the skin, the BM of postcapillaryvenules has a distinct appearance by transmis-sion electron microscopic, forming lamellae asopposed to the more homogeneous BM of ar-terioles and capillaries (Braverman 1989), im-plying a difference in composition and/or or-ganization that may influence EC functions.

The embryological origin of the three prin-cipal cell types of the microvasculature are dis-tinct. Almost all ECs in the adult organism cantrace their lineage back to angioblasts arising inthe blood islands of the aorto-gonadal-meso-nephric region that migrate and differentiateinto endothelial progenitor cells (EPCs) and/or fully differentiated ECs. Consequently, the

J.S. Pober and W.C. Sessa

2 Cite this article as Cold Spring Harb Perspect Biol 2015;7:a016345


tissue-specific features of capillary ECs developin response to local environmental cues ratherthan to distinct embryological origins (Aird2007a,b). The formation of new blood vesselsin settings of chronic inflammation begins withoutgrowth and replication of differentiated EClining preexisting microvessels, although EPCs,probably residing within the vessel wall, mayalso contribute. Vascular SMCs of large vesselsare more heterogeneous in their origin. Some,as in the proximal aorta, arise from the secondheart field or neural crest, whereas others de-velop from local mesenchyme, often maintain-ing embryological patterns of Hox gene expres-sion (Majesky 2007). Both SMCs and PCs of themicrovasculature are probably derived from lo-cal mesenchyme but SMCs and PCs within theCNS may derive from neural crest and, at somesites, PCsmay derive from ECs undergoing en-dothelial to mesenchymal cell transition (Ar-mulik et al. 2011). Little is known about tissue-specific differences among PCs and how thesemay affect the contributions of these mural cellsto inflammatory processes. Mesenchymal stemcells (MSCs, also called mesenchymal stromalcells) localize to the PC layer of microvessels(Crisan et al. 2008), but it is not clear whetherall PCs have multipotency.

HOMEOSTATIC FUNCTIONS OF THEVASCULATURE

The vascular systemunder normal circumstanc-es performs four major homeostatic functions:(1) it keeps blood fluid; (2) it regulates perfu-sion of different organs; (3) it prevents inappro-priate activation of leukocytes; and (4) it regu-lates permselective exchange ofmacromoleculesbetween blood and tissues. The prevention ofcoagulation is a general feature of the wholevascular system and will be discussed here.The other homeostatic properties of the vascu-lature are typically assigned to specific micro-vascular segments and will be discussed in sub-sequent sections.

ECs prevent intravascular coagulation byseveral mechanisms. First, ECs sequester phos-phatidylserine (PS) to the inner leaflet of theirplasma membrane, depriving circulating clot-

ting factors of a surface required for assemblyinto functional complexes. Second, resting ECsexpress proteins that inhibit coagulation, nota-bly tissue factor pathway inhibitor (TFPI) andthrombomodulin. TFPI prevents tissue factorsfrom capturing and accelerating the catalyticactivity of factor VIIa, the initiator of the intrin-sic coagulation cascade. Thrombomodulin cap-tures active thrombin and alters its substratespecificity from a procoagulant protease thatconverts fibrinogen to fibrin to an anticoagulantprotease that cleaves and activates proteinC. Activated protein C, when bound to its re-ceptor on ECs and in combination with proteinS (made by ECs among other cell types), cleavesand inactivates various coagulation factors,such as factor V (Bouwens et al. 2013). Third,ECs also express proteoglycans bearing heparinsulfate glycosaminoglycans (GAGs) that cap-ture and activate antithrombin III, creating asubstrate trap for active thrombin. Fourth, rest-ing ECs synthesize plasminogen activators(both tissue type and urokinase type) as wellas the receptor for urokinase-type plasminogenactivator, converting circulating plasminogen toplasmin, a protease that cleaves fibrin and lysesincipient thrombi. Fifth, ECs prevent plateletactivation by inhibiting thrombin, by prevent-ing contact with BM or interstitial collagens,and by degrading extracellular adenosine-50-tri-phosphate (ATP). Resting ECs synthesize vonWillebrand factor (vWf ), a platelet adhesivemolecule important for platelet adhesion in set-tings of high shear stress, but sequester it inter-nally within storage granules known asWeibelPalade bodies (WPB) where it is inaccessible toplatelets. Finally, resting ECs synthesize and re-lease small quantities of PGI2 and nitric oxide(NO) sufficient to inhibit platelet activation byraising intracellular cAMP and cGMP, respec-tively. Inhibition of platelet activation by thesemediators may act synergistically.

ARTERIOLES AND CONTROLOF BLOOD FLOW

Themajor roles of arterioles are to control pres-sure, flow, and nutrient delivery to the capillarybeds. Arborizing branches of arterioles provide

Microvessels and Inflammation

Cite this article as Cold Spring Harb Perspect Biol 2015;7:a016345 3


a physical buffer or resistance to normalize pres-sure gradients generated throughout the cardiaccycle. As pressure increases into proximal arter-ies during cardiac contraction, distal arterioleswill vasoconstrict to limit pressure and flow intothe microcirculation through a mechanismcalled myogenic vasoconstriction. This arteri-ole, smooth muscle intrinsic response is criticalfor maintaining constant flow to vital organs,such as the brain, kidney, and heart and servesas amechanism for reducing pressure intomorestructurally fragile capillary beds. In addition tomyogenic constriction in arterioles, ECs in ar-teries and arterioles can sense changes in flowand release local autacoids, such as NO, lipidmetabolites, and other mediators to increaseor decrease vessel diameter. During acute in-flammatory responses, leukocyte-derivedmedi-ators such histamine and bradykinin will causearteriolar dilation thereby increasing blood flowleading to rubor. Histamine-induced arteriolardilation is abrogated in mice lacking the endo-thelial nitric oxide synthase gene (eNOS) (Payneet al. 2003); however, other vasodilatory medi-ators, such as PGI2 and endothelium-derivedhyperpolarization factors may contribute toenhanced blood flow in response to proinflam-matory molecules. The increase in flow andpressure will increase intravascular hydrostaticpressure providing a gradient for the extravasa-tion of fluid and protein through postcapillaryvenules into tissue. This occurs simultaneouslywith retraction of EC lining the venules as high-lighted below.

POSTCAPILLARY VENULES, PLASMAPROTEIN EXTRAVASATION, ANDLEUKOCYTE RECRUITMENT

In their basal state, venular ECs form a barriersufficient to retain most proteins, including al-bumin (Dejana and Giampietro 2012). The on-cotic effect of these plasma proteins serves tolimit the extravasation of fluid. Resting ECsalso fail to recruit leukocytes, largely attributedto the absence of luminal molecules capableof capturing these cells from the circulation.Similarly, leukocyte-activating chemokines arealso absent on resting ECs. However, certain

leukocyte adhesion molecules (e.g., P-selectin[CD62P]) and some chemokines (e.g., interleu-kin [IL]-8 [CXCL8], MCP-1 [CCL2], and eo-taxin 3 [CCL26]), are expressed in resting ECs;they are sequestered within WPB and, thus, un-available to circulating leukocytes (Rondaijet al. 2006). Furthermore, basal NO produc-tion in ECs can potentially exert anti-inflamma-tory effects by inhibiting activation of leuko-cytes.

Changes in venular ECs alter the behaviorof these cells to promote inflammation; suchchanges have been denoted as type I and/ortype II activation (Pober and Sessa 2007).Many mediators of type I activation are vaso-active autacoids, such as histamine, that signalthroughG-protein-coupled receptors (GPCRs).Histamine receptors are more highly expressedon venular ECs than they are on arteriolar orcapillary ECs (Heltianu et al. 1982), althoughhistamine can stimulate vasodilator productionfrom arteriolar ECs. The EC signaling pathwaysactivated by thesemediators have been reviewedelsewhere (Pober and Sessa 2007). Venular ECresponses to histamine include: transient con-traction of ECs, creating intercellular gaps andleading to parcellular escape of plasma proteins(Majno et al. 1961); regulated exocytosis of thecontents of WPB, bringing molecules, such asvWf, P-selectin, and certain stored chemokines(IL-8,MCP-1, eotaxin-3) to the luminal cell sur-face (Lorant et al. 1991); and synthesis of lipidmediators, such as platelet-activating factor(PAF), a potent activator of leukocytes as wellas platelets. Coexpression of leukocyte-bindingproteins (P-selectin) and leukocyte-activatingmolecules (PAF, chemokines) on the EC plasmamembrane has been described as a juxtacrinesignaling that, invitro, cancapture and stimulatethe transendothelial extravasation of neutro-phils (Lorant et al. 1991). However, GPCR-in-duced signals are self-limited in duration to15 min and the same signals induce synthesisof other mediators that inhibit leukocyte activa-tion, such as NO. Consequently, the result ofhistamine injection is simply a transient vaso-dilation and vascular leak (wheal and flare)that quickly resolves without significant leuko-cyte recruitment.




Sustained tissue swelling and significantleukocyte recruitment is dependent on type IIactivation of venular EC, mediated by inflam-matory cytokines. The prototypical examples ofsuch cytokines are IL-1 (both IL-1a and IL-1b)and tumor necrosis factor (TNF, often designat-ed as TNF-a). ECs express both the signalingIL-1 receptor (IL-1R1, designatedCD121a) and,in the resting state, TNF receptor 1 (TNFR1 orCD120a) but not TNFR2 (CD120b) (Al-Lamkiet al. 2005). The binding of IL-1 or TNF to thesereceptors activates transcription factors, there-by increasing expression of mRNAs encodingproteins that capture and activate leukocytes.For example, TNF or IL-1 induce de novoexpression of leukocyte adhesion moleculesE-selectin (CD62E) and vascular cell adhesionmolecule (VCAM)-1 (CD104) and increase ex-pression of intercellular adhesion molecule(ICAM)-1 (CD54) from a low but detectablebasal state (Pober and Sessa 2007). In mice,but not humans, P-selectin transcription is in-duced as well (Pan et al. 1998). TNFalso inducesmiRs that feedback and limit E-selectin andICAM-1 expression (Suarez et al. 2010). TNFand IL-1 also induce synthesis of chemokines,notably of IL-8 and MCP-1, the principal hu-man chemokines that activate neutrophils andmonocytes, respectively. The de novo synthesisof these chemokines is independent of the re-lease of prestored chemokines fromWPB, a typeI activation responses, as inflammatory cyto-kines do not cause WPB exocytosis (Zavoicoet al. 1989). Chemokines released from EC orfrom leukocytes within the perivascular spacebind to GAG expressed on the EC luminal sur-face (Mortier et al. 2012), which, along withleukocyte adhesion molecules, provide sus-tained juxtacrine signaling that leads to leuko-cyte infiltration in vivo. Fractalkine (CX3CL-1)represents a variation on this theme in that thechemokine moiety that interacts with its recep-tor, predominantly expressed on a subset ofmonocytes, is synthesized and expressed as theamino-terminal domain of an EC proteoglycan(Imaizumi et al. 2004). Finally, it should benoted that, because the signaling pathways acti-vated by IL-1 binding to its receptor are alsoactivated by Toll-like receptors (TLRs), with

the exception of TLR3, and because human vas-cular EC express TLR1, TLR2, TLR4, TLR6, andTLR9, engagement of these receptors by appro-priate ligands, typically one or more pathogen-associated molecular patterns or damage-asso-ciated molecular patterns, can produce type IIactivation responses (Opitz et al. 2007).

The general model for leukocyte recruit-ment by type IIactivated venular ECs involvesamultistep cascade (Ley et al. 2007). In humans,E-selectin typically mediates the initial tether-ing of the circulating neutrophil or monocyteby recognizing specific carbohydrate determi-nants, which in the case of neutrophils are at-tached to L-selectin expressed on the tips ofmicrovillous projections. These adhesive inter-actions are of low affinity but are rapidlyformed. Flowing blood pushes the tethered leu-kocyte, breaking selectin-mediated attachmentsthat subsequently rapidly reform with new E-selectin molecules as the leukocyte is displacedin the direction of flow, resulting in leukocyterolling. Monocytes and some T cells may in-stead (or additionally) be tethered and roll us-ing leukocyte integrin VLA-4 (CD49d/CD29),that in its low affinity state, can also rapidlyform weak attachments to endothelial VCAM-1. Some birectional signals may be transmittedduring these adhesive interactions, but theseare generally thought to provide insufficient ac-tivation of the leukocyte to progress to the nextstep in the cascade, known as firm adhesion.Instead, specific GPCRs on the rolling leukocytemust encounter a cognate ligand, typically anEC-bound chemokine, resulting in signalingthat causes the leukocyte to increase affinity ofits integrins for EC ligands, specifically LFA-1(CD11a/CD18) andMac-1 (CD11b/CD18) forICAM-1orVLA-4 (CD49d/CD29) forVCAM-1,changing from a round cell to one spread out onthe surface of the ECs. These firmly adherentleukocytes acquire motility and use these sameintegrins and EC ligands to crawl toward ECjunctions. At or near the junctions, the EC mayproject its luminal plasma membrane upward,forming an adhesion cup or docking struc-ture that partially engulfs the bound leukocyteand within which EC adhesion molecules, spe-cifically ICAM-1 and VCAM-1, cluster (Bar-




reiro et al. 2004). The clustering of ICAM-1 andpossibly VCAM-1 induces the membrane re-modeling to occur at the EC junction, facil-itating a path for transendothelial migration.At the same time, the leukocyte projects cyto-solic processes (invadosomes) that push be-tween or through the ECs, initiating the processof extravasation (Carman et al. 2007).

Transit across the EC lining of the venuleinvolves sequential molecular interactions withplatelet/endothelial cell adhesion molecule PE-CAM-1 (CD31) andCD99, each of which formshomophilic adhesions with the same moleculesexpressed on the leukocyte (Muller 2011). VE-cadherin and certain tight junction proteins(e.g., junctional adhesion molecule [JAM] Aor C) on the ECmay also play a role at this stage.Unlike ICAM-1 and VCAM-1, the total levels ofEC expression of PECAM-1 and CD99 are notincreased by inflammatory cytokines. However,ECs sequester these molecules in perijunctionalvesicles denoted as the lateral border recyclingcompartment. Attachment of leukocytes at ornear the junctions induces the EC to bring PE-CAM-1 and CD99 to the surface of the junctionor to the membrane of transcellular channelsthat form near the junction, lining the paththrough which the leukocyte traverses the ECmonolayer. The signal transmitted to the ECsthat leads to increased surface expression of PE-CAM-1 and CD99 may be clustering of mole-cules, such as ICAM-1.

Once through theEC layer, the leukocyte stillmust traverse the PC layer and the BM in whichthe PCs are embedded. PCs also respond to in-flammatory cytokines, but their levels of in-duced expression of adhesionmolecules, largelyrestricted to ICAM-1 andVCAM-1, are less thanon ECs (Ayres-Sander et al. 2013). Consequent-ly, cultured PCmonolayers support only limitedtransit of neutrophils, although neutrophils thattraverse an EC monolayer become altered in anunspecified manner that increases their abilityto traverse PC monolayers. PCs do expressCD99, but not PECAM-1, and CD99 may alsobe involved in traversing the PC layer of the ven-ular wall. In addition, PCs are contractile cellsthat anchor to BM proteins and canmanipulatethe organization of the BM; PCs also steer ex-

travasating leukocytes to regions in which theBM is more attenuated, known as low expres-sion regions (Nourshargh et al. 2010). Howev-er, the detailed functions of PCs in the process ofleukocyte extravasation are less well understoodthan those of ECs.

During inflammation, plasma protein andleukocyte extravasation both occur at the samesites. The extravasation of leukocytes throughan EC monolayer in vitro can occur withoutinducing inter-EC gaps that result in increasedparacellular leak of macromolecules (Huanget al. 1988). The increased paracellular leak invivo probably results directly from inflammato-ry cytokine exposure rather than through gapsopened by leukocytes. The extravasated plasmaproteins form a provisional matrix within thetissue that can support attachment, survival,andmigration of subsequently extravasated leu-kocytes. In cell culture models, EC leaks in re-sponse to IL-1 or TNF occur in two stages: anearly but transient leak that may be mediated bycytoskeletal contraction induced by activationof a small G protein, likely Arf6 (Zhu et al.2012), and a later, more pronounced and sus-tained leak dependent on new protein synthesis(Clark et al. 2007). The newly synthesized pro-tein(s) responsible for increased leakiness havenot been defined.

Entry of leukocytes into the CNS has uniquefeatures that differ from extravasation elsewherein the body (Man et al. 2007). Within the CNS,ECs form numerous intercellular tight junc-tions, the PC to EC ratio reaches 1:1, the highestin the body, and astrocyte and glial foot process-es abut on the EC/PC BM, creating an addi-tional barrier referred to as the glia limitans.As a result, resting CNS microvessels effectivelyprevent all protein transit, creating the bloodbrain barrier. It is unknown to what extentthese features extend to the postcapillary ve-nules of the CNS, but basal trafficking of leuko-cytes is very low. Inflammation may disrupt thebloodbrain barrier and passage of leukocytesthrough the vessel wall seems to be particularlydependent on leukocyte integrin VLA-4 (capa-ble of recognizing cellular fibronectin as well asVCAM-1). Low-affinity VLA-4 interactions withVCAM-1 can support leukocyte tethering and




rolling, possibly explaining why knockout ofselectins has little effect on entry of leukocytesinto the CNS in mouse models of neuroinflam-mation. LFA-1 is also important for entry intothe CNS, and it may engage ICAM-2, a consti-tutively expressed EC molecule, as well asICAM-1 during crawling to the EC junctions.LFA-1 may also engage JAM-A, a tight junctionprotein during transendothelial migration.Those cells that do get through the EC/PC/BM barrier may still be prevented from enteringthe CNS because of the glia limitans, accumu-lating in the so-called VirchowRobin spaceadjacent to the blood vessels. Leukocytes mayalso bypass the CNS vasculature to enter intothe brain by crossing the more permeable cho-roid plexus where blood is filtered to producecerebrospinal fluid.

In skin as in most other tissues, inflamma-tory cytokines induce E-selectin and VCAM-1on venular ECs but not adjacent capillary ECs;capillary ECs are responsive to these same cyto-kines as shown by selective up-regulationICAM-1 expression (Enis et al. 2005). This dif-ference accounts for why leukocytes extravasatethrough venules and generally not capillaries,but the basis of this restriction of the responseto cytokines is unknown. Possible contributorsare differences in the shear stress detected by theECs (high in capillaries, low in venules) anddifferences in BM composition. Application ofshear stress to cultured microvascular ECs in-duces expression of the transcription factorKLF-4, which limits cytokine-induced adhesionmolecule expression (Clark et al. 2011). Asnoted previously, the venular BM in skinmicro-vessels has a distinct appearance by transmis-sion electron microscopy from that of capillar-ies and in psoriatic lesions, when capillary loopsin the dermal papillae are remodeled to becomevenules, these microvessels change the ap-pearance of their BM concomitantly with thecapacity of their ECs to express E-selectin andVCAM-1 (Petzelbauer et al. 1994). These dataare consistent with the hypothesis that attach-ment to BM of different composition or orga-nization can influence EC responses to cyto-kines. It should be noted that there areexceptions to the primary role of venules for

extravasation. For example, in the lung, leuko-cytes exit into the bronchial wall via venules ofthe bronchial circulation but enter into the al-veolus via alveolar capillaries of the pulmonarycirculation. The capillary tufts within the renalglomeruli and the sinusoidal capillaries of theliver can also express adhesion molecules andsupport leukocyte extravasation.

The various stimuli that evoke inflamma-tion may elicit inflammatory infiltrates that areenriched for particular types of leukocytes andthesemay change over time. Such selectivity andits evolution can often be explained by changesin the adhesion molecules and chemokines dis-played on the EC surface. For example, E-selec-tin is more rapidly synthesized and displayedthan is VCAM-1 or ICAM-1, but its expressionon the cell surface is generally more transient,peaking at 46 h and falling to low levels by24 h, and better correlates with neutrophilthanmononuclear cell capture and recruitment,perhaps because mononuclear cells are moreadept at tethering and rolling on VCAM-1. IL-4augments VCAM-1 expression by EC and alsocauses ECs to synthesize eotaxin-3, a chemokinethat favors recruitment of eosinophils. Thus, thecombination of TNF plus IL-4 can lead to in-flammatory infiltrates that are enriched for thiscell type (Briscoe et al. 1992). This kind of spe-cialized inflammation has historically been as-sociated with adaptive immunity in whichCD4, Th-1, Th-2, and, more recently, Th-17cells favor their own recruitment and specifictypes of effector cells and, hence, are outsidethe scope of this review. However, it has recentlybeen appreciated that other cell types, such asmast cells, basophils, and innate lymphocytes,also may display polarized cytokine profilesand elicit particular inflammatory patternsthat are independent of antigen (Bochner andSchleimer 2001). It is likely that ECs will play arole in shaping the nature of these types of in-flammatory reactions, although little is knownto date what signals are involved. IL-17, for ex-ample, has little effect on ECs by itself but maymodulate other cytokine responses (Berninket al. 2013). Interestingly, IL-17 does have directeffects on PC production of chemokines, simi-lar to previous observations using cultured




SMCs (Eid et al. 2009), and a more integratedview of venular activation rather than EC acti-vation may be required to understand how in-flammatory reactions evolve and differentiate.

VASCULAR DYSFUNCTION: FLOWDYSREGULATION, THROMBOSIS,AND CAPILLARY LEAK

As we noted earlier in this article, changes inECs induced by vasoactive autacoids (type I ac-tivation) and inflammatory cytokines (type IIactivation) contribute to inflammation. Be-cause resting ECs actively resist the developmentof inflammation, these responses may be seen asinterfering with normal (basal) EC function.Alternatively, such changes can be viewed as be-ing adaptive because local inflammation is animportant mechanism of host defense and ho-meostasis. However, there are changes that oc-cur in ECs that disrupt homeostasis withoutbeneficial effect to the host. Some of these aresimply exaggerated or inappropriate forms ofactivation, whereas others result from injuryand have been linked together under the cate-gory of endothelial dysfunction (Pober et al.2009). As discussed above, tissue perfusion iscontrolled by SMC tone in the terminal arte-rioles. SMCs normally respond to EC-derivedsignals, neural signals, and humoral (hormon-al) signals. Perfusion to a specific tissue canbe inappropriately diminished when arteriolarSMCs in the microvasculature of a particulartissue become refractory to vasodilatory signalsand/or because EC-derived vasodilatory signalsare decreased. A common pattern of EC dys-function occurs when EC lose the capacity tosynthesize NO because of diminished expres-sion of eNOS owing to TNF-mediated destabi-lization of its mRNAmediated by TNF-inducedmiR155 (Sun et al. 2012). Additionally, oxida-tive stress associated with inflammation canturn eNOS into a generatorof superoxide anion,a change called eNOS uncoupling thereby re-ducing NO bioactivity independent of changesin eNOS mRNA or protein levels (Forstermannand Sessa 2012). TNF may also cause ECs toincrease synthesis of the vasoconstrictor peptide

endothelin-1 (Marsden andBrenner 1992). Par-adoxically, because flow is regulated by bloodpressure in addition to vascular resistance, aglobal decrease in blood pressure caused bywidespread vasodilation throughout the circu-latory system, as occurs in septic shock,may alsolead to hypoperfusion. A possible explanationis that TNF increases production of PGI2 bySMCs, as well as ECs, possibly through induc-tion of cyclo-oxygenase 2 (Jimenez et al. 2005).

Thrombosis is another potential exampleof EC dysfunction (Pober et al. 2009). Localintravascular coagulation may be a means ofpreventing hematogenous dissemination of aninfection and thus part of innate immunity.However, thrombosis can also produce tissueinfarcts and, when widespread, paradoxicalbleeding caused by consumption of clotting fac-tors (Esmon 2004). ECs may contribute to thisby shedding microparticles, for instance plasmamembranederived vesicles with surface-ex-posed PS and thus can serve as a platform forassembly of coagulation factors. TNF-activatedECsmay also lose expression of thrombomodu-lin throughmRNAdestabilization andmay losetheir anticoagulant heparin sulfates through cy-tokine-induced enzymatic degradation. Tissuefactor pathway inhibitor may also be down-reg-ulated at the same time that tissue factor is syn-thesized and de-encrypted, the latter processdescribing transfer to a location, such as to shedmicroparticles, where it may encounter and cat-alytically enhance factor VIIa. The net effect ofthese changes is to enhance thrombin genera-tion and to allow thrombin to cleave fibrinogento fibrin. At the same time, ECs or other vascu-lar cells may enhance production of plasmin-ogen activator inhibitors, reducing the localcapacity of ECs to activate plasminogen to plas-min, and lyse fibrin thrombi as they are formed.

We noted earlier that venular leak of plasmaproteins plays an important role in the inflam-matory process, providing the components of aprovisional matrix to support extravasating leu-kocytes. We also noted that, in most tissues,capillaries form a much larger surface for ex-change of nutrients and wastes between tissuesand blood. Under normal circumstances, capil-laries (and arterioles) do not leak as a part of the




inflammatory process. This difference may re-late to the observation that capillaries (and ar-terioles) form tight junctions between adjacentECs, whereas venules do not (Simionescu et al.1975). The process of disassembly of tight junc-tions may be viewed as pathological and as anexample of dysfunction. Capillary leak is a char-acteristic feature of sepsis and its developmentis one of the causes of organ failure in thatsyndrome (Gustot 2011). The mechanisms bywhich mediators of sepsis open tight junctionsis unknown and may be distinct from the pro-cess by which venular adherens junctions aredisassembled. Capillary leak may also resultfrom EC injury and death (Joris et al. 1990).

ANGIOGENESIS AND THE VASCULARSYSTEM IN CHRONIC INFLAMMATION

Mononuclear leukocytes (monocytes) are re-cruited to sites of inflammation where theymay differentiate into macrophages and pro-mote angiogenesis. Angiogenesis is defined asthe migration and proliferation of ECs liningvenules into surrounding tissue resulting inthe formation of a capillary plexus. The recruit-ment ofmacrophages to sites of inflammation iscritical for the resolution of inflammation.However, if the signal inducing the acute in-flammatory response is not eradicated, chronicinflammation may ensue, and there is evidencethat the transition from acute to chronic in-flammation relies on an angiogenic responseas a means to provide blood supply to inflamedneotissue. In this context, sustained angiogene-sis may amplify the extent of macrophage infil-tration and edema, and worsen tissue damage.Indeed, strategies aimed at inhibiting macro-phage subsets and/or angiogenesis can reducethe degree of inflammation in several preclinicalmodel systems. Vascular endothelial cell growthfactor-A (VEGFA) and TNF derived frommacrophages are potent angiogenic factorsand inducers of inflammation-driven vascularremodeling in a variety of inflammatory diseas-es, such as psoriasis, rheumatoid arthritis, in-flammatory bowel disease, and asthma. BothVEGF and TNF are produced by activated mac-rophages and bind to their cognate receptors on

endothelium to stimulate angiogenesis. VEGFbinds to VEGF receptor 2 (VEGFR2) promot-ing angiogenesis (Baer et al. 2013), whereasTNF binds to TNFR2, which is induced onECs during the early stages of inflammationand mediates its proangiogenic function (Luoet al. 2006). Moreover, TNF can prime endothe-lium by up-regulating VEGF receptor 2 (Sain-son et al. 2008). Additional inflammatory me-diators can also promote angiogenesis includeIL-1, IL-6, IL-17, and IL-18, and there is sub-stantial cross talk between these cytokines andVEGF/TNF signaling pathways (Huggenbergerand Detmar 2011). Chemokines play an impor-tant role both as agents that recruit macro-phages and as agents produced by macrophages(or by other cells, including ECs and PCs) thatmay then promote or inhibit angiogenesis(Owen and Mohamadzadeh 2013).

CONCLUDING REMARKS

The blood vascular system has evolved to con-trol tissue homeostasis. Under normal con-ditions, the vasculature maintains a quiescentinterface between blood and tissue. On encoun-tering an inflammatory insult, the endotheliumactively participates in controlling blood flow,permeability, leukocyte infiltration, and tissueedema, changes that serve to eradicate the initialstimulus. PCs and SMCs also participate inthese processes, but their roles are less well un-derstood than those of ECs. If the stimulus per-sists, inflammation evolves into a chronic phase.Sustained activation of the endothelium causedby persistent inflammation can promote mac-rophage recruitment and neotissue angiogene-sis to sustain blood flow.Understanding the vas-cular changes that occur during the transitionfrom acute to chronic inflammation is crucialfor our development of novel therapeutic ap-proaches associatedwith debilitating inflamma-tory diseases.

REFERENCES

Aird WC. 2007a. Phenotypic heterogeneity of the endothe-lium: I. Structure, function, and mechanisms. Circ Res100: 158173.




Aird WC. 2007b. Phenotypic heterogeneity of the endothe-lium: II. Representative vascular beds. Circ Res 100: 174190.

Alitalo K. 2011. The lymphatic vasculature in disease. NatMed 17: 13711380.

Al-Lamki RS, Wang J, Vandenabeele P, Bradley JA, Thiru S,Luo D, Min W, Pober JS, Bradley JR. 2005. TNFR1- andTNFR2-mediated signaling pathways in human kidneyare cell type-specific and differentially contribute to renalinjury. FASEB J 19: 16371645.

Armulik A, Genove G, Betsholtz C. 2011. Pericytes: Devel-opmental, physiological, and pathological perspectives,problems, and promises. Dev Cell 21: 193215.

Ayres-Sander CE, Lauridsen H, Maier CL, Sava P, Pober JS,Gonzalez AL. 2013. Transendothelial migration enablessubsequent transmigration of neutrophils through un-derlying pericytes. PLoS ONE 8: e60025.

Baer C, Squadrito ML, Iruela-Arispe ML, De Palma M.2013. Reciprocal interactions between endothelial cellsand macrophages in angiogenic vascular niches. ExpCell Res 319: 16261634.

Barreiro O, Vicente-ManzanaresM, Urzainqui A, Yanez-MoM, Sanchez-Madrid F. 2004. Interactive protrusive struc-tures during leukocyte adhesion and transendothelialmigration. Front Biosci 9: 18491863.

Behringer EJ, Segal SS. 2012. Spreading the signal for vaso-dilatation: Implications for skeletal muscle blood flowcontrol and the effects of ageing. J Physiol 590: 62776284.

Bernink J, Mjosberg J, Spits H. 2013. Th1- and Th2-likesubsets of innate lymphoid cells. Immunol Rev 252:133138.

Bochner BS, Schleimer RP. 2001. Mast cells, basophils, andeosinophils: Distinct but overlapping pathways for re-cruitment. Immunol Rev 179: 515.

Bouwens EA, Stavenuiter F, Mosnier LO. 2013. Mechanismsof anticoagulant andcytoprotective actionsof theproteinCpathway. J Thromb Haemost 11: 242253.

Braverman IM. 1989. Ultrastructure and organization of thecutaneous microvasculature in normal and pathologicstates. J Invest Dermatol 93: 2S9S.

Briscoe DM, Cotran RS, Pober JS. 1992. Effects of tumornecrosis factor, lipopolysaccharide, and IL-4 on the ex-pression of vascular cell adhesion molecule-1 in vivo.Correlation with CD3 T cell infiltration. J Immunol149: 29542960.

Carman CV, Sage PT, Sciuto TE, de la Fuente MA, Geha RS,Ochs HD, Dvorak HF, Dvorak AM, Springer TA. 2007.Transcellular diapedesis is initiated by invasive podo-somes. Immunity 26: 784797.

Clark PR, Manes TD, Pober JS, Kluger MS. 2007. IncreasedICAM-1 expression causes endothelial cell leakiness, cy-toskeletal reorganization and junctional alterations. J In-vest Dermatol 127: 762774.

Clark PR, Jensen TJ, KlugerMS,MorelockM, Hanidu A, QiZ, Tatake RJ, Pober JS. 2011. MEK5 is activated by shearstress, activates ERK5 and induces KLF4 to modulateTNF responses in human dermalmicrovascular endothe-lial cells. Microcirculation 18: 102117.

Crisan M, Yap S, Casteilla L, Chen CW, Corselli M, Park TS,Andriolo G, Sun B, Zheng B, Zhang L, et al. 2008. A

perivascular origin for mesenchymal stem cells in multi-ple human organs. Cell Stem Cell 3: 301313.

Dejana E, Giampietro C. 2012. Vascular endothelial-cad-herin and vascular stability. Curr Opin Hematol 19:218223.

Dore-Duffy P, Cleary K. 2011. Morphology and propertiesof pericytes. Methods Mol Biol 686: 4968.

Eid RE, Rao DA, Zhou J, Lo SF, Ranjbaran H, Gallo A, SokolSI, Pfau S, Pober JS, Tellides G. 2009. Interleukin-17 andinterferon-g are produced concomitantly by human cor-onary artery-infiltrating T cells and act synergistically onvascular smooth muscle cells. Circulation 119: 14241432.

Enis DR, Shepherd BR, Wang Y, Qasim A, Shanahan CM,Weissberg PL, Kashgarian M, Pober JS, Schechner JS.2005. Induction, differentiation, and remodeling ofblood vessels after transplantation of Bcl-2-transducedendothelial cells. Proc Natl Acad Sci 102: 425430.

Esmon CT. 2004. The impact of the inflammatory responseon coagulation. Thromb Res 114: 321327.

Forstermann U, Sessa WC. 2012. Nitric oxide synthases:Regulation and function. Eur Heart J 33: 829837,837a837d.

Goddard LM, Iruela-Arispe ML. 2013. Cellular and molec-ular regulation of vascular permeability. Thromb Hae-most 109: 407415.

Gustot T. 2011. Multiple organ failure in sepsis: Prognosisand role of systemic inflammatory response. Curr OpinCrit Care 17: 153159.

Heltianu C, SimionescuM, Simionescu N. 1982. Histaminereceptors of the microvascular endothelium revealed insitu with a histamine-ferritin conjugate: Characteristichigh-affinity binding sites in venules. J Cell Biol 93:357364.

Huang AJ, Furie MB, Nicholson SC, Fischbarg J, LiebovitchLS, Silverstein SC. 1988. Effects of human neutrophilchemotaxis across human endothelial cell monolayerson the permeability of these monolayers to ions andmacromolecules. J Cell Physiol 135: 355366.

Huggenberger R, Detmar M. 2011. The cutaneous vascularsystem in chronic skin inflammation. J Investig DermatolSymp Proc 15: 2432.

Imaizumi T, Yoshida H, Satoh K. 2004. Regulation ofCX3CL1/fractalkine expression in endothelial cells. JAtheroscler Thromb 11: 1521.

Jimenez R, Belcher E, Sriskandan S, Lucas R, McMaster S,Vojnovic I, Warner TD, Mitchell JA. 2005. Role of Toll-like receptors 2 and 4 in the induction of cyclooxygen-ase-2 in vascular smooth muscle. Proc Natl Acad Sci 102:46374642.

Joris I, Cuenoud HF, Doern GV, Underwood JM, Majno G.1990. Capillary leakage in inflammation. A study by vas-cular labeling. Am J Pathol 137: 13531363.

Ley K, Laudanna C, CybulskyMI, Nourshargh S. 2007. Get-ting to the site of inflammation: The leukocyte adhesioncascade updated. Nat Rev Immunol 7: 678689.

Lorant DE, Patel KD, McIntyre TM, McEver RP, PrescottSM, Zimmerman GA. 1991. Coexpression of GMP-140and PAF by endothelium stimulated by histamine orthrombin: A juxtacrine system for adhesion and activa-tion of neutrophils. J Cell Biol 115: 223234.




Luo D, Luo Y, He Y, Zhang H, Zhang R, Li X, Dobrucki WL,Sinusas AJ, Sessa WC, Min W. 2006. Differential func-tions of tumor necrosis factor receptor 1 and 2 signalingin ischemia-mediated arteriogenesis and angiogenesis.Am J Pathol 169: 18861898.

Majesky MW. 2007. Developmental basis of vascularsmooth muscle diversity. Arterioscler Thromb Vasc Biol27: 12481258.

Majno G, Palade GE, Schoefl GI. 1961. Studies on inflam-mation: II. The site of action of histamine and serotoninalong the vascular tree: A topographic study. J BiophysBiochem Cytol 11: 607626.

Man S, Ubogu EE, Ransohoff RM. 2007. Inflammatory cellmigration into the central nervous system: A few newtwists on an old tale. Brain Pathol 17: 243250.

Marsden PA, Brenner BM. 1992. Transcriptional regulationof the endothelin-1 gene by TNF-a. Am J Physiol 262:C854C861.

Mortier A, Van Damme J, Proost P. 2012. Overview of themechanisms regulating chemokine activity and availabil-ity. Immunol Lett 145: 29.

MullerWA. 2011.Mechanisms of leukocyte transendothelialmigration. Annu Rev Pathol 6: 323344.

Nourshargh S,Hordijk PL, SixtM. 2010. Breachingmultiplebarriers: Leukocyte motility through venular walls andthe interstitium. Nat Rev Mol Cell Biol 11: 366378.

Opitz B, Hippenstiel S, Eitel J, Suttorp N. 2007. Extra- andintracellular innate immune recognition in endothelialcells. Thromb Haemost 98: 319326.

Owen JL, Mohamadzadeh M. 2013. Macrophages and che-mokines as mediators of angiogenesis. Front Physiol 4:159.

Pan J, Xia L,McEver RP. 1998. Comparison of promoters forthe murine and human P-selectin genes suggests species-specific and conserved mechanisms for transcriptionalregulation in endothelial cells. J Biol Chem 273: 1005810067.

Payne GW,Madri JA, SessaWC, Segal SS. 2003. Abolition ofarteriolar dilation but not constriction to histaminein cremaster muscle of eNOS2/2 mice. Am J PhysiolHeart Circ Physiol 285: H493H498.

Petzelbauer P, Pober JS, Keh A, Braverman IM. 1994. Induc-ibility and expression of microvascular endothelial adhe-sion molecules in lesional, perilesional, and uninvolvedskin of psoriatic patients. J Invest Dermatol 103: 300305.

Pober JS, Sessa WC. 2007. Evolving functions of endothelialcells in inflammation. Nat Rev Immunol 7: 803815.

Pober JS, Tellides G. 2012. Participation of blood vessel cellsin human adaptive immune responses. Trends Immunol33: 4957.

Pober JS, Min W, Bradley JR. 2009. Mechanisms of endo-thelial dysfunction, injury, and death. Annu Rev Pathol 4:7195.

Rondaij MG, Bierings R, Kragt A, van Mourik JA, VoorbergJ. 2006. Dynamics and plasticity of Weibel-Palade bodiesin endothelial cells. Arterioscler Thromb Vasc Biol 26:10021007.

Rous P, Smith F. 1931. The gradient of vascular permeability:III. The gradient along the capillaries and venules of frogskin. J Exp Med 53: 219242.

Sainson RC, Johnston DA, Chu HC, Holderfield MT, Na-katsu MN, Crampton SP, Davis J, Conn E, Hughes CC.2008. TNF primes endothelial cells for angiogenicsprouting by inducing a tip cell phenotype. Blood 111:49975007.

Simionescu M, Simionescu N, Palade GE. 1975. Segmentaldifferentiations of cell junctions in the vascular endothe-lium. The microvasculature. J Cell Biol 67: 863885.

Suarez Y, Wang C, Manes TD, Pober JS. 2010. Cutting edge:TNF-induced microRNAs regulate TNF-induced expres-sion of E-selectin and intercellular adhesion molecule-1on human endothelial cells: Feedback control of inflam-mation. J Immunol 184: 2125.

SunHX, Zeng DY, Li RT, Pang RP, Yang H, Hu YL, Zhang Q,Jiang Y, Huang LY, Tang YB, et al. 2012. Essential role ofmicroRNA-155 in regulating endothelium-dependentvasorelaxation by targeting endothelial nitric oxide syn-thase. Hypertension 60: 14071414.

Zavoico GB, Ewenstein BM, Schafer AI, Pober JS. 1989. IL-1and related cytokines enhance thrombin-stimulatedPGI2 production in cultured endothelial cells withoutaffecting thrombin-stimulated von Willebrand factor se-cretion or platelet-activating factor biosynthesis. J Immu-nol 142: 39933999.

Zhu W, London NR, Gibson CC, Davis CT, Tong Z, Soren-sen LK, Shi DS, Guo J, Smith MC, Grossmann AH, et al.2012. Interleukin receptor activates a MYD88ARNOARF6 cascade to disrupt vascular stability. Nature 492:252255.




October 23, 20142015; doi: 10.1101/cshperspect.a016345 originally published onlineCold Spring Harb Perspect Biol

Jordan S. Pober and William C. Sessa

Inflammation and the Blood Microvascular System

Subject Collection Innate Immunity and Inflammation

DiseaseGroup 2 Innate Lymphoid Cells in Health and

Brian S. Kim and David Artis

The Chemokine System in Innate ImmunityCaroline L. Sokol and Andrew D. Luster

SystemInflammation and the Blood Microvascular

Jordan S. Pober and William C. SessaIntracellular Nucleic Acid SensorsMicrobial Sensing by Toll-Like Receptors and

Surya Pandey, Taro Kawai and Shizuo Akira

Lymphohematopoietic InterfaceSinusoidal Immunity: Macrophages at the

Subhankar MukhopadhyaySiamon Gordon, Annette Plddemann and

Immunity and InflammationTumor Necrosis Factor Superfamily in Innate

John Sed, Vasileios Bekiaris and Carl F. Ware

Innately HomeostaticAllergic InflammationLaurence E. Cheng and Richard M. Locksley Transduction by Pattern-Recognition Receptors

Emerging Principles Governing Signal

Jonathan C. Kagan and Gregory M. Barton

Sensing and ResponseIntegration of Neural and Immune Pathogen Approaching the Next Revolution? Evolutionary

Kevin J. Tracey

Lipid Mediators in the Resolution of Inflammation

al.Charles N. Serhan, Nan Chiang, Jesmond Dalli, et

InflammasomesMarcel R. de Zoete, Noah W. Palm, Shu Zhu, et al. Responses

Transcriptional Control of Inflammatory

Stephen T. Smale and Gioacchino NatoliIL-6 in Inflammation, Immunity, and Disease

KishimotoToshio Tanaka, Masashi Narazaki and Tadamitsu

DNA Degradation and Its DefectsKohki Kawane, Kou Motani and Shigekazu Nagata

http://cshperspectives.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright 2015 Cold Spring Harbor Laboratory Press; all rights reserved


Documents

Cold Spring Harb Perspect Biol-2015-Pober