THE ENDOTHELIUM IN SEPSIS
Can Ince
*
, Philip R. Mayeux
†
, Trung Nguyen
‡
, Hernando Gomez
§
, John A. Kellum
§
, Gustavo
A. Ospina-Tascón
||
, Glenn Hernandez
¶
, Patrick Murray
**
, and Daniel De Backer
††
on behalf
of the ADQI XIV Workgroup
*
Department of Intensive Care, Erasmus MC, University Medical Center, Rotterdam, the
Netherlands
†
Department of Pharmacology and Toxicology University of Arkansas for Medical
Sciences, Little Rock, Arkansas
‡
Pediatric Critical Care Medicine, Department of Pediatrics,
Baylor College of Medicine, Texas Children’s Hospital, Houston, Texas
§
The Center for Critical
Care Nephrology Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh,
Pennsylvania
||
Department of Intensive Care, Fundación Valle del Lili, Universidad ICESI, Cali,
Columbia
¶
Department of Intensive Care Medicine, Pontificia Universidad Católica De Chile,
Santiago, Chile
**
University College Dublin, Dublin, Ireland
††
CHIREC Hospitals and Université
Libre de Bruxelles, Brussels, Belgium
Abstract
Sepsis affects practically all aspects of endothelial cell (EC) function and is thought to be the key
factor in the progression from sepsis to organ failure. Endothelial functions affected by sepsis
include vasoregulation, barrier function, inflammation, and hemostasis. These are among other
mechanisms often mediated by glycocalyx shedding, such as abnormal nitric oxide metabolism,
up-regulation of reactive oxygen species generation due to down-regulation of endothelial-
associated antioxidant defenses, transcellular communication, proteases, exposure of adhesion
molecules, and activation of tissue factor. This review covers current insight in EC-associated
hemostatic responses to sepsis and the EC response to inflammation. The endothelial cell lining is
highly heterogeneous between different organ systems and consequently also in its response to
sepsis. In this context, we discuss the response of the endothelial cell lining to sepsis in the kidney,
Address reprint requests to Can Ince, PhD, Department of Intensive Care, Erasmus MC University Hospital Rotterdam, s-
Gravendijkwal 2303015 CE Rotterdam, The Netherlands. c.ince@erasmusmc.nl.
ADQI XIV Workgroup: A complete list is provided in Appendix 1.
Can Ince and Daniel De Backer: Denotes workgroup facilitators.
Declaration of interest: C.I. has received research grants the Dutch Kidney Foundation (grant C 09.2290 and grant 14OIP11), Bussum,
The Netherlands. C.I. has also received honoraria and independent research grants from Fresenius-Kabi, Bad Homburg, Germany;
Baxter Health Care, Deerfield, Illinois and AM-Pharma, Bunnik, The Netherlands. C.I. has developed SDF imaging and is listed as
inventor on related patents commercialized by MicroVision Medical (MVM) under a license from the Academic Medical Center
(AMC). C.I. has received consultant fees from MVM in the past, but has not been involved with this company for more than five years
now, except that he still holds shares. Braedius Medical, a company owned by a relative of C.I., has developed and designed a hand-
held microscope called Cyto-Cam-IDF imaging; however, C.I. has no financial relation with Braedius Medical of any sort. P.M., T.N.,
G.O., H.G., and G.H. have no declared interests. J.K. has received consulting fees from Abbott, Aethlon, Alere, Alung, AM Pharma,
Astute Medical, Atox Bio, Baxter, Cytosorbents, venBio, Gambro, Grifols, Roche, Spectral Diagnostics, Sangart, and Siemens. J.K.
has also received research grants from Alere, Astute Medical, Atox Bio, Bard, Baxter, Cytosorbents, Gambro, Grifols, Kaneka, and
Spectral Diagnostics, and has licensed technologies through the University of Pittsburgh to Astute Medical, Cytosorbents, and Spectral
Diagnostics. P.M. has received consulting fees from AM Pharma, Abbvie, FAST Diagnostics. He has also received research funding
from Abbott, Alere, EKF Diagnostics. D.D.B. is a member of the Advisory Board Baxter-Gambro renal, and Nestlé Health Sciences.
He received research grants and material for studies from Edwards Lifesciences, Vytech, and Imacor. P.M., T.N., G.O., H.G., and G.H.
have no declared interests.
HHS Public Access
Author manuscript
Shock
. Author manuscript; available in PMC 2017 January 31.
Published in final edited form as:
Shock
. 2016 March ; 45(3): 259–270. doi:10.1097/SHK.0000000000000473.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
liver, and lung. Finally, we discuss evidence as to whether the EC response to sepsis is adaptive or
maladaptive. This study is a result of an Acute Dialysis Quality Initiative XIV Sepsis Workgroup
meeting held in Bogota, Columbia, between October 12 and 15, 2014.
Keywords
Barrier function; blood; endothelium; glycocalyx; hemostasis; inflammation; microcirculation;
sepsis
INTRODUCTION
The endothelial cell lining (ECL) of the vasculature is a unique cellular system that coats the
inside of blood vessels and forms the interface between the circulating blood and the
parenchymal cells responsible for organ function. It is critical for the regulation of
hemostasis, vasomotor control, and immunological function, by sensing and reaction
through secretion of molecules, which initiate transcellular and intra-cellular signaling. In
addition to these important functions, the endothelium forms the essential vascular barrier
for solute transport and osmotic balance. Sepsis is associated with severe endothelial cell
(EC) dysfunction leading to dysregulation of hemostasis and vascular reactivity, as well as
tissue edema. This failure of the ECL is considered central to the progression to organ
failure during sepsis.
This review discusses many of the latest insights into the physiological function of the ECL
and its dysfunction in sepsis. Since many excellent reviews on the septic endothelium have
preceded this study (1), we have focused our attention to studies published in the past 5
years. This endeavor arose as a part of an international Acute Dialysis Quality Initiative
(ADQI) XIV Sepsis Workgroup meeting held in Bogota, Columbia, between October 12 and
15, 2014, in which the authors participated. This meeting addressed the different cellular and
subcellular aspects of sepsis, and the working group of the present authors of this study
focused themselves on the role of endothelium in sepsis. In this review, we present a current
update on the central role of the endothelium in sepsis.
METHODS
Complete methods are available in the companion article to this series (2). Briefly, we
assembled a group of international experts with distinct clinical and scientific backgrounds;
this group included physicians; specialists in critical care, anesthesiology, nephrology,
surgery, and emergency medicine; and basic scientists with expertise in biology and
physiology, who were recruited on the basis of their expertise in sepsis and organ
dysfunction. The group consisted of 23 international experts from 5 continents. A set of
questions was generated through mutual agreement and we sought evidence to answer each
question by searching the Cochrane Controlled Trials Register, the Cochrane Library,
MEDLINE, and EMBASE, from 1966 to present. The search terms for questions regarding
epithelial dysfunction are provided in Appendix 2. Finally, we reviewed the evidence with
the group and used the Delphi method to achieve consensus.
Ince et al.
Page 2
Shock
. Author manuscript; available in PMC 2017 January 31.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
RESULTS
On the basis of the literature review and consensus among the workgroup members, the
following key questions were considered:
1.
How does sepsis affect EC function and integrity?
2.
What different techniques are available to assess EC function at the bedside?
3.
What is the relationship between endothelial altered function and organ function?
4.
Impact of usual and microvascular targeted therapies?
5.
Is EC dysfunction adaptive or maladaptive?
The endothelial glycocalyx in sepsis
The ECL contains fenestrations and pores, which are heterogeneous between organs and the
different vascular generations (3) (Fig. 1). The integrity of the ECL as a barrier and
transporter of solutes is determined largely by the endothelial cytoskeleton and the
glycocalyx, which are tightly regulated. The glycocalyx is a 0.2 to 0.5-μm thick gel-like
layer lining the luminal membrane of the ECL, thought to compromise some 20% of the
intravascular volume. It is a multicomponent layer consisting of proteoglycans (of which
50% to 90% is heparin sulfate) and glycoproteins, anchored to ECs by glycosaminoglycans
(4). Shedding of the glycocalyx occurs in the presence of oxidants, hyperglycemia,
cytokines, and bacterial endotoxins (5, 6), and is associated with many states of disease
including sepsis (Fig. 2). The glycocalyx mediates several key physiological processes such
as the vascular barrier function, hemostasis, leukocyte and platelet adhesion, the
transmission of shear stress to the endothelium (4), and anti-inflammatory and antioxidant
defenses. The main instigators for glycocalyx shedding are thought to be reactive oxygen
species (ROS) such as hydrogen peroxide, hydroxyl anions, and superoxide, but other
mediators include tumor necrosis factor-alpha (TNF-α) and heparanase (7, 8). Their action
results in an increase in shedding products of the endothelium and disruption of the barrier
function—an effect that can be reversed experimentally by treatment with antioxidant
enzymes such as catalase and superoxide dismutase (SOD) (9). Loss-of-barrier function
induced by glycocalyx shedding is associated with the formation of edema (6) and is a key
contributor to sepsis-induced organ failure. Shedding of the glycocalyx may also hamper the
ability to sense and transduce blood flood-induced sheer stress, resulting in the endothelial
release of nitric oxide (NO) or endothelin (ET), which regulates smooth muscle cell
contraction and constitutes the basis of a process referred to as myogenic control of vascular
regulation (4). Increased plasma concentrations of NO and ET metabolites have been
reported in endotoxic shock (10). In addition, loss of sheer stress flow monitoring can alter
further regional vascular control because such signals are communicated to proximal
vascular structures by interendothelial communication (11) through gap junctions, resulting
in upstream vasogenic control. Finally, neutrophil extracellular trap (NET)—a mechanism
implicated in host defense against infection—can also contribute to endothelial damage and
impair microvascular perfusion (12).
Ince et al.
Page 3
Shock
. Author manuscript; available in PMC 2017 January 31.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Hemostasis and the endothelium
Sepsis is not only a state of systemic inflammation, it is also a state of deregulated
hemostasis. Hemostasis is a complex system mediated by the endothelium, soluble plasma
molecules, platelets, and leukocytes, which not only regulates the balance between pro and
anticoagulant forces, but it also directs platelet and fibrin clotting to areas of focal vascular
injury. The endothelium synthesizes and expresses molecules that are vital in regulating
hemostasis, such as von Willebrand factor (VWF), tissue factor (TF), and plasminogen
activator inhibitor type 1 (PAI-1). VWF—the largest multimeric glycoprotein in human
plasma (molecular masses from 500 to 20,000 kDa)—mediates initial platelet adhesion to
the damaged vessel wall by bridging platelet receptor platelet glycoprotein GPIB-IX
COMPLEX (GPIb-IX) to exposed subendothelial collagen. VWF is secreted by either the
constitutive pathway of lower molecular mass dimers, or the inducible pathway
[inflammatory stimulation: TNF-α, interleukin (IL)-6, IL-8] of the larger and ultralarge
multimers. The ultralarge multimers (ULVWF) are highly thrombotic and so they are rapidly
cleaved to less active forms by a disintegrin and metalloproteinase with a thrombospondin
type 1 motif, member 13 (ADAMTS-13) (or VWF-cleaving protease) as they are released
into the plasma. Thus, during normal physiology, plasma VWF binds and aggregates
platelets only in the presence of modulators such as ristocetin or under conditions of high
shear stress. Recently, a previously unrecognized role for GPIb-IX was reported, suggesting
that platelets may actually exert an anti-inflammatory action in some models of experimental
sepsis.
In normal hemostasis, coagulation and fibrinolysis are tightly regulated and kept in balance
by the endothelium, such that they allow blood to flow freely without systemic bleeding or
clotting. TF is a procoagulant transmembrane glycoprotein synthesized by the endothelium
and leukocytes (13), which, by creating complexes with factor VIIa activates factors IX and
X, which ultimately leads to clot formation. The endothelium regulates TF by producing TF
pathway inhibitor (TFPI), which limits fibrin deposition by binding to factor Xa, and inhibits
TF-factor VIIa complex. In addition, the endothelium further regulates anticoagulation by
activating protein C via thrombomodulin and endothelial protein C receptor, which inhibits
factor V, factor VIII, and PAI-1. PAI-1—another glycoprotein synthesized by the
endothelium and the liver—regulates fibrinolysis by inhibiting tissue plasminogen activator
(tPA) in health, but is incrementally released during inflammation.
Sustained inflammation during severe sepsis drives hemostasis toward a prothrombotic and
antifibrinolytic state, which can lead to disseminated microvascular thrombosis, organ
ischemia, and multiple organ dysfunction syndrome (MODS). Clinically, this phenomenon
can manifest as one of the following phenotypes—disseminated intravascular coagulation
(DIC), thrombotic thrombocytopenic purpura/hemolytic uremic syndrome (TTP/HUS), or
thrombocytopenia-associated multiple organ failure (TAMOF) (13, 14). Inflammatory
mediators during sepsis such as IL-6, plasma-free hemoglobin, VWF proteolytic fragments,
shiga toxin, and neutralizing autoantibodies can inactivate ADAMTS-13 (15, 16), the
proteolytic enzyme in charge of cleaving ULVWF into smaller less thrombogenic multimers.
In addition, plasmin, thrombin, products of activated coagulation, and granulocyte elastase
released by activated neutrophils can proteolyze ADAMTS-13 into inactive fragments, and
Ince et al.
Page 4
Shock
. Author manuscript; available in PMC 2017 January 31.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
neutrophil-derived ROS can inhibit ADAMTS-13–mediated cleavage, which leads to an
acquired ADAMTS-13 deficiency, and thus increased risk for disseminated platelet/VWF-
rich microvascular thrombosis (14, 17). Furthermore, the normal anticoagulant system
regulated by TFPI and protein C is defective in sepsis because TFPI is decreased due to
reduced synthesis and proteolytic inactivation, and protein C activation is dysfunctional. In
addition, the fibrinolytic pathway is suppressed in sepsis by the increased release of PAI-1
by the endothelium. These imbalances can ultimately lead to the dissemination of fibrin-rich
microvascular thrombi as observed in DIC, which occurs in 25% to 50% of septic patients.
Drugs have been tested in attempts to address the hemostatic imbalances associated with
sepsis. For example, recombinant human (rh) TFPI, rh-activated protein C, rh-soluble
thrombomodulin, protein C concentrate, heparin, antithrombin III, and platelet-activating
factor antagonist have all been unsuccessful in large randomized control trials as
monotherapy for sepsis (35). Because the hemostatic system is so complex with many
molecules being altered during sepsis, blood purification offers a chance to achieve
hemostatic balance by removing molecules that are causing harm and replenishing deficient
soluble plasma factors. Therapeutic plasma exchange has been tried with various successes
in patients with sepsis-induced disseminated microvascular thromboses (DIC, TTP/HUS,
and TAMOF) (18). A future approach of employing combination therapy would more likely
achieve success once clinicians have appropriate tools/biomarkers to fully assess the
underlying pathogenic mechanism of disseminated microvascular thrombosis.
The inflamed endothelium
Neutrophil/monocyte–EC interaction plays an important role in the pathogenesis of sepsis,
leading to organ failure. This interaction is mediated by adhesion molecules to which
leukocytes anchor themselves, allowing them to eventually extravagate into the tissues cells
—a process referred to as diapedesis. There, they can release inflammatory mediators and
reactive molecules to destroy pathogens, but at the same potentially causing tissue damage.
The integrity of the glycocalyx is of prime importance in this process. The adhesion
molecules responsible for leukocyte adhesion leading to extravasation in conditions of health
are embedded in the glycocalyx, shielding them from adherence to the leukocytes (Fig. 1).
In conditions of inflammation and sepsis, cytokines and reactive species induce glycaclyx
shedding, exposing the adhesion molecules initiating leukocyte adhesion, leading to
transmigration to the tissues (Fig. 2).
Of the inflammatory mediators, oxidative and nitrosative stress (ROS, reactive nitrogen
species [RNS]) resulting from oxidant release by mitochondria, xanthine oxidase,
nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) oxidase, and by
uncoupled endothelial NO synthase (eNOS) damages the ECL glycocalyx and alters
endothelial function. Shedding of the glycocalyx exposes the otherwise hidden adhesion
molecules to circulating leukocytes, which in turn facilitates adhesion and ultimately
transmigration through the ECL and into the parenchyma, in addition to altering NO and ET
production and contributing to the loss of vascular reactivity. Selectins (E, L, and P) mediate
sticking and rolling, whereas integrins such as intercellular adhesion molecule 1 (ICAM-1)
and vascular cell adhesion molecule 1 (VCAM-1) mediate firm adhesion and transcellular
Ince et al.
Page 5
Shock
. Author manuscript; available in PMC 2017 January 31.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript