Angiopoietin-1 and -2 in Infectious Diseases Associated with Endothelial Cell Dysfunction

by

Andrea Vaughn Page, B.Sc(H), MD, FRCPC

A thesis submitted in conformity with the requirements for the degree of Master of Science

Institute of Medical Science University of Toronto

© Copyright by Andrea Vaughn Page 2012

ii

Angiopoietin-1 and -2 in Infectious Diseases Associated with

Endothelial Cell Dysfunction

Andrea Vaughn Page, B.Sc(H), MD, FRCPC

Master of Science

Institute of Medical Science

University of Toronto

2012

Abstract

Normal endothelial cell function is controlled in part by a tightly regulated balance between

angiopoietin-1 and -2 (Ang-1 and Ang-2). Angiopoietin dysregulation (decreased Ang-1 and

increased Ang-2) leads to an activated endothelium that is contractile, adhesive, and

prothrombotic. Since an activated endothelial phenotype is seen in invasive group A

streptococcal infection, E. coli O157:H7-induced hemolytic-uremic syndrome (HUS), and sepsis,

we hypothesized that angiopoietin dysregulation might also be present in these syndromes, and

to that end, measured angiopoietin levels in several well-characterized patient cohorts. Decreased

Ang-1 and/or increased Ang-2 were found in all three syndromes, and were predictive of clinical

outcome in HUS and sepsis. The prognostic utility of Ang-2 in sepsis was further enhanced by

combination with biomarkers of inflammation. Angiopoietin dysregulation may therefore

represent a shared final common pathway to endothelial activation as well as a clinically useful

prognostic biomarker in streptococcal toxic shock, HUS, and sepsis.

iii

Acknowledgments

The author would like to acknowledge Dr. W. Conrad Liles (supervisor) for his guidance and

mentorship, Drs. Kevin C. Kain, James W. Scholey, and Susan E. Quaggin (members of the

Program Advisory Committee) for their advice and direction, Drs. Malak Kotb, Allison McGeer,

Donald E. Low, Phillip I. Tarr, Shevin T. Jacob, Christopher C. Moore, and W. Michael Scheld

for their collaboration and for their provision of banked patient specimens, Ms. Nimerta Rajwans

for her technical expertise and her generous and unfailing support throughout the work, and Dr.

Laura Erdman, Dr. Andrea Conroy, and the other members of the Kain-Liles laboratory for their

encouragement. This work was funded in part by the University of Toronto Department of

Medicine Clinician-Scientist Training Program (A.V.P.), the McLaughlin Centre for Molecular

Medicine (W.C.L.), the Canadian Institutes of Health Research (CIHR) Canada Research Chair

program (W.C.L., K.C.K.), CIHR Team Grant in Malaria (K.C.K.), and CIHR MOP-13721

(K.C.K.), Genome Canada through the Ontario Genomics Institute (K.C.K.), and the Defense

Advanced Research Projects Agency (K.C.K.).

iv

Table of Contents

Acknowledgments .......................................................................................................................... iii

Table of Contents ........................................................................................................................... iv

List of Publications 2008-2011 ..................................................................................................... vii

List of Abbreviations ................................................................................................................... viii

List of Tables ............................................................................................................................... xiv

List of Figures ............................................................................................................................... xv

Chapter 1 Literature Review ......................................................................................................... 1

1.1 The Vascular Endothelium ................................................................................................. 1

1.1.1 Streptococcal Toxic Shock Syndrome .................................................................... 3

1.1.2 Hemolytic-Uremic Syndrome ................................................................................. 3

1.1.3 Sepsis ...................................................................................................................... 4

1.2 Angiogenesis ....................................................................................................................... 6

1.3 Molecular Structure and Function of the Angiopoietins and their Receptor(s) .................. 7

1.3.1 Regulation of Angiopoietin Expression and Release .............................................. 7

1.3.2 Angiopoietin Receptor(s) and Signalling ................................................................ 8

1.3.3 Downstream effects of Angiopoietin-1 and -2 ...................................................... 10

1.4 Angiopoietin-1 and -2 in Development ............................................................................ 13

1.5 Angiopoietin-1 and -2 in Mature Endothelial Cells .......................................................... 15

1.6 Angiopoietin-3 and -4, and Angiopoietin-like proteins .................................................... 16

1.7 Dysregulation of Angiopoietin-1 and -2 in Human Disease ............................................. 17

1.7.1 Angiopoietins in Critical Illness ........................................................................... 17

1.7.2 Angiopoietins in Malaria ...................................................................................... 18

1.7.3 Angiopoietins in Cardiovascular Disease ............................................................. 19

1.7.4 Angiopoietins in Chronic Kidney Disease ............................................................ 20

v

1.7.5 Angiopoietins in Pregnancy and Preeclampsia ..................................................... 21

1.7.6 Angiopoietins in Malignancy ................................................................................ 22

1.7.7 Angiopoietins in Respiratory Diseases ................................................................. 23

1.7.8 Angiopoietins in Autoimmune Connective Tissue Diseases ................................ 24

1.7.9 Angiopoietins in Miscellaneous Diseases and Syndromes ................................... 25

1.8 Angiopoietin-based Therapies .......................................................................................... 26

1.9 Biomarkers of Endothelial Dysfunction in Sepsis ............................................................ 27

1.9.1 Procalcitonin ......................................................................................................... 27

1.9.2 C-reactive protein .................................................................................................. 36

1.9.3 sTREM-1 ............................................................................................................... 36

1.9.4 Chi3L1 (YKL-40) ................................................................................................. 37

1.9.5 IP-10 ...................................................................................................................... 38

1.9.6 PF4 ........................................................................................................................ 38

1.9.7 vWF ....................................................................................................................... 39

1.9.8 sICAM-1 ............................................................................................................... 39

1.9.9 VEGF .................................................................................................................... 40

1.9.10 sFlt-1 ..................................................................................................................... 40

1.9.11 Miscellaneous Markers of Endothelial Dysfunction in Sepsis ............................. 41

1.9.12 Conclusion ............................................................................................................ 42

Chapter 2 Research Aims and Hypotheses .................................................................................. 43

2.1 Research Aims .................................................................................................................. 43

2.2 Hypotheses ........................................................................................................................ 43

Chapter 3 Angiopoietin-1 and -2 in Invasive Group A Streptococcal Infection ......................... 44

3.1 Introduction ....................................................................................................................... 44

3.2 Methods ............................................................................................................................. 44

3.3 Results ............................................................................................................................... 46

vi

3.4 Discussion/Conclusion ...................................................................................................... 50

Chapter 4 Angiopoietin-1 and -2 in the Hemolytic-Uremic Syndrome ...................................... 53

4.1 Introduction ....................................................................................................................... 53

4.2 Methods ............................................................................................................................. 53

4.3 Results ............................................................................................................................... 55

4.4 Discussion/Conclusion ...................................................................................................... 59

Chapter 5 Prognostic Biomarkers for Sepsis in a Predominantly HIV-infected Population in

a Resource-limited Setting ....................................................................................................... 62

5.1 Introduction ....................................................................................................................... 62

5.2 Methods ............................................................................................................................. 62

5.3 Results ............................................................................................................................... 65

5.4 Discussion/Conclusion ...................................................................................................... 79

Chapter 6 Discussion ................................................................................................................... 84

Chapter 7 Conclusion .................................................................................................................. 89

Chapter 8 Future Directions ........................................................................................................ 93

8.1 Effect of Ventilatory Strategy on Angiopoietin levels in ARDS ...................................... 94

8.2 Angiopoietins in HIV ........................................................................................................ 96

8.3 Angiopoietins in Obstructive Sleep Apnea ....................................................................... 97

8.4 Procalcitonin in patients with Hematologic Malignancy .................................................. 98

8.5 Systematic review of procalcitonin for the prediction of mortality in sepsis and severe

bacterial infection ............................................................................................................ 100

References ................................................................................................................................... 101

vii

List of Publications 2008-2011

1. Petruzziello TN, Yuen DA, Page AV, et al. Functional contribution of the

CXCR4/CXCR7/SDF-1 pathway to E. coli O157:H7 pathogenesis. J Clin Invest. 2011; In

press.

2. Page AV, Liles WC. Colony-stimulating factors (CSFs) in the prevention and management of

infectious diseases. Infect Dis Clin North Am. 2011; 25: 803-817.

3. Page AV, Kotb M, McGeer A, et al. Systemic dysregulation of angiopoietin-1/2 in

streptococcal toxic shock syndrome. Clin Infect Dis. 2011; 52: e157-e161.

4. Makhani N, Morris S, Page AV, et al. A Twist on Lyme: The challenge of diagnosing

European Lyme Neuroborelliosis. J Clin Micro. 2011; 49: 455-457.

5. Boggild AK, Page AV, Keystone JS, et al. Delay in diagnosis: malaria in a returning

traveller. CMAJ. 2009; 180: 1129-1131.

6. Page AV, Liles WC. Immunomodulators (Chapter 42). Principles and Practices of Infectious

Diseases. Ed. GL Mandell et al. 7th

ed. 2009.

7. Page AV, Liles WC. Granulocyte Colony-Stimulating Factor (G-CSF), Granulocyte-

Macrophage Colony-Stimulating Factor (GM-CSF), and other immunomodulatory therapies

for the treatment of infectious diseases in solid organ transplant recipients. Curr Opin Organ

Transplant. 2008; 13: 575-580.

viii

List of Abbreviations

ABC Abacavir

ACCP American College of Chest Physicians

ACS Acute Coronary Syndrome

ALI Acute Lung Injury

AMI Acute Myocardial Infarction

Ang-1 Angiopoietin-1

Ang-2 Angiopoietin-2

Ang-3 Angiopoietin-3

Ang-4 Angiopoietin-4

APACHE Acute Physiology and Chronic Health Evaluation

ARDS Acute Respiratory Distress Syndrome

ART Antiretroviral Therapy

AUC Area under the Curve

CAP Community-acquired Pneumonia

CaRT Classification and Regression Tree

CHF Congestive Heart Failure

Chi3L1 Chitinase-3-like protein-1

CI Confidence Interval

CKD Chronic Kidney Disease

ix

CLL Chronic Lymphocytic Leukemia

COPD Chronic Obstructive Pulmonary Disease

CPAP Continuous Positive Airway Pressure

CRP C-reactive Protein

DIC Disseminated Intravascular Coagulation

Dll4 Delta-like ligand 4

DMARDs Disease-modifying Anti-rheumatic Drugs

ED Emergency Department

EGF Epidermal Growth Factor

EGR-1 Early Growth Response-1

ELISA Enzyme-linked Immunosorbent Assay

ESR Erythrocyte Sedimentation Rate

FDP Fibrinogen Degradation Product

FGF-2 Fibroblast Growth Factor-2

GAS Group A Streptococcus (Streptococcus pyogenes)

Gb3 Globotriaosylceramide

HAP Hospital-acquired Pneumonia

Hb Hemoglobin

HbA1c Hemoglobin A1c

HFO High Frequency Oscillation

x

HIF-2α Hypoxia-inducible Factor-2α

HIV Human Immunodeficiency Virus

HMG-CoA 3-hydroxy-3-methyl-glutaryl-CoA (reductase inhibitor)

HUS Hemolytic-uremic Syndrome

ICAM-1 Intercellular Adhesion Molecule-1

ICH Intracranial Hemorrhage

ICU Intensive Care Unit

IGRA Interferon-γ Release Assay

IL Interleukin

IP-10 Interferon-γ-inducible Protein - 10 kDa

IQR Interquartile Range

LPS Lipopolysaccharide

MAPK Mitogen-Activated Protein Kinase

MCP-1 Monocyte Chemoattractant Protein-1

MEDS Mortality in Emergency Department Sepsis (score)

MHC Major Histocompatibility Complex

MMP Matrix Metalloproteinase

mRNA messenger Ribonucleic Acid

NF-κB Nuclear Factor-κB

NIHSS National Institutes of Health Stroke Scale

xi

NO Nitric Oxide

NS Non-significant

OR Odds Ratio

OSA Obstructive Sleep Apnea

PaCO2 Partial pressure of Carbon Dioxide (arterial)

PAH Pulmonary Arterial Hypertension

PAI-1 Plasminogen Activator Inhibitor-1

PaO2 Partial pressure of Oxygen (arterial)

PCR Polymerase Chain Reaction

PCT Procalcitonin

PF4 Platelet Factor 4

PI3k Phosphatidylinositol-3-kinase

POCT Point-of-care Test

PROWESS recombinant human activated Protein C Worldwide Evaluation in Severe Sepsis

PSI Pneumonia Severity Index

RA Rheumatoid Arthritis

ROC Receiver Operating Characteristic (curve)

RR Relative Risk

S1P Sphingosine-1-phosphate

SAPS Simplified Acute Physiology Score

xii

SCCM Society of Critical Care Medicine

sFlk-1 soluble Fms-like tyrosine kinase receptor-1 (VEGFR2 - Vascular Endothelial

Growth Factor Receptor 2)

sFlt-1 soluble Fms-like tyrosine kinase receptor-1 (VEGFR1 - Vascular Endothelial

Growth Factor Receptor 1)

Shh Sonic hedgehog

SIRS Systemic Inflammatory Response Syndrome

siRNA small interfering Ribonucleic Acid

SLE Systemic Lupus Erythematosus

SN Sensitivity

SOFA Sequential Organ Failure Assessment

SP Specificity

STEC Shiga Toxin-producing Escherichia coli (E. coli)

sTie-2 soluble Tyrosine kinase with immunoglobulin and epidermal growth factor

homology domain-2

sTREM-1 soluble Triggering Receptor Expressed on Myeloid cells-1

STSS Streptococcal Toxic Shock Syndrome

TB Tuberculosis

Th1 T helper cell, subset 1

Tie-1 Tyrosine kinase with immunoglobulin and epidermal growth factor homology

domain-1

xiii

Tie-2 Tyrosine kinase with immunoglobulin and epidermal growth factor homology

domain-2

TLR4 Toll-like Receptor 4

TNF Tumour Necrosis Factor

tPA tissue Plasminogen Activator

UC Ulcerative Colitis

VAP Ventilator-associated Pneumonia

VCAM-1 Vascular Cell Adhesion Molecule-1

VE-cadherin Vascular Endothelial-cadherin

VEGF Vascular Endothelial Growth Factor

VILI Ventilator-induced Lung Injury

vWF von Willebrand Factor

WPB Weibel-Palade Body

xiv

List of Tables

Table 1. Selected studies evaluating the use of procalcitonin to predict mortality in 30

adult patients with sepsis or severe bacterial infection

Table 2. Significant predictors of mycobacterial infection in Ugandan patients with 77

sepsis

xv

List of Figures

Figure 1. Angiopoietin-1 and -2 (Ang-1 and Ang-2) in patients with invasive Group A 47

streptococcal disease with and without streptococcal toxic shock syndrome (STSS)

Figure 2. Angiopoietin-1 and -2 (Ang-1 and Ang-2) concentrations, and the ratio 49

between the two (Ang-2:Ang-1), in matched acute and convalescent plasma samples

from patients with STSS

Figure 3. Timeline of blood sampling and division into study groups of participants 54

at various stages of E. coli O157:H7 infection

Figure 4. Serum angiopoietin concentration in E. coli O157:H7 infection 56

Figure 5. Serum angiopoietin concentration in individual patients with 58

E. coli O157:H7 infection before and after the diagnosis of HUS

Figure 6. Plasma biomarkers associated with in-hospital mortality in Ugandan 67

patients with sepsis

Figure 7. Plasma biomarkers associated with 28-day mortality in Ugandan patients 70

with sepsis

Figure 8. Classification and Regression Tree (CaRT) analysis of biomarker 72

combinations to predict mortality following an episode of sepsis

Figure 9. Plasma biomarkers associated with positive aerobic blood cultures 74

Figure 10. CaRT analysis for the prediction of positive mycobacterial cultures 78

1

Chapter 1 Literature Review

1.1 The Vascular Endothelium

The vascular endothelium, once conceptualized as little more than the inert and passive lining of

blood vessels, is now known to be a complex organ responsible for interacting with, and

responding to, its environment.1 As reviewed by Aird, bloodborne mediators (including

cytokines, pro-inflammatory molecules, and growth factors), surrounding and supporting cells,

mechanical shear stress, and physico-chemical factors (including temperature and pH) can all act

on endothelial cells to influence their function. This influence is revealed in the endothelial cell

phenotype, typically dichotomized as either activated or resting/quiescent, although intermediary

states are also possible. Quiescent endothelial cells are anticoagulant, anti-adhesive, and

vasodilatory, while activated endothelial cells are pro-coagulant, adhesive, and contractile. These

properties are reflected in cell surface marker expression, secreted molecules, and the integrity of

the endothelial cell barrier.

As reviewed by Levi and van der Poll, as well as Schouten et al, resting endothelial cells express

thrombomodulin and the endothelial protein C receptor on their surface, both of which facilitate

the generation of the crucial anticoagulant molecule, activated protein C.2, 3

Downregulation of

thrombomodulin expression, as occurs in activated endothelial cells, leads to reduced protein C

generation and subsequently, increased thrombin generation. Likewise, tissue factor pathway

inhibitor, tissue-type plasminogen activator, heparan sulfates, prostacyclin, and antithrombin are

either expressed or released by the anticoagulant, resting endothelium, while tissue factor,

plasminogen-activator inhibitor (PAI)-1 and von Willebrand Factor (vWF) are upregulated in the

procoagulant, activated endothelium. Differential cell surface molecule expression between

quiescent and activated endothelial cells influences not only the relative balance between pro-

and anti- coagulant activity, but also the degree of adhesion of circulating hematologic cells. E-

selectin is expressed on activated endothelial cells, where, in combination with P-selectin, it

facilitates rolling of leukocytes along the endothelial layer as a prelude to leukocyte adhesion

(facilitated by the upregulation of ICAM-1 (intercellular adhesion molecule-1) and VCAM-1

(vascular cell adhesion molecule-1) to activated endothelial cells and subsequent transmigration

across the endothelial barrier to a site of injury or inflammation.4 It should be noted that although

2

this is the classical paradigm of endothelial activation and leukocyte response, there is some

organ-specific variability as described in a second review by Aird.4

In addition to cell surface marker expression, activation of the endothelium also influences its

secretory function. As reviewed by Lowenstein et al, endothelial cells contain Weibel-Palade

bodies (WPBs), intracellular structures that store a variety of molecules for rapid release or

externalization upon exposure of the cell to inflammatory or noxious stimuli, including

mechanical factors (hypoxia, trauma, or shear stress), lipids (sphingosine-1-phosphate, ceramide,

and oxidized low-density lipoprotein), inflammatory mediators (thrombin, fibrin, terminal

complement components, and leukotrienes), and angiogenic factors (VEGF).5 Nitric oxide (NO)

impairs or inhibits WPB release.6 WPB exocytosis occurs within minutes, in contrast to the

transcriptional response which requires hours, and liberates the following: vWF, P-selectin,

tissue plasminogen activator, Interleukin (IL)-8, endothelin-1, and angiopoietin-2 (Ang-2),

among others. Consistent with the opposing actions of certain of these molecules, differential

regulation of WPB contents and exocytosis is known to exist.7 Nonetheless, the majority of the

molecules contained in WPBs are prothrombotic, inflammatory, or vasoconstrictive,

characteristics associated with an activated endothelium.

Finally, endothelial cell activation may manifest as abnormal barrier function. The resting

endothelium relies on stable endothelial cell morphology and intact inter-endothelial cell

adherens junctions (comprised largely of the protein vascular endothelial (VE)-cadherin) and

tight junctions (composed of the occludin, claudin, and junctional adhesion molecule proteins) to

limit vascular leak.8 Junctional protein stability is maintained in part by close association with

actin bundles. Under the various conditions which activate the endothelium (described in the

preceding paragraphs), junctional proteins may be downregulated, mis-located, or

phosphorylated, all leading to loss of junctional integrity. Furthermore, many vaso-active agents

activate the Rho GTPases (RhoA or Rac1), which in turn activate Rho kinase, leading to myosin

light chain phosphorylation, stress fibre (as opposed to the usual actin fibre) formation, and

increased cell contractility. Ultimately, this sequence of events produces inter-endothelial cell

gaps, and consequently, vascular leak.9

Under normal circumstances, all of the above responses are appropriately adaptive in the setting

of tissue inflammation or injury. However, these responses can be detrimental when they are

3

exaggerated or excessive, as is the case in certain infectious disease syndromes associated with

prominent endothelial cell activation.

1.1.1 Streptococcal Toxic Shock Syndrome

According to the Working Group on Severe Streptococcal Infections, Group A Streptococcal

(GAS; Streptococcus pyogenes) toxic shock (STSS) is defined by the isolation of Group A

streptococcus from a normally sterile site and hypotension (systolic blood pressure ≤ 90 mmHg),

as well as at least 2 of the following 6 signs of major organ dysfunction: acute kidney injury,

coagulopathy, necrotizing fasciitis, erythematous rash that may desquamate, elevated liver

enzymes, or acute respiratory distress syndrome (ARDS).10

STSS results largely from the action

of a bacterial-derived exotoxin that functions as a superantigen to activate a polyclonal subset of

T cells far in excess of that recruited during a normal immune response.11

Massive pro-

inflammatory cytokine release follows. Ultimately, endothelial activation leads to profound

hypotension and vascular leak, however, the mechanism by which this occurs has not yet been

fully explored.12

Antibiotics are administered in STSS but have little effect on pre-formed toxin. Therefore,

patients require aggressive supportive care, commonly in the intensive care unit (ICU).13

In

addition, mortality from STSS remains high, reported at 43% in the most recent epidemiologic

study.14

As a result, detailed exploration of all potential mediators of endothelial activation in

STSS will be crucial to the discovery of the new therapeutic targets necessary to improve patient

outcome.

1.1.2 Hemolytic-Uremic Syndrome

Shiga (or vero-) toxin-producing Escherichia coli (STEC) are a potentially serious cause of

foodborne infectious diarrhea and were responsible for a recent and well-publicized European

epidemic centered in Germany in 2011. During that outbreak, at least 4075 patients were

infected, and 908 subsequently developed the hemolytic uremic syndrome (HUS).15

Thirty-four

of those patients died.

HUS is defined by the triad of non-immune hemolytic anemia, thrombocytopenia, and acute

renal failure. Clinical signs and symptoms result from circulating bacterial-derived Shiga toxin,

since STEC infections are rarely associated with frank bacteremia.16

Shiga toxin binds to its

4

receptor, globotriaosylceramide (Gb3), which is concentrated on renal glomerular endothelial

cells, and induces the characteristic pathophysiologic changes associated with HUS, namely

endothelial cell swelling and detachment from the basement membrane.17

Endothelial

dysfunction, characterized by increased procoagulant activity, vWF release, and leukocyte

adhesion, is present even earlier in the disease process, before the clinical signs and symptoms of

HUS are manifest.18-20

Whether this occurs directly as a result of toxin interaction with the

endothelial cells or secondarily as a result of elevated pro-inflammatory cytokines is not clear, as

both mechanisms have been documented.21-23

Furthermore, although the inciting factors (Shiga

toxin and/or cytokines) and sequelae (procoagulant and adhesive endothelium) are well-

described, the intervening mediators of endothelial dysfunction have not been determined.

1.1.3 Sepsis

Prior to 1992, proposed definitions of sepsis were often vague, contradictory, and fluid. With the

publication of consensus definitions by the American College of Chest Physicians (ACCP) and

Society of Critical Care Medicine (SCCM), early patient diagnosis and treatment, as well as

clinical trial enrolment or inclusion criteria, could be standardized.24

The Systemic Inflammatory

Response Syndrome (SIRS) was defined as the existence of two or more of: temperature >38oC

or < 36oC, heart rate > 90 beats per minutes, respiratory rate > 20 breaths per minute or PaCO2

(partial pressure of carbon dioxide) < 32 mmHg, and leukocyte count > 12 x 109 cells/L, < 4000

x 109 cells/L, or > 10% immature (band) forms. SIRS may result from a variety of both

infectious and non-infectious conditions, with pancreatitis, ischemia, trauma, and burns being

among the more common causes in the latter group. Sepsis was then defined as SIRS occurring

as a result of infection. As elegantly illustrated in a Venn diagram from the original manuscript,

not all patients with infection have SIRS, and not all patients with SIRS have infection, however,

all patients with sepsis have both (suspected infection and SIRS). Bacteremia was recognized as

the most common, but not sole, cause of sepsis. Severe sepsis was defined as sepsis associated

with organ dysfunction, hypoperfusion, or hypotension, while septic shock was defined as

sepsis-induced hypotension, hypoperfusion, or need for inotropic or vasopressor support despite

adequate fluid resuscitation. Upon re-evaluation in 2001, the basic concepts behind each

definition were maintained, and more specific guidance provided regarding vital sign,

inflammatory, hemodynamic, organ dysfunction, and perfusion parameters that might be used to

5

make the diagnosis of sepsis.25

Despite concerns about specificity, the original definitions of

SIRS and sepsis remain in common use.

In 2009, the Canadian Institute for Health Information released updated statistics on the

epidemiology of sepsis in Canada (excluding Quebec).26

In 2008-2009, there were more than

30 000 sepsis-related hospitalizations in Canada, an increase of more than 4000 in 4 years, with

an associated mortality rate of 30.5%. Severe sepsis accounted for 39% of those admissions, and

45% of the deaths. Patients with sepsis remained in hospital for a median of 9 days longer than

did patients hospitalized for other indications, and 45% of patients with sepsis required ICU care,

further adding to the healthcare cost. This problem is mirrored in other high-income countries in

which sepsis remains a leading cause of healthcare expenditure and patient mortality.27-30

As recently reviewed by Lee and Liles, the clinical manifestations of sepsis ultimately derive

from diffuse and profound endothelial activation, with all of its associated characteristics:

hypercoagulability, inflammation, and vascular leak.31

In a mouse model of sepsis, disruption of

endothelial-specific NF-κB reduced thrombin-antithrombin complex formation (a marker of

coagulation), diminished neutrophil infiltration, attenuated pulmonary edema and major organ

dysfunction, and reversed hypotension, all without enhancing bacterial clearance, suggesting that

the documented benefits were a result of improved endothelial function.32

In vivo studies of

sepsis have documented impaired endothelium-dependent microvascular reactivity and elevated

levels of circulating markers of endothelial activation (including E-selectin, soluble ICAM-1,

soluble VCAM-1, and PAI-1), directly proportional to disease severity.33, 34

The stimuli for

endothelial activation in sepsis are bacterial products, such as lipopolysaccharide (LPS), detected

by pattern recognition molecules, such as toll-like receptor 4 (TLR4) on endothelial cells and

monocytes and macrophages of the innate immune system. Endothelial activation via an

LPS/TLR4 signalling pathway has been shown to augment neutrophil recruitment and

extravasation, endothelial contractility and vascular leak, and upregulation of procoagulant

genes, while activation of the innate immune system leads to production of pro-inflammatory

cytokines and subsequent activation of endothelial cells.35-38

Typically, such responses are

appropriate, adaptive, and localized, and the inciting infection is cleared. However, when these

responses are exaggerated, excessive, or prolonged, sepsis and critical illness result.

6

1.2 Angiogenesis

Angiogenesis, the development and maturation of new blood vessels or their repair following

injury, is a critical function of the healthy endothelium and is mediated by both positive and

negative regulatory factors. Given the complexity of the process (reviewed in detail by Carmeliet

and Jain) and the intent of this review to focus specifically on the angiopoietins, only a limited

discussion of angiogenesis will be included here.

The quiescent endothelium is maintained in a resting state by cell survival signals produced by

the pericytes that underlie and support the endothelial cell monolayer.39

These signals include

VEGF, which is anti-apoptotic via the phosphatidylinositol-3-kinase (PI3k)/Akt pathway, as well

as Dll4/Notch, and angiopoietin-1 (Ang-1). Ang-1 has been shown to upregulate Delta-like

ligand 4 (Dll4) in endothelial cells to facilitate basal signalling through the Notch receptor and

promote continued endothelial quiescence.40

In response to hypoxic (hypoxia-inducible factor

(HIF)-2α), inflammatory, or mechanical stimuli, angiopoietin-2 (Ang-2) is released from

endothelial cells, and mediates endothelial cell detachment from the basement membrane via

matrix metalloproteinases.41-43

Endothelial permeability is further increased by VEGF, which

allows plasma proteins to establish a subendothelial extracellular matrix. The sprouting

angiogenesis that follows requires the migration of Dll4-expressing “tip” endothelial cells in

response to a concentration gradient of VEGF on the extracellular matrix, followed by the

proliferation of Notch1-expressing “stalk” endothelial cells in response to high levels of the

Notch1 ligand Jagged 1. Akin to its function in resting endothelial cells, Dll4 on the tip cell is

thought to interact with Notch on the stalk cells to prevent concurrent sprouting or branching of

the stalk cells, thereby facilitating vessel growth in a single direction.44

Following migration, the

endothelial cells form tubules and return to quiescence to complete maturation of the new vessel,

a process guided in part by Ang-1.45

7

1.3 Molecular Structure and Function of the Angiopoietins and their Receptor(s)

Ang-1 and Ang-2 are glycoproteins with a similar size of 70 kDa and coiled-coil domains at the

amino termini that facilitate oligomerization.46

Ang-1 must, at minimum, form a tetramer in

order to activate its receptor, but is in fact usually found in superclusters of oligomers.47

Ang-2,

in contrast, typically exists in dimeric form in vivo. The carboxy terminus of both molecules

consists of a fibrinogen-related domain that regulates binding to their shared receptor, Tie-2.

Traditionally, Ang-1 has been considered an agonist, and Ang-2 an antagonist, at the Tie-2

receptor, inducing opposite effects on endothelial cell function.

1.3.1 Regulation of Angiopoietin Expression and Release

Ang-1 is produced primarily in the pericytes and smooth muscle cells that surround the

endothelial cell monolayer, but can also derive from platelets.48, 49

More recent evidence suggests

that Ang-1 is also expressed in neutrophils in vitro, but the in vivo significance of this finding is

not clear.50

Ang-1 production by pericytes is constitutive, but can be upregulated by Sonic

hedgehog (Shh) and downregulated by Fibroblast Growth Factor-2 (FGF-2).51, 52

In contrast,

FGF-2 upregulates Ang-2 expression in endothelial cells and VEGF expression in stromal cells,

illustrating the positive/negative regulatory factor interplay common in vascular physiology.

Ang-2 is produced by endothelial cells and stored in the Weibel-Palade bodies for rapid release

upon exposure to various physico-chemical, mechanical, and inflammatory stimuli (discussed in

Chapter 1.1). Of particular relevance to the angiogenic system, WPB exocytosis can be induced

by VEGF and sphingosine-1-phosphate (S1P).53, 54

S1P is a phospholipid signalling molecule

that is secreted from platelets, activated monocytes and mast cells, and vascular endothelial

cells,55

providing evidence of an additional degree of interaction between the inflammatory and

angiogenic pathways. The net effect of S1P on the vasculature may be context-dependent, as S1P

not only has a role in Ang-2 release via WPB exocytosis, but is also implicated as a mediator of

Ang-1 signalling. Ang-1 has been found to increase intra-endothelial cell concentrations of S1P

and to be at least partly dependent on S1P for its effects on vascular permeability.56

8

1.3.2 Angiopoietin Receptor(s) and Signalling

The primary angiopoietin receptor, Tie-2 (tyrosine kinase with immunoglobulin and epidermal

growth factor homology domain-2), as well as Tie-1 (tyrosine kinase with immunoglobulin and

epidermal growth factor homology domain-1), both belong to a family of vascular tyrosine

kinase receptors expressed primarily, but not exclusively, in endothelial cells.57

The extracellular

domain of Tie-2 contains three immunoglobulin-like domains and three epidermal growth factor

(EGF)-like domains.58

Binding of the Tie-2 agonist Ang-1 leads to phosphorylation of the receptor and downstream

signalling. Ang-2 binds at the same site, but its effects may be agonistic or antagonistic

depending on the microenvironment.59

In cultured endothelial cells, Ang-2, in the absence of

Ang-1, can act as a partial agonist at Tie-2, leading to phosphorylation of the receptor and to the

downstream effects characteristic of Ang-1/Tie-2 signalling.60

However, Yuan et al propose that

since Ang-1 binds the Tie-2 receptor with 20 times the affinity of Ang-2 (perhaps due to the

multimerization of Ang-1 which would support clustering and activation of multiple Tie-2

molecules on the cell surface), Ang-2 acts as a relative antagonist in the presence of Ang-1,

leading to relatively weaker signalling when it (Ang-2) binds to Tie-2. Consistent with this, Ang-

2 induces a lesser degree of signalling through Tie-2 than does Ang-1, with reduced downstream

phosphorylation of Akt. Interestingly, release kinetics prior to receptor internalization also differ

between the two molecules, with Ang-2 being released more rapidly than Ang-1.61

Taken

together, the above studies indicate that the degree of activation of Tie-2 is determined by the

relative balance between Ang-1 and Ang-2.

Like other members of the tyrosine kinase receptor family, Tie-2 exists on the cell surface in pre-

formed clusters of oligomers.62

Since tetrameric Ang-1 is the smallest oligomer that can activate

Tie-2, receptor grouping likely facilitates ligand binding. Tie-2 is located on both the apical and

basolateral membranes, thereby allowing it to be bound by circulating Ang-1/Ang-2, as well as

by Ang-1/Ang-2 secreted by pericytes and endothelial cells, respectively.63

Upon activation by

Ang-1, Tie-2 is rapidly internalized through clathrin-coated pits, ubiquitylated, and degraded,

likely to regulate the duration of signalling, as is common within the receptor family.62, 64

As

noted above, ligands appear to be released prior to internalization.61

In keeping with its largely

antagonist activity and its rapid rate of release after binding, Ang-2 is only a weak stimulus for

9

the internalization and degradation of Tie-2. Tie-2 is also regulated by constitutive and VEGF-

inducible ectodomain cleavage, leading to a soluble Tie-2 (sTie-2) fragment that can bind Ang-1

(and Ang-2) and inhibit signalling through intact, transmembrane Tie-2 receptors.65

The function of the alternate angiopoietin receptor, Tie-1, has only recently been determined.

Ang-1 can induce Tie-1 phosphorylation in a Tie-2-dependent manner, however, Ang-1-induced

activation of Tie-1 downregulates the Akt and 42/44 MAPK (mitogen-activated protein kinase)

pathways, resulting in inhibition of Tie-2-mediated effects.66

Tie-1 has been shown to form

heterodimers with Tie-2 on the endothelial cell surface, and experiments using inactive forms of

Tie-1 or Tie-2 tyrosine kinases suggest that it is Tie-2 tyrosine kinase activity that results in Tie-

1 phosphorylation.67

In keeping with their opposing actions, Ang-1 has been found to

preferentially bind non-Tie-1-associated Tie-2.68

In addition to acting as a regulator of Ang-

1/Tie-2 signalling, Tie-1 has also been proposed as a second mechanism by which Ang-2 may

exert its context-dependent agonist/antagonist effects. In the presence of Tie-1, Ang-2 functions

as an antagonist, binding but not signalling at Tie-2, and in the absence of Tie-1, Ang-2 functions

as an agonist, binding and signalling through Tie-2.69

Tie-1 also mediates, in part, the cross-talk

between the angiopoietins and VEGF. VEGF indirectly phosphorylates Tie-2 by facilitating

proteolytic cleavage of the extracellular domain of Tie-1, thereby enhancing access of Ang-1 to

the binding site of Tie-2.70, 71

In addition to VEGF, inflammatory stimuli and shear stress can

also rapidly induce cleavage of the Tie-1 ectodomain.72, 73

Taken together, these findings suggest

that endothelial function can be precisely regulated in response to changes in the

microenvironment, and that this regulation can occur by a variety of means, including changes in

receptor cell surface expression, cleavage of the Tie-1 extracellular domain, and the relative

agonist/antagonist concentrations.

The function of Tie-1 independent of Tie-2 remains controversial. Disruption of the Tie-1 gene is

embryonic lethal at day E13.5 due to edema, hemorrhage, and microvascular rupture.74

Prior to

death, endothelial cells in these animals demonstrate increased filopodia and resulting enhanced

capillary density, as well as abnormal lymphangiogenesis.75

Double Tie-1 and Tie-2 knockout

(-/-, null) mice display a phenotype more severe than that seen with knockout of either gene

alone.76

Beyond the developmental stage, however, the Tie-2-independent function of Tie-1 is

less clear. In a mouse model of atherosclerosis, Tie-1 was overexpressed at regions of shear

stress, while Tie-1 knockdown attenuated atherosclerotic changes.77

In vitro studies have

10

suggested that Tie-1 may have its own function in the endothelial response to inflammation.

Overexpression of Tie-1 in endothelial cells lead to auto-phosphorylation and subsequent

upregulation of cell surface adhesion molecules, including E-selectin, VCAM-1, and ICAM-1,

while knockdown of Tie-1 by siRNA (small interfering RNA) lead to downregulation of certain

pro-inflammatory molecules, including IL-1β.78, 79

However, these findings have yet to be

duplicated in higher order models. Furthermore, other authors have not detected an effect of

siRNA knockdown of Tie-1 on Ang-1 function.80

In summary, the net signalling that occurs through Tie-2 is regulated by the balance in local

concentrations of Ang-1 and Ang-2, as well as by the degree of co-localization and ectodomain

cleavage of Tie-1. Through the latter mechanism, VEGF and inflammatory stimuli can modulate

the magnitude of Ang-1 and Ang-2 signalling (or lack thereof) through Tie-2, and influence

endothelial cell function.

1.3.3 Downstream effects of Angiopoietin-1 and -2

Consistent with their classic roles as competitive antagonists of the same receptor, Ang-1 and

Ang-2 typically exert opposite effects on signalling through Tie-2. Via Tie-2, Ang-1 activates

both the Erk and Akt pathways. Erk is thought to be important for cell migration and

proliferation, while Akt modulates cell survival and inhibits apoptosis.81

Multimeric Ang-1

results in relocation of Tie-2 to areas of cell-to-cell contact, subsequently leading to Ang-1

bridging and trans-activation of Tie-2, preferentially activating Akt, which subsequently

activates endothelial Nitric Oxide Synthase (eNOS) and phosphorylates the forkhead

transcription factor Foxo1.82

This results in Foxo1 exclusion from the nucleus, and reduced

transcription of its dependent genes. The net result is vascular quiescence in the setting of cell-to-

cell contact. Sensing of neighbouring endothelial cells may be facilitated through Tie-2 receptors

clustered in lipid rafts that also contain VE-cadherin, and appears to hinge on the expression of

the transcription factor Krüppel-like factor 2.83, 84

In the absence of cell-to-cell contact, Ang-1

can be bound to fibronectin, collagen, vitronectin, and other components of the extracellular

matrix to anchor Tie-2 to the cell surface in contact with the substratum. In this situation,

signalling through Tie-2 results in preferential activation of the Erk pathway, resulting in

increased cell mobility and enhanced likelihood of forming a functional endothelial cell

monolayer.82

Also necessary for cell migration is Tie-2-dependent caveolin-1 polarization, and

11

Ang-1-induced Early Growth Response-1 (Egr-1).67

Egr-1 is a transcription factor that

upregulates multiple growth factors, cytokines, adhesion molecules, and pro-angiogenic factors.

In siRNA knockdown studies, loss of Egr-1 abolished Ang-1-induced endothelial cell

migration.85

Gene expression profiling of Ang-1-activated endothelial cells reveals upregulation

of genes involved in endothelial cell proliferation, differentiation, migration, and survival, and

downregulation of those involved in apoptosis and inhibition of transcription.86

Besides its anti-apoptotic effect, Ang-1 exerts an anti-permeability effect on the endothelium, at

least partly through increasing the depth and glycosaminoglycan content of the endothelial

glycocalyx.87

Ang-1 also activates Rac1 and phosphorylates RhoGAP, while at the same time

inhibiting RhoA, which is responsible for Ang-2 mediated cytoskeletal rearrangements and

vascular permeability (see below).88

Ang-1 also has anti-thrombotic properties: downregulation

of Ang-1 leads to enhanced VEGF- and tumour necrosis factor (TNF)- induced tissue factor

expression through loss of Ang-1 negative regulation.89, 90

Finally, Ang-1 is anti-inflammatory. It

has been shown to downregulate endothelial cell production of the pro-inflammatory molecule

endothelin-1, as well as LPS- and VEGF- induced cell surface expression of the leukocyte

adhesion markers E-selectin, ICAM-1, and VCAM-1.91-93

Consistent with its role as a Tie-2 antagonist, Ang-2 typically opposes the effects of Ang-1 and

therefore induces pro-apoptotic, pro-permeability, and pro-adhesive changes in the endothelium.

Blockade of signalling through Tie-2 leads to endothelial cell apoptosis via inhibition of the Akt

pathway, as well as increased expression of the antiangiogenic molecule thrombospondin.94

Ang-

2 increases thrombin-induced vascular permeability by impairing VE-cadherin organization and

increasing intercellular gap formation, possibly through actions on Rho kinase and myosin light

chain phosphorylation, or by inducing αvβ3 integrin internalization and degradation.9, 95, 96

Tie-2

knockdown mimics the effect of Ang-2 on endothelial barrier function. Treatment with

exogenous Ang-2 has been shown to result in endothelial detachment from the basement

membrane within 4 hours.97

Finally, in a transgenic mouse model, inducible endothelial cell-

specific Ang-2 expression enhanced myeloid cell recruitment, adhesion to ICAM-1 on the

endothelial cell surface, and tissue infiltration.98

12

A simplified summary of these effects casts Ang-1 as an effector of vascular quiescence, and

Ang-2 as an effector of vascular activation, although there are certainly context-dependent

nuances to this interpretation of the function of both molecules.

13

1.4 Angiopoietin-1 and -2 in Development

Disruption of the Ang-1 gene in mice is embryonic lethal at day E12.5.99

Prominent

developmental defects include a simplified endocardium detached from the myocardial wall, loss

of myocardial trabeculations (suggestive of diminished cross-talk between myocardial and

endocardial cells), and generalized dilatation and loss of complexity (diminished branching) of

vascular networks. Ultrastructural examination of these abnormal vessels reveals rounded

endothelial cells separated from the underlying supportive cells. Disruption of Tie-2 is also

embryonic lethal in mice, resulting in death by day E10.5.74

The Tie-2 null phenotype is one of

large, unbranched blood vessels, and a simplified vascular network, similar to that seen with

Ang-1 knockout.

Overexpression of Ang-2 in the vascular endothelium (transgenic mice created using Tie-2

transcriptional regulatory elements) is embryonic lethal at day E9.5-10.5.100

Pathology studied at

day E9 revealed a disorganized and discontinuous vascular network, and confirmed the basic

concept that Ang-2 antagonizes the effect of Ang-1; detachment of the endocardium from the

underlying myocardium with overexpression of Ang-2 was identical to that seen in Tie-2 and

Ang-1 knockout mice. Similarly, detachment of the endothelial layer and rounding of individual

endothelial cells was seen in the vessel walls and was equivalent in Ang-2 overexpression and

Ang-1 or Tie-2 underexpression.

In contrast to all of the above phenotypes, knockout of Ang-2 does not result in embryonic

lethality. Instead of being a key component of vascular development, Ang-2 appears to be crucial

for vascular remodelling. Using the retinal vasculature as a representative vascular bed, Ang-2 -/-

mice were found to have grossly normal vessels at birth.101

However, they failed to show

evidence of angiogenesis in response to hypoxia-induced VEGF, and similarly failed to

demonstrate the usual and necessary hyaloid vascular regression. Similarly, in a rat pup model of

angiogenesis, Ang-2 mediated new vessel sprouting in the presence of VEGF, and vessel

regression and endothelial apoptosis in its absence.102

That Ang-2 is involved in the coordinated

vascular sprouting and regression necessary to produce a mature vascular network is consistent

with its known actions as a context-dependent antagonist/agonist at Tie-2. Of note, Ang-2 -/-

mice typically die within two weeks of birth with profound chylous ascites ± pleural effusions.101

These mice were found to have a leaky and disorganized lymphatic network thought to result

14

from the lack of Ang-2 agonist activity at Tie-2 in the presence of VEGF. This is consistent with

more recent in vitro work that confirms a prominent agonist effect of Ang-2 on lymphatic

endothelial cells.103

Finally, Ang-2 -/- mice are unable to mount an appropriate inflammatory

response to either chemical or bacterial insult, partly as a result of the loss of Ang-2-induced

sensitization of endothelial cells to the effects of TNF, in particular to TNF-induced expression

of cell adhesion molecules.104

15

1.5 Angiopoietin-1 and -2 in Mature Endothelial Cells

In healthy adult human tissue, Ang-1 is widely expressed in most organs while Ang-2 mRNA is

detectable primarily in those anatomic sites with significant angiogenesis, namely the ovary

(which must create a densely vascularized corpus luteum after follicular rupture), uterus, and

placenta.100

It was early studies involving the rat ovary that characterized the influence of VEGF

on the vascular action of Ang-2: in the presence of VEGF, as is found in the evolving corpus

luteum, excess Ang-2 promoted vessel sprouting by blocking the stabilizing effect of Ang-1,

while in the absence of VEGF, as is found in follicular atresia, excess Ang-2 promoted vessel

regression.100

Reflecting its role in destabilizing the mature endothelium (whether physiologic, in

the promotion of angiogenesis, or pathologic, in the initiation of endothelial dysfunction),

administration of exogenous Ang-2 leads to severe vascular leak and pulmonary edema in mice.9

In contrast, transgenic mice that overexpress Ang-1 were found to have leak-resistant vessels,

even in the presence of inflammatory mediators that typically increase vascular permeability.105

Furthermore, while isolated VEGF overexpression enhances angiogenesis but produces

pathologically permeable vessels, co-expression of VEGF with Ang-1 leads to the development

of new, functional, appropriately permeable vessels. In addition, while adenoviral-delivered

VEGF lead to diffuse edema and death in mice, this was blocked by the pre-administration of

adenoviral-delivered Ang-1.106

Further parsing out the role of Ang-1 in the adult vasculature, a

murine Cre-Lox model of conditional Ang-1 gene knockout demonstrated no obvious phenotype

during vascular quiescence with Ang-1 gene disruption initiated after E13.5, but significantly

altered vascular remodelling in response to injury.107

Taken together, the studies of the angiopoietins in embryonic development and in the mature

endothelium indicate that Ang-1 signalling through the Tie-2 receptor is indispensible for the

formation of new blood vessels whether during embryogenesis or during repair of the adult

vasculature. Although Ang-1 plays a role in the maintenance of vascular quiescence, it is not an

absolute pre-requisite for the resting endothelial state. In contrast, Ang-2 is not absolutely

required during embryogenesis, but is necessary for vascular remodelling thereafter, during

which its function is context-dependent as per the presence or absence of VEGF.

16

1.6 Angiopoietin-3 and -4, and Angiopoietin-like proteins

While Ang-1 and Ang-2 are the best characterized, they are not the only members of the

angiopoietin family. Angiopoietin-3 and -4 (Ang-3 and Ang-4) are Tie-2 receptor agonists,

capable of inducing Tie-2 phosphorylation and subsequent angiogenesis.108

Both are species-

specific, with Ang-3 limited to mice and Ang-4 to humans, and represent mouse-human

orthologues based on equivalent chromosomal loci.109

The functional role of Ang-4 remains

unclear: some in vitro studies have found that Ang-4 inhibits endothelial cell migration and

supports endothelial barrier function,110

while others have described increased angiogenesis and

tumour progression in glioblastoma multiforme,111

suggesting that the effect of Ang-4 may be

context-dependent.

The physiologic role of the angiopoietin-like proteins is currently under active study.

Angiopoietin-like proteins 1-7 all have a coiled-coil domain at the N-terminus and a fibrinogen-

like domain at the C-terminus, akin to the angiopoietins, but none bind Tie-2 or Tie-1, and most

appear to have important in vivo functions unrelated to angiogenesis.112

As an example of the

pleiotropic effects of this family of molecules, Angiopoietin-like protein-4 participates both in

increased endothelial permeability through direct interactions with integrin α5β1 (and subsequent

Rac signalling), VE-cadherin, and claudin-5, and in hypertriglyceridemia in response to

glucocorticoids.113, 114

Angiopoietin-like protein-2 is highly expressed in adipose tissue, increases

in response to chronic hypoxia and inflammation, and is correlated with insulin resistance in

obese patients.115

Angiopoietin-like protein-3 increases plasma levels of cholesterol and

triglycerides, and mutations in its gene have recently been implicated in familial combined

hypolipidemia.116

Angiopoietin-like protein-6 has also been implicated in obesity, although the

literature on the subject is conflicting.117

17

1.7 Dysregulation of Angiopoietin-1 and -2 in Human Disease

Because of the central role of angiogenesis in the development, proper function, defence, and

subsequent repair of every organ system, it is not surprising that the angiopoietins, important

regulators of angiogenesis, have been widely studied and implicated in a variety of human

diseases and syndromes. Although the occasional study has attempted to explore angiopoietin

levels in parenchymal tissue or non-hematogenous bodily fluids, the vast majority of the studies

presented below have examined serum, plasma, or whole blood concentrations of Ang-1 and

Ang-2, in keeping with their vascular origins and targets. Because of the breadth and depth of the

literature in this field, the following review will focus primarily on clinical studies, with

reference to in vitro or animal models where required to provide supporting evidence or

rationale. It should be noted that, in healthy individuals, the serum/plasma concentration of Ang-

1 is greater than that of Ang-2, and that angiopoietin dysregulation refers to a perturbation of this

typical ratio (decreased Ang-1, increased Ang-2, and an increased Ang-2:Ang-1 ratio).118

1.7.1 Angiopoietins in Critical Illness

The majority of work in this area has focused on the association of the angiopoietins with sepsis

and its diagnosis, treatment, complications, and outcome. Angiopoietin dysregulation, and in

particular a net excess of Ang-2, would be predicted to create a pro-inflammatory, pro-coagulant,

and pro-adhesive endothelium that would in turn manifest as tissue inflammation, microvascular

thrombosis, and hypotension, all frequently observed in septic patients. In healthy volunteers

who received an infusion of LPS, serum Ang-2 levels began to rise 2 hours after the infusion and

peaked at 4.5 hours, while neither sTie-2 nor Ang-1 was appreciably different from baseline.119

The increase in Ang-2 occurs coincident with that of IL-6, 30 minutes after that of TNF, and 30

minutes before that of IL-8. Of note, Ang-2 levels declined more slowly than did those of the

cytokines, yet faster than both E-selectin and ICAM-1. An accompanying study of 21 septic

patients found higher Ang-2 levels in non-survivors than survivors (28-day mortality), both on

admission and at 72 hours. Likewise, in a study of 50 critically ill patients, plasma Ang-2 levels

taken within 12 hours of the diagnosis of sepsis were predictive of ICU and 28-day mortality.120

Ang-2 levels on the day of enrolment and throughout the ICU admission correlated with fluid

balance, with standard indices of pulmonary dysfunction, and with mortality, suggesting a

potential role for serial measurements of Ang-2. vWF was not as uniformly associated with

18

illness severity, prompting the authors to speculate that, despite co-storage in the Weibel-Palade

bodies, differential regulation might be relevant.7 Other studies have found a similar relationship

between Ang-2 concentration and the severity of illness in sepsis: Orfanos et al described a

correlation between serum Ang-2, severe sepsis, and TNF,121

Siner et al, Davis et al, and

Kümpers et al between serum Ang-2, Sequential Organ Failure Assessment (SOFA) and Acute

Physiology And Chronic Health Evaluation (APACHE) II scores, and mortality,122-124

and Parikh

et al between serum Ang-2 (but not Ang-1) and impaired pulmonary gas exchange, but not with

survival or APACHE II score.9 Ang-2 has been found to be similarly predictive of mortality in

certain subsets of critically ill patients, including those requiring renal replacement therapy and

in children with septic shock.125, 126

In contrast, decreased plasma Ang-1 at ICU admission predicted 28-day mortality in 70 patients

with severe sepsis, while decreased Ang-1 levels were also associated with mortality in a study

of Malawian children with severe bacterial infection.127, 128

Furthermore, activated protein C,

which until recently was approved for the treatment of severe sepsis, has been found to

upregulate both Ang-1 and Tie-2, thereby improving endothelial barrier function in a manner

which may have been responsible for some of the therapeutic benefit documented in early

clinical trials.129

Importantly, the clinical utility of Ang-1 as a prognostic biomarker may be

limited by the frequent fluctuations in circulating levels seen in individual patients with hourly

sampling over a 24-hour time period.118

The Ang-2:Ang-1 ratio, reflecting perturbations in one

or both of the angiopoietins, has been found to be predictive of the eventual development of

septic shock at the time of fever onset in neutropenic patients.130

Finally, although the soluble

Tie-2 receptor does increase in sepsis, it has no apparent effect on Ang-1/2 function or

circulating levels, and was not found to be a significant indicator of pulmonary edema nor a

significant predictor of mortality in a multivariate analysis.131

1.7.2 Angiopoietins in Malaria

In malaria, endothelial activation leads to adherence of parasitized erythrocytes, microvascular

obstruction, and tissue ischemia. Ang-2, but not VEGF, was found to be elevated in severe

malaria (and a better predictor of outcome than lactate) in association with reduced NO

bioavailability.132

Subsequent studies, however, have all identified Ang-1 as a more consistent

biomarker in malaria, discriminating between cerebral malaria and uncomplicated malaria in

19

Thai adults and Ugandan children, between cerebral, severe, and uncomplicated malaria in Thai

adults, and between cerebral malaria with retinopathy and febrile, non-malarial decreased level

of consciousness in Malawian children.133-135

Furthermore, decreased Ang-1 was associated with

Plasmodium falciparum malaria in pregnancy, while an increased Ang-2:Ang-1 ratio was

associated with both placental malaria and low birth weight infants.136

1.7.3 Angiopoietins in Cardiovascular Disease

Given the prominence of the angiopoietin system in vascular remodelling, Ang-1 and Ang-2

have been proposed as both mediators and potential therapeutic targets for a variety of primary

cardiovascular diseases. Higher levels of Ang-2 were associated with systolic blood pressure,

diabetes, and the metabolic syndrome in 3778 third-generation cohort participants of the

Framingham Heart Study.137

Likewise, David et al found a higher level of plasma Ang-2 in

hypertensive patients as compared to healthy controls, and in those patients with atherosclerotic

disease as compared to those without.138

Despite purported anti-inflammatory effects separate

from their lipid-lowering properties, a six-week course of therapy with a 3-hydroxy-3-methyl-

glutaryl (HMG)-CoA reductase inhibitor (“statin”) did not lower Ang-2 levels in this population.

It should be noted that other studies have documented an elevation in plasma angiogenic factors,

including VEGF, Ang-1, Ang-2, and sTie-2, in treated hypertensive patients with high

Framingham risk scores, attributed to the relative hypoxia resulting from rarefaction or loss of

the microcirculation that occurs in hypertension.139

As noted in the Framingham study, angiopoietin dysregulation has also been associated with

disorders of glucose and lipid metabolism. In vitro studies involving human cardiac

microvascular endothelial cells and human macrovascular endothelial cells have demonstrated

that exposure to high concentrations of glucose leads to specific inhibition of Ang-1/Tie-2-

induced Akt phosphorylation, resulting in a predicted loss of cell survival signalling.140

The Ang-

1/Tie-2-induced phosphorylation of Erk 1/2 was not affected, leading to a predicted net signal

favouring endothelial cell activation and migration. In clinical studies, Ang-2 is higher in

patients with diabetes than in healthy control subjects, and elevated Ang-2 is independently

associated with an elevated hemoglobin A1c (HbA1c), a marker of longterm glycemic control, as

well as with diabetic retinopathy and a composite marker of macrovascular disease that included

cardiovascular and cerebrovascular disease as well as neuropathy.141-143

20

Given its association with multiple cardiovascular risk factors, it is not surprising that elevated

levels of Ang-2 have been associated with coronary artery disease. Plasma levels of Ang-2 at

presentation are significantly higher in patients with acute myocardial infarction (AMI) than in

patients with unstable or stable angina, in whom Ang-2 levels are significantly higher than in

healthy controls.144

Microarray analysis of myocardial biopsies from patients undergoing urgent

coronary artery bypass grafting for AMI confirms significant upregulation of the Ang-2 gene.145

Ang-2 levels are also elevated in acute, versus chronic, congestive heart failure (CHF),146

and at

a baseline routine health examination in patients who experienced AMI within the next ten

years.147

The latter association was independent of the traditional cardiovascular risk factors.

Since vascular stabilization, repair, and regeneration are important prerequisites for functional

recovery following stroke, high post-event Ang-1 levels in the cerebral vasculature would be

predicted to have a beneficial effect. Indeed, in experimental animal models and in one

prospective clinical trial involving 95 patients, elevated Ang-1 levels were associated with

improved outcome after intracerebral hemorrhage (ICH).148

A serum Ang-1 level of ≥ 35 ng/mL

at 72 hours post-ICH yielded an odds ratio (OR) of 14.7 for the occurrence of a good outcome at

three months, defined as an improvement in the NIHSS (National Institutes of Health Stroke

Scale) of ≥ 50%.149

Consistent with these observations, patients with subarachnoid hemorrhage

who exhibited post-event cerebral vasospasm on transcranial Doppler ultrasound were found to

have more prolonged and severe suppression of serum Ang-1 levels when compared with both

healthy controls and patients with SAH but without cerebral vasospasm.150

Higher Ang-1 levels

likewise correlate with more rapid middle cerebral artery recanalization after tPA (tissue

plasminogen activator) therapy, while higher Ang-2 levels have been associated with risk of

recurrence following lacunar infarct.151, 152

Elevated levels of Ang-2 have also been implicated in

post-stroke breakdown of the blood-brain barrier and subsequent cerebral edema.153

1.7.4 Angiopoietins in Chronic Kidney Disease

A role for the angiopoietins in chronic kidney disease (CKD) was first suggested by animal

studies. Podocyte-specific overexpression of Ang-2 lead to albuminuria and apoptosis of the

glomerular endothelial cells, and to decreased expression of VEGF,154

an important local

regulator of the antagonist/agonist effects of Ang-2. In contrast, exogenous Ang-1 has been

found to attenuate renal dysfunction in a variety of models of acute and chronic kidney

21

injury.93,

155, 156

Angiopoietin dysregulation appears to be greatest in patients on hemodialysis

(although the comparison with those on peritoneal dialysis trended towards but did not reach

statistical significance), and resolves with renal transplantation.157

Neither time of sampling (pre-

or post- dialysis) nor transition to nocturnal hemodialysis for a period of 6 months altered

angiopoietin dysregulation. Furthermore, in patients with CKD on dialysis (either hemo- or

peritoneal), serum Ang-2 levels correlated with the presence of coronary heart disease and

peripheral arterial disease. The same authors described a progressive increase in serum Ang-2

with each successive stage of chronic kidney disease, and attributed angiopoietin dysregulation

in this setting to the reduction in nitric oxide, a known inhibitor of Weibel-Palade body

exocytosis, seen in CKD.158

Given the relatively high molecular weight of the angiopoietins, and

their existence as multimers, they are not candidates for glomerular filtration, nor are they

detectable in urine (Page AV, unpublished observations). Therefore, elevated serum levels of

Ang-2 in CKD are likely related to increased release, rather than decreased renal clearance.

Further work by Kümpers et al determined that elevated Ang-2 was an independent predictor of

mortality in CKD.159

1.7.5 Angiopoietins in Pregnancy and Preeclampsia

The angiopoietins are frequently implicated in the vascular complications of pregnancy, but

published reports differ with regard to the direction of change of Ang-1, Ang-2, and soluble Tie-

2. Serum Ang-2 levels are higher in healthy pregnant women than in non-pregnant or post-

partum women.160

Circulating Ang-2 levels were initially reported to be lower, and Ang-1 levels

higher, in women with preeclampsia than in those with normal pregnancies.161

Further study

revealed that this difference might be present in the third trimester only: the two groups of

patients are indistinguishable in the first trimester, and those patients who will subsequently

develop preeclampsia typically have higher serum Ang-2 concentrations in the second

trimester.162, 163

More recently, several authors reported a high Ang-2:Ang-1 ratio, primarily as a

result of lower Ang-1, in women in the second trimester of pregnancy who would later develop

preeclampsia, and in those with established preeclampsia.164, 165

Low soluble Tie-2 levels have

also been found in women with preeclampsia.166

In summary, while angiopoietin dysregulation

as a marker or contributor to preeclampsia and other hypertensive disorders in pregnancy is

certainly plausible, additional research is needed to clarify the relevant timing, direction, and

magnitude of change of Ang-1, Ang-2, and soluble Tie-2.

22

Of note, Ang-2 has also been detected in the amniotic fluid of healthy women in the second and

third trimesters of pregnancy, and is elevated in those with an intrauterine infection, consistent

with its known links to inflammation.167

1.7.6 Angiopoietins in Malignancy

Consistent with what is known about the role of Ang-2 in development, tumours from transgenic

mice that overexpress Ang-2 have increased microvessel density, an immature vascular

phenotype with a deficit of pericytes, and an enhanced rate of endothelial cell apoptosis.168

Various malignant cells, including chronic lymphocytic leukemia (CLL) cells, can secrete Ang-

2.169

In addition, gastrointestinal stromal tumours, leiomyomas, schwannomas, colorectal

adenocarcinomas, and to a lesser extent Kaposi‟s sarcomas and cutaneous angiosarcomas, have

all been shown to harbour Ang-1, Ang-2, Ang-4, Tie-1, and Tie-2 in the tumour cell

cytoplasm.170-172

Although the clinical significance of this finding is not known, the presence of

both the ligand and receptor in the same cell suggests that the tumour cells may be able to

respond to autocrine or paracrine signalling through the angiopoietin/Tie system.

Further evidence of the interaction between the angiopoietin/Tie system and various

malignancies is derived from cohort studies. Elevated circulating Ang-2 levels have been

associated with reduced time to first treatment and higher mortality in CLL,169

increased

mortality in acute myeloid leukemia (14.7% three-year survival in patients with Ang-2 levels ≥

1495.6 pg/mL versus 64.7% in those with Ang-2 levels < 1495.6 pg/mL),173

stage of disease and

bone involvement in multiple myeloma,174

baseline disease and cardiac involvement in primary

light chain (AL) amyloidosis,175

baseline disease and lymph node metastases in early gastric

cancer,176

baseline disease in castration-resistant prostate cancer,177

reduced survival in

pancreatic ductal adenocarcinoma,178

baseline disease, residual disease after debulking surgery,

and poor overall and disease-free survival in ovarian carcinoma,179

increased mortality at the

time of diagnosis of colorectal carcinoma,180

baseline disease, tumour burden, and prognosis in

neuroendocrine tumours,181, 182

disease stage, metastases, and overall survival in non-small cell

lung cancer,183

survival and disease progression in metastatic malignant melanoma,184

baseline

disease in liver cirrhosis and hepatocellular carcinoma,185

increased tumour size in soft tissue

sarcoma,186

baseline disease in breast cancer,187

and inversely with tumour burden in response to

treatment in metastatic renal cell carcinoma.188

Interestingly, high bone marrow concentrations

23

of Ang-1 (studied following reports of increased bone marrow microvessel density) have

recently been reported to be independently associated with a higher rate of transformation to

acute leukemia and with overall mortality in myelodysplastic syndromes.189

Of note, the angiopoietin/Tie system has also been implicated in adverse effects associated with

cancer therapy. Vascular leak syndrome occurs in more than half of the patients treated with

high-dose IL-2 for metastatic renal cell carcinoma or melanoma, and in these patients, Ang-2

levels rose with each consecutive day of therapy and declined with discontinuation.190

Furthermore, when applied to endothelial cells in culture, patient serum from the final day of IL-

2 therapy resulted in endothelial gaps and increased actin stress fibres, pathophysiologic

correlates of the vascular leak syndrome in vivo. In patients receiving the anti-VEGF antibody

bevacizumab, upregulation of Ang-1 as a result of VEGF inhibition limits tumour hypoxia and

protects the tumour vasculature from regression, thereby potentially contributing to tumour

recurrence despite ongoing VEGF blockade.191

1.7.7 Angiopoietins in Respiratory Diseases

The rationale for the study of the angiopoietins in obstructive airway diseases derives from

clinicopathologic studies that have documented an increase in the number and permeability of

blood vessels in the airway walls in chronic obstruction as recently reviewed by Zanini et al, and

from the growing need for novel, non-steroid-based therapies.192

Expression of Tie-2 is

upregulated in airway epithelial cells and selected macrophages in the lungs of ovalbumin

-sensitized and -challenged mice, a recognized animal model for airway hyperresponsiveness,

and is correlated with the degree of airway remodeling.193

Ang-2 expression was increased, and

Ang-1 expression decreased, in the same cells. In clinical studies, asthmatics who smoked had

significantly higher levels of Ang-2 in induced sputa than did asthmatics who did not smoke, and

these levels were positively correlated with the airway vascular permeability index.194

In patients

with chronic obstructive pulmonary disease (COPD), serum Ang-2 levels were significantly

higher in those with stable disease as compared to healthy controls, and higher still in those with

acute exacerbations.195

Ang-2 levels in this instance were inversely correlated with PaO2.

Furthermore, both former and current smokers had higher serum levels of Ang-2 than did those

patients who had never smoked. Since VEGF is also increased in obstructive airway diseases,196

and since Ang-2 is known to promote new vessel sprouting with endothelial migration and

24

proliferation in the presence of VEGF, the predicted impact of the net increase in Ang-2 is

increased angiogenesis, thereby providing a plausible explanation and potential therapeutic target

for the airway vascular proliferation characteristic of chronic obstructive lung diseases.

Chronic pulmonary vascular remodelling also occurs in pulmonary hypertension. In

experimental models of pulmonary arterial hypertension (PAH), Ang-1 has been found to be

protective against vascular remodelling, while Ang-2 is elevated in hypoxia-induced PAH,

consistent with the ability of hypoxia to trigger Weibel-Palade body release from endothelial

cells.197, 198

In patients with PAH, elevated plasma Ang-2 has been found to be inversely

correlated with the response to treatment, and to be an independent predictor of mortality.199, 200

1.7.8 Angiopoietins in Autoimmune Connective Tissue Diseases

Angiopoietin dysregulation has been associated with a variety of autoimmune disorders. When

compared to healthy controls, patients with newly diagnosed rheumatoid arthritis (RA) had

elevated levels of serum Ang-2, and these levels correlated with the severity of disease and

measures of disease activity.201

Interestingly, patients who later developed cardiovascular disease

(at a mean interval of 12.5 years after the onset of RA) had higher levels of Ang-2 at arthritis

diagnosis than did those patients who never developed cardiovascular disease over the same

length of follow-up. Elevated serum Ang-2 levels (and the Ang-2:Ang-1 ratio) have also been

found to correlate with disease activity and proteinuria in patients with Systemic Lupus

Erythematosus (SLE), and with disease activity and/or renal involvement in patients with various

vasculitides.202-204

Furthermore, Ang-2 expression was upregulated in the glomeruli of patients

with SLE-associated renal disease. Additional data suggesting that the angiopoietin system may

contribute to the pathogenesis of autoimmune diseases is derived from patients with autoimmune

thyroid disease, in whom monocyte overexpression of the Tie-2 receptor appears to enhance the

chemotactic response to Ang-1 and Ang-2 overproduction by thyroid follicular cells, leading to

an inflammatory cell infiltrate in the thyroid parenchyma.205

Serum Ang-1 levels in patients with early RA (at the time of first clinic visit) were found to be

elevated when compared with those in patients with other collagen vascular diseases, a

discriminatory effect not seen with serum C-reactive protein or rheumatoid factor.206

Serum

Ang-1 levels are likewise elevated in patients with Behçet‟s disease as compared to healthy

controls,207

suggesting that angiopoietin dysregulation is not a uniform process across

25

rheumatologic disorders, and therefore, that its clinical utility should be carefully evaluated with

particular attention to the type, stage, and degree of complications in each autoimmune disease.

1.7.9 Angiopoietins in Miscellaneous Diseases and Syndromes

Plasma levels of Ang-2 have been shown to rise within 30 minutes of major trauma, and to

correlate with the severity of injury, the degree of tissue hypoperfusion, and in-hospital mortality

in trauma victims.208

Circulating Ang-2 is elevated in asymptomatic patients with sickle cell

disease as compared to healthy controls (HbSS > HbSC > HbAA), and is further elevated in both

groups during painful crises.209, 210

Since vascular remodelling is prominent in the

pathophysiology of inflammatory bowel disease, serum Ang-2 was investigated and found to be

elevated in patients with Crohn‟s disease and ulcerative colitis (UC).211

Epithelial cells from

crypt abscesses in biopsy specimens from UC patients were found to overexpress Ang-1, Ang-2,

and Tie-2, while the neutrophils in those abscesses overexpressed Tie-2, suggesting the ability to

respond to Ang-1 or Ang-2 as chemotactic factors.212

These findings are strikingly similar to

those of autoimmune thyroid disease (discussed above). In patients with acute pancreatitis,

admission serum Ang-2 levels were found to be predictive of later organ failure that persisted for

more than 48 hours.213

Finally, serum Ang-2 levels have been found to be increased in patients

with hepatitis C-induced liver disease as compared to healthy controls, and to decline in patients

with a sustained virologic response to therapy.214

Consistent with these findings, in vitro

experiments have shown that Ang-2 is upregulated in hepatocytes infected with either hepatitis C

or hepatitis B.215

26

1.8 Angiopoietin-based Therapies

Treatment strategies targeting the angiopoietin/Tie-2 system have been proposed for a wide

variety of disorders and diseases, although most have not yet progressed beyond testing in

experimental models. In vitro or animals models using Ang-1 mimetic or Tie-2 stimulatory

therapies have been successful in the treatment of: acute lung injury (ALI),216

pulmonary

hypertension,217

diabetic wounds,218

fractures, ischemic necrosis, and destructive bone diseases

(through enhancement of osteoblast differentiation and to promote angiogenesis in ex vivo bone

tissue engineering),219-221

cardiac, cerebral, and limb ischemia,222-224

ischemia-reperfusion-

induced acute kidney injury,155

radiation-induced apoptosis of the microcapillary endothelial

cells of the intestinal villi,225

and spinal cord injury,226

while those using Ang-2 inhibitory

therapies have demonstrated positive results in the treatment of hematologic and solid tumour

malignancies.227, 228

Not all studies have been positive, however. Despite multiple animal and

human studies documenting a link between angiopoietin dysregulation and pulmonary edema,

treatment with Ang-1 did not attenuate ventilator-induced lung injury in mice.229

The sole

clinical study of angiopoietin therapy involved AMG 386, a peptide-Fc fusion protein that blocks

the Tie-2 receptor. In a Phase 1 study of 22 patients with various advanced solid tumour

malignancies, AMG 386 was well-tolerated and appeared to exert some anti-tumour effect,

although all patients were receiving concomitant full-dose standard chemotherapy.230

While the

angiopoietin/Tie-2 system is an attractive therapeutic target, further studies are clearly needed to

determine when, how, and in what patient population anti-angiogenic therapy is best deployed.

27

1.9 Biomarkers of Endothelial Dysfunction in Sepsis

The manifestations of sepsis are notoriously protean (the 2001 International Definitions

Conference limned 24 criteria, some or all of which may be used to diagnose sepsis in

association with documented or suspected infection), yet early diagnosis and treatment is crucial,

with delay directly associated with mortality.25, 231, 232

The results of confirmatory diagnostic

tests, and in particular microbiologic cultures, can take hours to days. Furthermore, clinical

scores and traditional markers of inflammation (leukocyte count) and tissue hypoxia (lactate) are

non-specific or reflect the end result of sepsis rather than the underlying pathophysiology.

Therefore, the ideal biomarker for use in sepsis would derive from one of the two major

dysregulated systems (either the inflammatory or endothelial/vascular system), would

discriminate between septic and non-septic patients with SIRS, would allow prognostication at

the time of first contact with the healthcare system, and would be reproducible across various

patient populations. What follows is a discussion of the evidence for and against several

proposed diagnostic and/or prognostic biomarkers, representative of either the inflammatory or

vascular pathway, and their use in sepsis.

1.9.1 Procalcitonin

A prodigious literature has been generated describing the rationale for and clinical utility of

procalcitonin (PCT) as a diagnostic and prognostic biomarker in a variety of infectious diseases,

and has been most recently reviewed by Gilbert.233

Procalcitonin is the 12.6 kDa prohormone of

calcitonin, itself produced by the medullary C cells of the thyroid gland in response to

hypercalcemia. Plasma levels in healthy adults are low, but procalcitonin levels rise within 2-4

hours of bacteremia or endotoxemia, earlier than other markers of inflammation such as C-

reactive protein (CRP) or the erythrocyte sedimentation rate (ESR). PCT appears to be relatively

specific for infection, and does not increase in response to non-infectious inflammatory states

such as autoimmune connective tissue disease or inflammatory bowel disease.234, 235

Interestingly, PCT production by endothelial-adherent monocytes is transient, and has only been

documented in vitro.236

PCT in patients with infection appears to arise from monocyte-stimulated

adipocytes, and has been documented to act as an in vitro monocyte chemoattractant.237, 238

Human neutrophils increase production of pro-inflammatory cytokines in response to PCT in

vitro, but demonstrate decreased migration.239

Hence, the pathophysiologic role of PCT in the in

28

vivo response to infection is not clear. Further obscuring its true in vivo effect are studies

demonstrating that exogenous procalcitonin produces no effect when administered to healthy

animals, but increases mortality when administered during sepsis.240

Despite its questionable role in the immune response, much interest has been generated in the

potential utility of PCT to differentiate bacterial from non-bacterial infection. A 2004 meta-

analysis by Simon et al included 12 prospective studies that compared PCT and CRP, and

calculated sensitivities of 85% and 82%, and specificities of 83% and 88%, for the use of PCT to

differentiate between bacterial infection and either non-infectious inflammatory states or viral

infections, respectively.241

In both cases, PCT outperformed CRP. A second meta-analysis

focused on PCT in the Emergency Department or at admission for the diagnosis of bacteremia,

and included 17 studies (none included in the 2004 analysis) with a total of 2008 patients.242

PCT

was found to be moderately predictive of bacteremia, with a negative likelihood ratio of 0.3

helping to rule out the diagnosis. PCT levels have also been found to discriminate between Gram

positive and Gram negative bacteremia in septic patients in the ICU.243

A cut-off value of 16

ng/mL yielded a positive predictive value of 83% and a negative predictive value of 74% for the

diagnosis of Gram negative bacteremia in this population. Neither CRP, nor leukocyte count, nor

the SOFA score were discriminative in this setting. Notably, however, mortality was also higher

in those patients with Gram negative bacteremia, and the positive likelihood ratio of 4.2

indicated only moderate clinical utility. When followed routinely in the ICU, a decrease in PCT

levels between days 2 and 3 is predictive of appropriate empiric antibiotic therapy,244

suggesting

that PCT may have a role in monitoring therapeutic response. PCT was able to distinguish

between patients with community-acquired pneumonia and those with non-infectious

exacerbations of asthma or COPD, however, CRP was slightly more predictive than PCT in this

setting.245

PCT at a cut-off value of 0.8 ng/mL had a sensitivity of 91% and specificity of 68%

for the diagnosis of secondary bacterial infection during the 2009 H1N1 pandemic influenza

season.246

It should be noted, however, that not all studies have confirmed the specificity of PCT

for bacterial infection.247

PCT has been evaluated in at least two meta-analyses, one including 33 studies and the other 18,

for its utility in distinguishing sepsis from non-infectious inflammatory causes of SIRS in

critically ill patients. However, the two studies reported widely divergent results. The first, by

Uzzan et al, evaluated 3943 patients and concluded that PCT had good predictive value and

29

outperformed CRP for the differentiation of sepsis versus SIRS in critically ill patients after

surgery or trauma, while the second, by Tang et al, evaluated 2097 patients and concluded that

PCT was insufficiently sensitive (71%) for the differentiation of sepsis versus SIRS in a mixed

population (medical and surgical) of critically ill adults.248, 249

The diagnostic odds ratio for PCT

reported by Tang et al was less than half that reported by Uzzan et al. In at least one study

reported since, of 539 consecutive adult patients presenting to the Emergency Department with

suspected infection, PCT could distinguish between severe sepsis and other infectious/non-

infectious presentations, but not amongst non-severe sepsis, non-septic bacterial infection, and

non-infectious presentations, the clinical situations in which it would be most helpful.250

Based

on these results, the utility of PCT seems to vary significantly according to the clinical setting

and patient population, making generalization difficult and reinforcing the need for specific

study in each population in which its use is intended.

Importantly, procalcitonin levels are not affected by advanced human immunodeficiency virus

(HIV) disease, and can distinguish between respiratory infection with Pneumocystis jirovecii,

Mycobacteria tuberculosis, and community-acquired bacteria in this population, as well as

between bacteremia and disseminated viral, mycobacterial, and fungal infections, localized

bacterial infections, and cerebral toxoplasmosis.251-253

In addition to its potential role in the diagnosis of bacterial infection, procalcitonin has also been

proposed as a prognostic marker in sepsis and severe bacterial infection. Table 1 presents a

thorough, though not exhaustive, compilation of positive and negative studies examining

procalcitonin for the prediction of mortality at Emergency Department presentation, hospital or

ICU admission, onset of illness, or up to 1 week into the course of illness. Some studies have

also advocated the use of the PCT trajectory over time as a means to increase predictive value.

Regardless, as is illustrated by Table 1, the most appropriate use of PCT (in terms of timing,

duration of sampling, and target patient population) for the prediction of mortality in sepsis

remains controversial.

30

Author Study Population Significant PCT

Measure

Predicted

Outcome

Notes

Bloos F et

al 254

175 patients requiring

mechanical ventilation for CAP,

HAP, and (new) VAP

PCT on days 1-14,

starting within 48 hours

of diagnosis

28-day mortality Daily PCT levels higher in non-

survivors from days 1-14; Initial cut-off

value 1.1 ng/mL, peak cut-off value 7.8

ng/mL

Predictive value equal to APACHE II

Boeck L et

al 255

101 patients with VAP Serum PCT at diagnosis 28-day mortality PCT levels failed to decline over time

in non-survivors

Boussekey

N et al 256

110 patients requiring ICU

admission for CAP

Serum PCT within 48

hours of ICU admission

ICU mortality CRP not significant

Charles PE

et al 244

180 patients with sepsis in the

ICU

PCT on day 3 ICU mortality PCT levels on day 1 not predictive

Clec‟h C et

al 257

75 consecutive ICU patients

with shock

Serum PCT on day of

diagnosis

ICU mortality SN 87.5%, SP 54%, 6 ng/mL cut-off

31

Author Study Population Significant PCT

Measure

Predicted

Outcome

Notes

Dahaba AA

et al 258

69 patients requiring ICU

admission for severe sepsis

within 24 hours of surgery

Plasma PCT at day 6 28-day mortality

due to sepsis

No difference in PCT between

survivors and non-survivors on days 1-

4; SN 85%, SP 89%, 3.2 ng/mL cut-off

day 6

Giamarellos-

Bourboulis

EJ et al 259

1156 hospitalized patients with

sepsis (234 in the ICU)

PCT with 24 hours of

diagnosis of sepsis

In-hospital

mortality

Cut-off values 0.12 ng/mL in non-ICU

patients and 0.85 ng/mL in ICU

patients

Gibot S et

al 260

63 critically ill patients with

sepsis

None ICU mortality Only sTREM-1 remained predictive of

mortality in a multivariate analysis

Plasma PCT within 12 hours of ICU

admission was measured

Guan J et

al 261

37 patients with septic shock and

initial PCT levels > 10 ng/mL

Δ PCT from days 1-5 28-day mortality

Hillas G et

al 262

45 consecutive patients with

VAP

PCT and ΔPCT on days

1 and 7

28-day mortality OR 7.23 for ΔPCT1-7

32

Author Study Population Significant PCT

Measure

Predicted

Outcome

Notes

Huang DT et

al 263

1651 patients presenting to the

ED with CAP

PCT at presentation 30-day mortality PCT identifies low-risk patients from

amongst those identified as high risk by

PSI and CURB-65

Karlsson S

et al 264

242 critically ill patients with

sepsis

None In-hospital

mortality

Serum PCT at day 0 measured

Lee CC et

al 265

525 consecutive patients

presenting to the ED with sepsis

Serum PCT on

admission

Early (≤ 5 d) and

late (6-30 d)

mortality

MEDS score > PCT > CRP for

predicting either outcome

Menendez R

et al 266

453 patients hospitalized for

CAP

None 30-day mortality Only CRP remained a significant

predictor of mortality in a multivariate

analysis

PCT on admission measured

Meng FS et

al 267

86 patients with severe sepsis

requiring admission to the ICU

Serum PCT on the day

of ICU admission

28-day mortality Outperformed CRP and APACHE II

for prediction of mortality

33

Author Study Population Significant PCT

Measure

Predicted

Outcome

Notes

Novotny A

et al 268

160 patients with sepsis

following major visceral surgery

Serum PCT on the day

of diagnosis of sepsis

Sepsis-related

mortality

SN 71%, SP 77% for PCT + APACHE

II at diagnosis

Oberholzer

A et al 269

124 patients with severe sepsis None 28-day mortality Neither baseline plasma PCT nor

change over days 1-4 were predictive

of mortality

Phua J et

al 270

72 consecutive patients with

septic shock within 24 hours of

admission to the ICU

Increase in plasma

lactate + serum PCT

between days 1 and 2

28-day mortality APACHE II was the only significant

variable at admission; IL-1β, IL-6, IL-

10, TNF not significant

Rau BM et

al 271

82 patients with secondary

peritonitis, within 96 hours of

symptom onset

PCT ≥ 1 ng/mL after

the first week

In-hospital

mortality

Outperformed CRP

Ruiz-

Alvarez MJ

et al 272

103 consecutive patients

requiring ICU admission for

suspected sepsis

None ICU mortality Neither serum PCT nor CRP within 24

hours of admission were predictive of

mortality

34

Author Study Population Significant PCT

Measure

Predicted

Outcome

Notes

van

Langevelde

P et al 273

464 febrile patients presenting to

the Medicine ED

PCT at presentation In-hospital

mortality

RR 5.0 (95% CI 1.7 - 15) for PCT level

≥ 0.5 ng/mL

Viallon A et

al 274

131 consecutive patients

presenting to the ED with sepsis

PCT at presentation 30-day mortality SAPS II and lactate significant; serum

IL-6, IL-8, TNF, and CRP not

significant

Wunder C et

al 275

33 consecutive ICU patients

with severe sepsis

Plasma PCT 28-day mortality APACHE III score also significant, but

not IL-6 or IL-10

Table 1. Selected studies evaluating the use of procalcitonin to predict mortality in adult patients with sepsis or severe bacterial

infection. Sepsis defined as per ACCP/SCCM standard definitions unless otherwise indicated.24

Community-acquired pneumonia (CAP),

hospital-acquired pneumonia (HAP), ventilator-associated pneumonia (VAP), Procalcitonin (PCT), Acute Physiology And Chronic

Health Evaluation (APACHE), Intensive Care Unit (ICU), C-reactive protein (CRP), Sensitivity (SN), Specificity (SP), soluble Triggering

Receptor Expressed on Myeloid cells-1 (sTREM-1), Emergency Department (ED), Pneumonia Severity Index (PSI), CURB-65 (alternate

pneumonia severity score), Mortality in Emergency Department Sepsis (MEDS) score, Odds Ratio (OR), Tumour Necrosis Factor (TNF),

Relative Risk (RR), Simplified Acute Physiology Score (SAPS), Change in (Δ).

35

Procalcitonin has also been proposed for use as a prognostic indicator in non-septic patients. In a

study of 472 consecutive patients admitted to the ICU of a tertiary care hospital in Denmark,

routine daily procalcitonin measurements revealed that a maximum procalcitonin level of ≥ 1

ng/mL and a rising procalcitonin level were both independent predictors of 90-day all-cause

mortality.276

It should be noted that leukocyte count and C-reactive protein (CRP) were also

measured but were not similarly predictive. Schneider et al studied 220 unselected patients who

required post-operative admission to an intensive care unit for therapy or monitoring and found

that, in a multivariate analysis, admission procalcitonin values were independently predictive of

in-hospital mortality (sensitivity 81% and specificity 80% for a cut-off value of 1.44 ng/mL) and

were superior in this regard to the APACHE II score and to IL-6.277

In children, procalcitonin levels have been associated with severity of illness in

meningococcemia and sepsis post-bone marrow transplant, and performed better than CRP in

both populations.278, 279

In the latter group, procalcitonin was also associated with mortality.

Similarly, in a study of 75 children with septic shock, ICU admission levels of procalcitonin,

TNF, and IL-10 were all associated with mortality, however, none performed as well as a

pediatric clinical risk score unless followed serially.280

In a recent meta-analysis, procalcitonin

proved to have good discriminative capacity for the differentiation of sepsis from other critical

illnesses in neonates.281

Carrol et al evaluated 377 Malawian children with meningitis or

pneumonia, of whom 279 had a confirmed bacterial etiology.282

Plasma procalcitonin was higher

in severe bacterial infection regardless of HIV status (as was CRP), and predicted mortality. Of

note, however, the area under the ROC curve for procalcitonin in the prediction of mortality was

only 0.61 (95% CI 0.50 - 0.71). Furthermore, the ability of both PCT and CRP to distinguish

between viral and bacterial pneumonia in Mozambican children was abolished in the presence of

detectable Plasmodium falciparum (malaria) parasitemia,283

indicating the need for caution in

both the use and study of PCT in malaria-endemic areas.

In summary, PCT may be useful in the diagnosis of bacterial infections, and may be a more

specific predictor of outcome than cytokines and other non-specific markers of inflammation

such as CRP. However, it may need to be measured over time, and it is not clear that it is a better

predictor of outcome in all cases than are standard clinical prediction scores. Furthermore, PCT

36

shows considerable variability in its utility in different populations, particularly amongst those

with malarial co-infection, indicating the need for further study.

1.9.2 C-reactive protein

C-reactive protein (CRP) is a 21 kDa acute phase reactant protein synthesized by the liver in a

process upregulated by inflammation.241

When evaluated for use as a biomarker, CRP has often

yielded contradictory results in patients with sepsis. As illustrated in Table 1, CRP, when

compared to PCT, has variously been found to be more, less, or equally predictive of mortality in

sepsis. CRP levels in ALI/ARDS have been reported in at least one study to be higher in

survivors than non-survivors, even when adjusted for potential confounding factors such as

corticosteroid use and hepatic failure.284

Therefore, despite the fact that CRP assays are widely

available for clinical use, the appropriate interpretation of the results in any given clinical context

will require further study.

1.9.3 sTREM-1

Triggering Receptor Expressed on Myeloid cells (TREM)-1 is found on monocytes,

macrophages, and neutrophils, and consistent with its distribution on cells of the innate immune

system, is upregulated by TLR signalling.285

TREM-1 is thought to amplify the immune response

to pathogens, but whether it is beneficial, detrimental, or potentially both (in a context-dependent

manner) is in question. In a mouse model of pneumococcal pneumonia, TREM-1 signalling

enhances cytokine production, bacterial clearance, and survival.286

Consistent with this

protective effect, soluble TREM-1 (sTREM-1) levels on the day of ICU admission in 63 patients

with sepsis were lower in non-survivors than survivors, and remained predictive of outcome in a

multivariate analysis (in contrast to procalcitonin and CRP).260

In contrast, blockade of TREM-1

has been variously reported to increase both survival and mortality in murine models of

polymicrobial sepsis induced by cecal ligation and puncture, but has also been shown to improve

survival in a murine model of endotoxemia.287, 288

Consistent with the latter finding, sTREM-1

has been shown to correlate with severity of illness in a study of 52 critically ill patients with

sepsis, while PCT and CRP did not.289

sTREM-1 levels were found to be higher in the non-

survivors than survivors at days 10 and 14 of admission. Other authors have also documented an

association between elevated levels of sTREM-1 and severity of illness in sepsis.290

Further

complicating the interpretation of sTREM-1 levels in sepsis are the findings from a study of 65

37

septic patients in a surgical ICU, in whom plasma sTREM-1 measured within 24 hours of

diagnosis could not distinguish between SIRS and sepsis, nor between survivors and non-

survivors (28-day mortality).291

sTREM-1 has also been proposed as a means to distinguish infectious from non-infectious

inflammatory states.292

In a small study of patients with febrile neutropenia, both sTREM-1 and

PCT were associated with bloodstream infections and mortality, with levels of both markers

higher in non-survivors, while in a larger study of patients with febrile neutropenia, sTREM-1

performed better for the diagnosis of microbiologically documented infection than did PCT, with

sensitivities of 88% and 48%, respectively.293, 294

sTREM-1 has been reported to discriminate

between infectious and non-infectious pulmonary infiltrates in hospitalized, non-critically ill

patients with acute respiratory illness, yielding a sensitivity of 83% and specificity of 63%.295

Furthermore, sTREM-1 performed better than usual clinical markers in predicting bacteremia in

patients with community-acquired pneumonia, with low admission sTREM-1 levels

demonstrating a negative predictive value of 99%.296

However, at least one study has reported

that sTREM-1 performs less well than CRP in distinguishing septic versus non-septic patients

with SIRS,297

underscoring the variability found in clinical studies of sTREM-1.

1.9.4 Chi3L1 (YKL-40)

Chitinase-3-like protein 1 (Chi3L1) is a 40 kDa protein secreted by activated neutrophils and

macrophages in response to IL-6. Levels of Chi3L1 peak 6 to 24 hours following IL-6

infusion.298

The function of Chi3L1 is not known, and although it is a secreted protein, no cell

surface or soluble receptors have yet been identified.

In hospitalized patients with bacterial community-acquired pneumonia, serum Chi3L1 levels are

elevated as compared to healthy controls, peak on day 1, and decline to baseline levels by day

3.299

Interestingly, patients with pneumococcal pneumonia demonstrated the greatest elevation in

Chi3L1, while those with pneumonia due to H. influenzae or the “atypical” organisms had

Chi3L1 levels only modestly above baseline. In keeping with this finding, serum Chi3L1 levels

are elevated in patients with pneumococcal bacteremia as compared to healthy controls, and

relate to the severity of infection (defined as need for hemodialysis, vasopressor medications, or

mechanical ventilation) and in-hospital mortality in these patients.300

In a multivariate analysis,

Chi3L1 was found to be an independent predictor of mortality in the same population. Elevated

38

Chi3L1 levels have also been documented in association with Gram negative sepsis. In

experimental human Escherichia coli (E. coli) endotoxemia, Chi3L1 levels rose significantly

above baseline at 2 hours and peaked at 24 hours post-infusion.301

The Chi3L1 response trailed

that of TNF and IL-6, whose levels peaked at 1.5 and 2 hours post-infusion, respectively. In

addition, proteomic analysis identified Chi3L1 as one of three proteins uniquely upregulated in

septic, as compared to non-septic, critically ill patients.302

Serum levels at the time of ICU

admission were significantly higher in septic versus non-septic patients, and correlated with

severity of illness and circulating levels of IL-6.

1.9.5 IP-10

Interferon-γ-inducible protein-10 kDa (IP-10 or CXCL10) is a pro-inflammatory chemokine that

is involved in the chemotaxis of monocytes and T cells to sites of inflammation. IP-10 was found

to correlate with severity of illness and mortality in a univariate analysis of a small study of 16

patients with sepsis, and in addition was found to have excellent sensitivity and specificity in the

diagnosis of bacterial sepsis in very low birth weight neonates.303, 304

IP-10 has been shown to

effectively discriminate between survivors and non-survivors of severe malaria in a study

conducted in Ugandan children,305

a finding in keeping with evidence from animal models in

which IP-10 gene knockout reduced peripheral parasitemia and mortality.306

IP-10 is known to be elevated in both latent tuberculosis (TB) and active TB, regardless of

immunosuppression due to corticosteroids and disease-modifying anti-rheumatic drugs

(DMARDs).307

IP-10 levels in the pleural fluid are significantly different between patients with

tuberculous pleural effusions and those with pleural effusions of other etiologies.308, 309

Furthermore, IP-10 has been studied and found to be effective as a novel measurable end-product

of a modified Interferon-γ Release Assay (IGRA) in which patient plasma is incubated ex vivo

with purified specific M. tuberculosis proteins.310

1.9.6 PF4

Platelet Factor 4 (PF4) is a marker of platelet activation after release from platelet alpha

granules, and is upregulated in septic shock.311

By binding to both protein C and

thrombomodulin, PF4 can augment activated protein C generation in vivo, and improve survival

in animal models of sepsis.312, 313

In a small study of ten patients with sepsis, PF4 levels were

39

elevated in the acute phase of illness and declined significantly with effective therapy,314

but the

utility of PF4 as a biomarker in sepsis has not been further investigated.

1.9.7 vWF

von Willebrand Factor (vWF) is stored in the Weibel-Palade bodies of endothelial cells for

release upon endothelial cell activation.5 It acts to stabilize the adhesion of platelets at sites of

vascular injury, and its release has been associated with microvascular thrombosis in animal

models of experimental endotoxemia, although not in cecal ligation and puncture-induced

polymicrobial sepsis.315, 316

In a small study of 14 patients with sepsis, bacteremia, and

disseminated intravascular coagulation (DIC), plasma levels of vWF were significantly elevated

when compared to non-septic hospitalized patients, and correlated with the plasma level of

fibrinogen degradation products (FDPs), a marker of the severity of DIC and therefore indirectly

of sepsis.317

There was no correlation with lactate or platelet count. Other studies have also

documented an association between vWF and severity of illness in sepsis.318

A study of 50

mechanically ventilated patients in the ICU, enrolled within 12 hours of the diagnosis of sepsis,

found that the vWF concentration on the day of enrolment, but not thereafter, was predictive of

mortality during the ICU admission.120

However, not all studies have confirmed the association

between vWF levels and mortality in sepsis.319

In addition, there is conflicting literature as to

whether vWF is predictive of the development of ARDS in at-risk patients,320-323

although

elevated vWF levels have been associated with mortality in patients with ALI.324

1.9.8 sICAM-1

ICAM-1 plays a crucial role in neutrophil recruitment to sites of inflammation and injury.

Expressed constitutively at low levels on the endothelial cell surface but substantially

upregulated during endothelial cell activation, ICAM-1 interacts with β2-integrin on the surface

of the neutrophil to mediate firm adhesion, a necessary prerequisite to eventual neutrophil

extravasation.325

A soluble form of ICAM-1 can be detected in plasma and increases in concert

with the cell surface form during inflammation. Hence, soluble ICAM-1 (sICAM-1) has been

studied as a marker of inflammation.

In a small study of 25 critically ill patients with sepsis, plasma levels of sICAM-1 at the time of

diagnosis/study entry were significantly higher than those in non-septic hospitalized patients and

40

in eventual non-survivors versus survivors.326

In a multivariate analysis, sICAM-1 remained

predictive of survival. Similar results were seen in other studies, including a large trial of 221

patients with sepsis, however, sICAM-1 performed less well in this setting than did other

markers of endothelial dysfunction such as sFlt-1 (see below).34, 327

1.9.9 VEGF

VEGF plays a pivotal role in angiogenesis as outlined in Chapter 1.2, and is also linked to the

inflammatory pathway.328

In a mouse model of sepsis, circulating VEGF increases in a time-

dependent manner.329

Furthermore, reducing VEGF signalling via a soluble form of its receptor

attenuated the clinical signs of sepsis in the same model. Despite such promising results, the

findings in human studies have been contradictory. While most authors agree that circulating

levels of VEGF are elevated in sepsis,330

studies have yielded conflicting results as to whether

the degree of elevation correlates with disease severity.123, 331-333

Although VEGF has been

associated with the development of ALI/ARDS in septic patients, it is less predictive of this

outcome than are the angiopoietins.320

Also unexpected given the results of the animal model

was a large study of 293 Malawian children with severe bacterial infection, in whom plasma

VEGF concentrations were higher in survivors than in non-survivors.128

However, this difference

did not remain significant in a multivariate analysis. Other studies of sepsis have also failed to

detect an association between VEGF level and outcome.334

1.9.10 sFlt-1

Fms-like tyrosine kinase receptor 1 (sFlt-1) is the first of two VEGF receptors (VEGFR1). Both

the transmembrane and soluble form of Flt-1 may act to trap VEGF and prevent binding to Fms-

like tyrosine kinase receptor (Flk-1; VEGFR2), which is capable of transducing a much stronger

signal.39

This role is consistent with the interactions of sFlt-1 and the Dll4/Notch system. During

sprouting angiogenesis, VEGFR1 is upregulated, and VEGFR2 downregulated, by Notch

signalling in stalk cells, thereby limiting their responsiveness to the surrounding high

concentrations of VEGF required to guide the tip cells, and promoting organized, unidirectional

angiogenesis.335

In this way, sFlt-1 may function to fine-tune the actions of VEGF.

In 83 patients with suspected infection, sFlt-1 levels upon presentation to the Emergency

Department were higher in those with than without shock, and correlated with the APACHE II

41

and SOFA scores.331

A similar study, this time including 221 patients with sepsis of varying

severity, confirmed those findings.34

In 10 patients with febrile neutropenia, sFlt-1 measured 48

hours after fever onset was significantly higher amongst those patients who would develop septic

shock than amongst those who would continue to manifest only uncomplicated sepsis.332

Nonetheless, neither the pathophysiologic role nor the source of sFlt-1 in sepsis is clear.

Of note, sFlt-1 has also been implicated in malaria, and plasma levels were predictive of

mortality in a large study of 156 Ugandan children with malaria.305

1.9.11 Miscellaneous Markers of Endothelial Dysfunction in Sepsis

As might be expected, many other circulating factors have been explored as potential prognostic

indicators in sepsis. A recent review documented 3370 references detailing studies on 178

different biomarkers.336

Few of those that pertain to endothelial dysfunction have as well-defined

a pathophysiologic role and as reproducible a clinical prognostic utility as the angiopoietins. For

instance, although pro-inflammatory cytokines would seem an obvious choice for prognostic

indicators in sepsis, their utility has not been confirmed in large studies. In a study of 60 patients

with severe sepsis, plasma levels of pro-inflammatory cytokines, including IL-1β, IL-6, IL-8, and

TNF, were higher at diagnosis in patients with shock than in those without.337

Plasma levels of

IL-8 and monocyte chemoattractant protein (MCP)-1 correlated best with the SOFA score, and

also predicted 48-hour and 28-day mortality, however, only MCP-1 was associated with outcome

in a multivariate analysis. One of the largest studies, involving 1690 patients at enrolment in the

PROWESS (recombinant human activated Protein C Worldwide Evaluation in Severe Sepsis)

study, measured IL-1β, TNF, IL-6, and IL-8, the pro-inflammatory cytokines most commonly

associated with sepsis.338

While IL-8 and TNF correlated with APACHE II score, few patients

had detectable IL-1β, TNF, or IL-10 (an anti-inflammatory cytokine) levels at study entry,

reflecting their early and transient release during sepsis. Furthermore, serum IL-6 levels at the

time of enrolment did not predict 28-day mortality in the 840 patients randomized to the study‟s

placebo arm. In addition, anti-cytokine therapies have not proven useful in the treatment of septic

patients.339, 340

Many other candidate biomarkers have also been studied in sepsis. E-selectin has been associated

with severity in sepsis and mortality in critically ill patients.34, 342

While some studies have

42

documented an association with outcome in severe infection, others have not.326, 342-344

Endothelial progenitor cells are known to play a role in vascular repair, are elevated in sepsis,

and have been found to correlate with the outcome of sepsis in at least one small study.345-347

Circulating endothelial microparticles have also been evaluated as markers of severity in sepsis,

however, controversy still exists as to whether they exert a pro- or anti- inflammatory effect.348,

349 Finally, serum levels of various matrix metalloproteinases (MMPs) were found to be

abnormal at the time of diagnosis of sepsis, although the direction of change was dependent on

the particular MMP studied. MMP-10 was higher in septic than non-septic patients, while MMP-

9 was lower in non-survivors than survivors of sepsis.350

Since conflicting data regarding MMPs

in sepsis exist, further study is needed.351

1.9.12 Conclusion

In summary, although multiple candidate biomarkers have been proposed for use in sepsis, none

have been uniformly and consistently shown to be clinically useful as either diagnostic or

prognostic indicators in sepsis. Nor have any met the prespecified criteria delineated in Chapter

1.9 for the ideal biomarker: one that would derive from one of the two major dysregulated

systems, discriminate between septic and non-septic patients with SIRS, allow prognostication at

the time of first contact with the healthcare system, and be reproducible across various patient

populations. Thus there is the need for further study in this area, and importantly, for the direct

comparison of performance characteristics of candidate molecules in a single patient population.

43

Chapter 2 Research Aims and Hypotheses

2.1 Research Aims

The specific aims of the project are:

1) To determine whether angiopoietin-1 and -2 are dysregulated in human diseases

characterized by diffuse endothelial cell activation,

2) To define the clinical consequences (morbidity and mortality) associated with

abnormal angiopoietin levels, and the prognostic or predictive value of these levels,

3) To explore the clinical utility of biomarker combinations to predict mortality in septic

patients in a resource-limited setting.

2.2 Hypotheses

1) Since the angiopoietins are prominent regulators of endothelial cell function,

dysregulation of the angiopoietins (decreased Ang-1 and increased Ang-2) will

correlate with disease severity and outcome in infectious diseases associated with

significant endothelial cell dysfunction, including STSS, HUS, and sepsis, and will be

predictive of disease progression (reflecting the severity of endothelial dysfunction)

when measured early in illness.

2) As outlined in Chapter 1, multiple pathways mediate reciprocal interactions between

the immune system and the endothelium. Since sepsis is characterized by both

immune and endothelial cell activation, combining biomarkers from the inflammatory

and vascular pathways will more closely approximate the underlying pathophysiology

and therefore will more accurately predict mortality in sepsis.

44

Chapter 3 Angiopoietin-1 and -2 in Invasive Group A Streptococcal Infection

Originally published in Clinical Infectious Diseases, 2011 352

3.1 Introduction

Streptococcal toxic shock syndrome (STSS) remains a serious disease associated with poor

clinical outcomes despite prompt recognition, timely surgical debridement, and rapid initiation of

appropriate antimicrobial therapy.13

Much of the morbidity and mortality results from the release

of toxins by invading group A streptococcal bacteria. GAS pyrogenic exotoxins function as

superantigens to activate T cells in a Major Histocompatibility Complex (MHC) class II

unrestricted manner, initiating a series of inflammatory reactions that result in diffuse endothelial

cell activation, vascular leak, and the clinical characteristics of STSS: hypotension, major organ

dysfunction, and coagulopathy.13

While massive pro-inflammatory cytokine release has been

implicated in the pathogenesis of STSS, we hypothesized that angiopoietin-1 and angiopoietin-2,

critical regulators of endothelial cell activation, might also play central roles in the development

of STSS.

3.2 Methods

Beginning in 1992, prospective, population-based surveillance for invasive group A

streptococcal disease has been undertaken in Ontario, Canada via mandatory laboratory reporting

of S. pyogenes isolates cultured from normally sterile sites.353

Thirty-seven patients, enrolled

between 1999 and 2009, were included in the current study. Informed consent was obtained to

collect bacterial isolates and plasma samples, as well as selected clinical data from interviews

with the attending physicians and patient chart review. Patients with invasive disease were

considered to have streptococcal toxic shock if they met the consensus definition: hypotension in

combination with at least two of coagulopathy, acute kidney injury, elevated serum

aminotransferases, ARDS, rash, or necrotizing fasciitis.13

Of the 37 patients included in this study, 16 had invasive streptococcal infection and toxic shock

(STSS), while 21 had invasive streptococcal infection alone (no STSS). The median age was not

significantly different between these groups (41.5 years with STSS versus 38 years without;

45

range from 10 to 74 years; P = NS). The patients were predominantly female (56% in the group

with STSS and 67% in the group without, P = NS), and otherwise healthy, with only one-third of

the patients in either group describing a pre-existing illness. Only 3 patients (2 with STSS and 1

with invasive disease alone) reported pre-existing illnesses potentially associated with

angiopoietin dysregulation. The underlying source of infection was similar in each group, with

the majority of patients in both groups having skin and soft tissue infections (7 patients (44%)

with STSS and 12 patients (57%) with invasive streptococcal infection alone, P = NS). Amongst

the patients with skin and soft tissue infections, 5 (71%) with STSS and 4 (33%) with invasive

infection alone had necrotizing fasciitis, P = NS. Presenting group A streptococcal infections in

the remaining patients included respiratory tract infections, bacteremia without an identified

source, post-partum infections, and peritonitis, and did not differ significantly between the

groups. The two groups were significantly different only in the diagnostic criteria for STSS;

hypotension was present in 100% of patients with STSS and 33% of patients without (P <

0.0001). Five patients with STSS died as compared to one patient with invasive infection alone

(31% versus 5%, P = 0.06).

Acute phase plasma samples were collected in heparinised tubes upon study enrolment and

stored at -70oC until use. Convalescent plasma samples were collected at least one month post-

enrolment and once all signs and symptoms of infection had resolved. Plasma concentrations of

angiopoietin-1 and -2 were measured by enzyme-linked immunosorbent assay (ELISA; R&D

Systems, Minneapolis MN) according to the manufacturer‟s instructions. Each sample was tested

in duplicate and all samples were assayed simultaneously. The upper and lower limits of

detection for the assays were 10 000 pg/mL and 9.77 pg/mL for Ang-1 and 2520 pg/mL and 2.46

pg/mL for Ang-2, respectively. Samples were diluted in assay diluent (1:20 for Ang-1 and 1:4

for Ang-2) to fall within the range of the standard curves. Laboratory testing was performed by

an investigator blinded to all clinical data.

Statistical analysis was performed using GraphPad Prism v4.03 and Instat v3.06 (GraphPad

Software, San Diego, CA). To describe the study population, categorical variables were analyzed

by the Fisher‟s exact test and continuous variables with the Student‟s t-test. Differences in

median serum angiopoietin levels were assessed by the Kruskal-Wallis test followed by Dunn‟s

test for multiple comparisons. Matched acute and convalescent samples were compared using the

Wilcoxon matched pairs test. Receiver operating characteristic (ROC) curves were generated

46

using patients with STSS as “cases” and those with invasive streptococcal disease alone as

“controls”, with the null hypothesis that the area under the curve (AUC) equals 0.5.

3.3 Results

Angiopoietin dysregulation (decreased Ang-1 and increased Ang-2) was associated with disease

severity in patients with invasive group A streptococcal infection (Figure 1A). The median

plasma concentration of Ang-1 was lower during the acute phase of illness in patients with

invasive infection and STSS than in those with invasive streptococcal infection alone (13 915

pg/mL, [interquartile range (IQR): 5295 - 23 876 pg/mL] vs. 29 084 pg/mL [IQR: 11 855 -

54 803 pg/mL]), while the median plasma concentration of Ang-2 was higher (5752 pg/mL,

[IQR: 1250 - 7695 pg/mL] vs. 1337 pg/mL [IQR: 539 - 2371 pg/mL]). As a result, the normally

low Ang-2:Ang-1 ratio was significantly higher amongst patients with invasive infection and

STSS as compared to those with invasive streptococcal infection alone (0.437 [IQR: 0.122 -

0.775] versus 0.048 [IQR: 0.014 - 0.327], P < 0.05).

Ang-1/2 dysregulation declined with convalescence in both groups of patients (Figure 1A). In the

cohort of patients with STSS, the median plasma concentration of Ang-1 rose from 13,915

pg/mL to 21 115 pg/mL (IQR: 3517 - 41 373 pg/mL), the median plasma concentration of Ang-2

decreased from 5752 pg/mL to 378 pg/mL (IQR: 225 - 1097 pg/mL), P < 0.01, and the median

Ang-2:Ang-1 ratio fell from 0.437 to 0.019 (IQR: 0.011 - 0.139), P < 0.05.

Receiver operating characteristic (ROC) curves were generated for Ang-1, Ang-2, and the Ang-

2:Ang-1 ratio, and the area under the ROC curves indicated that the degree of magnitude of Ang-

1/2 dysregulation accurately differentiated between those individuals with STSS and those

without STSS (Figure 1B). The ROC curves for plasma Ang-2 (AUC: 0.759; 95% confidence

interval (CI): 0.596 - 0.921; P = 0.009) and for the Ang-2:Ang-1 ratio (AUC: 0.791; 95% CI:

0.643 - 0.938; P = 0.003) discriminated between patients with STSS and those with invasive

streptococcal infection alone (no STSS). The ROC curve for plasma Ang-1 concentration

trended towards, but did not reach, statistical difference (AUC: 0.683; 95% CI: 0.506 - 0.860; P

= 0.07).

47

Figure 1. Angiopoietin -1 and -2 (Ang-1 and Ang-2) in patients with invasive Group A streptococcal disease with and without

streptococcal toxic shock syndrome (STSS). A. Absolute and median concentrations of Ang-1 and Ang-2, as well as Ang-2:Ang-1

expressed as log base 10, in acute and convalescent plasma from patients with or without STSS. * P < 0.05; ** P < 0.01. = survived; =

died. B. Receiver operating characteristic curves for each marker: Ang-1 (AUC: 0.683; 95% confidence interval (CI): 0.506 - 0.860; P =

0.07), Ang-2 (AUC: 0.759; 95% CI: 0.596 - 0.921; P = 0.009), and Ang-2:Ang-1 (AUC: 0.791; 95% CI: 0.643 - 0.938; P = 0.003).

48

In individual patients with STSS, the matched acute and convalescent plasma Ang-2

concentrations and the Ang-2:Ang-1 ratios also differed significantly (Figure 2). Although the

same pattern was observed in the cohort of patients with invasive streptococcal disease without

STSS, the changes in Ang-1/2 concentrations were more modest. The median plasma

concentration of Ang-1 in this group increased only from 29 084 pg/mL to 31 743 pg/mL (IQR:

8443 - 54 645 pg/mL), while the Ang-2 concentration declined from 1337 pg/mL to 535 pg/mL

(IQR: 203 - 1702 pg/mL), and the Ang-2:Ang-1 ratio decreased slightly from 0.048 to 0.027

(IQR: 0.008 - 0.091), P = NS for all comparisons.

49

Figure 2. Angiopoietin-1 and -2 (Ang-1 and Ang-2) concentrations, and the ratio between the two (Ang-2:Ang-1), in matched acute and

convalescent plasma samples from patients with STSS. P = 0.01 for Ang-2 in acute vs. convalescent plasma, and P = 0.03 for Ang-2:Ang-

1 in acute vs. convalescent plasma by the Wilcoxon matched pairs test.

50

3.4 Discussion/Conclusion

This study represents the first report of an association between angiopoietin dysregulation and

disease severity in invasive streptococcal infection. Ang-1/2 dysregulation, as reflected by an

increased Ang-2:Ang-1 ratio, was significantly greater during the acute phase of illness in

individuals with STSS than in those without (Figure 1). This finding confirms that angiopoietin

dysregulation in invasive streptococcal disease is specifically related to the degree of endothelial

dysfunction present, and not simply associated with the underlying infection. Further, the

observation is consistent with our current understanding of angiopoietin dysregulation in other

infectious disease syndromes associated with prominent endothelial cell activation. Ang-2 and/or

the Ang-2:Ang-1 ratio have been found to be markers of disease severity in sepsis, ALI/ARDS,

febrile neutropenia, and malaria.9, 130, 133, 134, 320

In our study, ROC curves suggested that both the

absolute Ang-2 concentration and the Ang-2:Ang-1 ratio could effectively discriminate between

patients with and without STSS (Figure 1).

The sequence of events leading to angiopoietin dysregulation during STSS is not yet known. S.

pyogenes produces a variety of factors, including the M proteins, that have been shown to

mediate endothelial cell function.354

Streptococcal superantigens may also be involved, whether

through direct action on the vascular endothelium or through intermediary molecules.

Streptococcal superantigens have long been known to mediate the release of pro-inflammatory

cytokines, including TNF, which might function as a link to the angiopoietin system. Circulating

and tissue-infiltrating mononuclear cells from patients with STSS have been shown to produce

significantly higher levels of TNF and other inflammatory cytokines as compared to cells from

patients with invasive streptococcal infection without shock, and TNF itself has been associated

with the upregulation of Ang-2 expression in human endothelial cells in vitro, although not

definitively with Ang-2 release.7, 11, 355

In turn, Ang-2 is known to sensitize endothelial cells to

the pro-inflammatory effects of TNF,104

reflecting the type of complex interplay between the

cytokine and angiopoietin systems that may occur during STSS. Ultimately, increased Ang-2

release in STSS leads to increased Ang-2 binding to the endothelial Tie-2 receptor, with

subsequent induction of the activated, pro-thrombotic endothelial cell phenotype that contributes

to fluid extravasation, shock, organ dysfunction, and coagulopathy.356

A similar mechanism has

been proposed to explain angiopoietin dysregulation in sepsis.356

Together, our findings as well

51

as those of others are consistent with the possibility of a shared final common pathway of

endothelial activation and diffuse vascular leak in the pathogenesis of sepsis, ALI/ARDS, and

STSS.120, 320, 357

Although the current study cannot prove causation, it suggests that angiopoietin

dysregulation may contribute to disease severity, and provides a rationale for future translational

studies in this area.

Consistent with this hypothesis, convalescence and clinical recovery were associated with

progressive normalization of the balance between Ang-1 and Ang-2, most dramatically in

survivors of STSS (Figure 2). Similar results have been reported in sepsis, in which Ang-2 levels

were also found to decline with convalescence.9 In agreement with the observation that

endothelial cell function (activation versus quiescence) reflects the balance between local

concentrations of both Ang-1 and Ang-2, we observed that the Ang-2:Ang-1 ratio was a

consistent marker of both STSS and convalescence.

There are some notable limitations to this study. The small sample size restricted our ability to

detect potentially important differences, particularly in Ang-1 levels that displayed a trend

towards statistical significance. In addition, the overall mortality rate was low and not

significantly different between those with STSS and those without, which limited the potential

for conclusions regarding the prognostic value of plasma concentrations of Ang-1 and/or Ang-2

for clinical outcome. Finally, since serial samples were not available during the acute phase of

illness in the current study, it is not possible to comment upon the predictive value of early

angiopoietin dysregulation for disease progression and the development of STSS in invasive

streptococcal infection. Elevation of the Ang-2:Ang-1 ratio and serum Ang-2 concentration have

recently been reported to be predictive of the development of septic shock in patients with febrile

neutropenia when measured at or within 48 hours of fever onset, respectively.130

It is therefore

possible that close monitoring of the kinetics of angiopoietin dysregulation in the plasma of

patients with invasive streptococcal disease could provide a reliable prognostic biomarker for

disease progression and severity, and an indicator of response to various therapeutic

interventions.

Despite the above limitations, our study adds to the growing body of evidence implicating

endothelial cell dysfunction and angiopoietin dysregulation in life-threatening infectious disease

syndromes, and suggests that angiopoietin dysregulation may be both a contributor to, and a

52

clinically informative biomarker of, disease severity in invasive streptococcal infection. On this

basis, further studies exploring the potential role of angiopoietin dysregulation in the

pathogenesis of STSS and the predictive value of angiopoietin dysregulation for both the

development of STSS and mortality in invasive streptococcal infection are warranted.

53

Chapter 4 Angiopoietin-1 and -2 in the Hemolytic-Uremic Syndrome

Submitted for Publication, 2012

4.1 Introduction

The haemolytic-uremic syndrome (HUS) is defined by the triad of non-immune hemolytic

anemia, thrombocytopenia, and renal failure.16

Once thought to be a consequence of toxin-

induced glomerular endothelial cell ribosomal arrest and apoptosis, recent evidence suggests a

more complex pathogenesis in HUS, involving Shiga toxin-induced microvascular endothelial

cell activation.23

Although studies have shown that Shiga toxin can induce a prothrombotic and

adhesive endothelial cell phenotype,358

the precise mechanism for endothelial cell

activation/dysfunction in HUS has yet to be definitively determined.

We hypothesized that angiopoietin dysregulation (decreased Ang-1 and increased Ang-2) might

be present and contribute to the pathogenesis of HUS. The bacterial serotype most commonly

implicated in typical HUS is E. coli O157:H7, and those most at risk of developing HUS as a

consequence of STEC hemorrhagic colitis are children.359

Therefore, we studied serum

angiopoietin concentrations at various time points throughout illness in children with E. coli

O157:H7 infection, with and without HUS.

4.2 Methods

Study population and specimen collection

Beginning in 1997, population-based surveillance for E. coli O157:H7 infection in children less

than 10 years of age was undertaken in Washington, Oregon, Idaho, and Wyoming through

mandatory laboratory reporting of positive stool cultures as previously described.10

Children

from whom a positive culture was obtained within the first seven days after the onset of diarrhea

were eligible for enrolment. Seventy-eight children, enrolled between 1998 and 2005, were

included in the current study. Written informed consent from the parents, and assent from the

child where appropriate, was obtained at study entry for the collection of epidemiologic and

clinical data and for phlebotomy both at enrolment and as clinically indicated thereafter. HUS

was diagnosed in the setting of hemolytic anemia (a hematocrit < 30% with evidence of

54

schistocytes on peripheral blood film), thrombocytopenia (platelet count < 150 000 cells/mm3),

and renal insufficiency (serum creatinine above the age-adjusted upper limit of normal);

participants who had not met these criteria by day 14 following the onset of diarrhea were

considered to have had uncomplicated infection.18

In total, 84 serum samples were tested (Figure

3): 26 from patients on the day of diagnosis of HUS, 8 from patients who would subsequently be

diagnosed with HUS but had not yet met diagnostic criteria (pre-HUS), and 50 from patients

with uncomplicated infection. Six patients had samples taken both prior to (pre-HUS) and on the

day of, HUS diagnosis.

Figure 3. Timeline of blood sampling and division into study groups (uncomplicated infection,

pre-HUS, HUS) of participants at various stages of E. coli O157:H7 infection.

Measurement of serum Ang-1 and Ang-2

Serum samples were stored in aliquots at -80oC until use. Serum concentrations of Ang-1 and

Ang-2 were measured by ELISA (R&D Systems, Minneapolis MN) as per the manufacturer‟s

instructions. Each sample was tested in duplicate and all samples were assayed simultaneously.

The technical upper limits of detection for the assay were 10 000 pg/mL for Ang-1 and 2520

Uncomplicated Infection

Pre-HUS and HUS

55

pg/mL for Ang-2, yielding effective upper limits of detection of 200 000 pg/mL and 10 080

pg/mL, respectively, for the dilutions employed in the analysis. Lower limits of detection for the

assay were 9.77 pg/mL for Ang-1 and 2.46 pg/mL for Ang-2.

Statistical analyses

Statistical analysis was performed using GraphPad Prism v.4.03 (San Diego, CA). Median serum

angiopoietin levels were compared using the Kruskal-Wallis test followed by Dunn‟s test for

multiple comparisons. Matched pre-HUS and HUS samples were compared using the Wilcoxon

matched pairs test. Receiver operating characteristic (ROC) curves were generated using pre-

HUS patients as “cases” and those with uncomplicated infection as “controls”, with the null

hypothesis that the area under the curve (AUC) equals 0.5.

4.3 Results

In this study of 78 children with E. coli O157:H7 infection, angiopoietin dysregulation

(decreased Ang-1 and increased Ang-2) was associated with illness severity (Figure 4A). The

median serum Ang-1 concentration in those with uncomplicated infection was significantly

higher than in those with HUS (77 357 pg/mL [interquartile range (IQR): 53 437 - 114 889

pg/mL] versus 10 622 pg/mL [IQR: 3464 - 43 523 pg/mL]), P < 0.001. Conversely, the median

serum Ang-2 concentration was significantly lower in those with uncomplicated infection than in

those with HUS (1140 pg/mL [IQR: 845 - 1492 pg/mL] in patients with uncomplicated infection

versus 1959 pg/mL [IQR: 1057 - 2855 pg/mL] in those with HUS), P < 0.05. Finally, the Ang-

2:Ang-1 ratio was 0.014 (IQR: 0.011 - 0.023) in patients with uncomplicated infection, and more

than 10-fold higher, at 0.18 (IQR: 0.066 - 0.51) in those with HUS, P < 0.001.

56

Figure 4. Serum Angiopoietin Concentration in E. coli O157:H7 Infection. A. Angiopoietin-1

(Ang-1), Angiopoietin-2 (Ang-2), and the Ang-2:Ang-1 ratio expressed as log base 10 in

children with uncomplicated E. coli O157:H7 infection (infected), children prior to the diagnosis

of HUS (pre-HUS), and children at the time of diagnosis of HUS (HUS). *P<0.05; ** P<0.001,

o = outlier (1.5 x interquartile range [IQR]); = extreme outlier (3 x IQR). B. Receiver

Operating Characteristic (ROC) curves comparing Ang-1, Ang-2, and the Ang-1:Ang-2 ratio

amongst children with uncomplicated infection and those in the pre-HUS phase of illness, with

the null hypothesis that the area under the curve is 0.5. P = 0.01 for Ang-1.

57

Furthermore, serum Ang-1 concentration at the time of presentation to hospital effectively

discriminated between two populations of clinically indistinguishable children: 1) those with

uncomplicated hemorrhagic colitis and 2) those with hemorrhagic colitis who would eventually

develop HUS (Area under the ROC curve [AUC]: 0.785, 95% confidence interval (CI): 0.641 -

0.923; P = 0.01) (Figure 4B). A serum Ang-1 cut-off value of less than 48 000 pg/mL yields a

sensitivity of 75% and a specificity of 80% for the diagnosis of the pre-HUS phase of illness.

The ROC curves for the serum Ang-2 concentration and the Ang-2:Ang-1 ratio trended towards,

but did not reach, statistical difference (AUC: 0.600, 95% CI: 0.357 - 0.843; P = 0.37) and AUC:

0.685, 95% CI: 0.499 - 0.872; P = 0.09, respectively).

Data from individual patients closely mirrored the aggregate data presented above. Paired

samples from patients on the day of diagnosis of infection (day of positive stool culture, or pre-

HUS) and on the day of diagnosis of HUS, revealed the same trend: serum concentrations of

Ang-1 fell, while the Ang-2:Ang-1 ratio rose (Figure 5).

58

Figure 5. Serum Angiopoietin Concentration in Individual Patients with E. coli O157:H7

Infection before and after the diagnosis of HUS. Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-

2), and the Ang-2:Ang-1 ratio in matched serum samples from individual patients at the time of

diagnosis of E. coli O157:H7 infection (pre-HUS) and at the time of diagnosis of HUS (HUS).

59

4.4 Discussion/Conclusion

The finding of angiopoietin dysregulation in HUS is consistent with the known biologic effects

of Ang-1 and Ang-2, and with their predicted impact on endothelial cell function during the

course of E. coli O157:H7 infection. Ang-1 and Ang-2 compete for binding to the shared Tie-2

receptor.356

In healthy individuals, constitutive release and binding of Ang-1 results in

phosphorylation of the Tie-2 receptor, inhibition of the NF-κB pathway, and endothelial cell

quiescence.356

In patients with HUS, inflammation-induced release and binding of Ang-2 is

predicted to block the action of Ang-1, release the NF-κB pathway from inhibition, and activate

the endothelium.356

The serum Ang-1 and Ang-2 concentrations reported here for children with

uncomplicated infection are comparable to those found in the serum of healthy children and

adults included as control subjects in other studies, and are in keeping with the clinical

observation that there is little if any endothelial dysfunction present in these patients.124, 133

In

contrast, the relative deficit of Ang-1 and excess of Ang-2 in children with HUS would be

expected to produce significant endothelial cell activation. As a consequence, constitutive Ang-1

inhibition of the actions of TNF and VEGF, upregulated by Shiga toxin or in HUS

respectively,360, 361

would be diminished, leading to increased expression of tissue factor,

VCAM-1, ICAM-1, and E-selectin on the endothelial cell surface.81, 90

Furthermore, Ang-2 itself

has been shown to sensitize endothelial cells to the effects of TNF.104

The endothelial phenotype

that is predicted to result from an increased Ang-2:Ang-1 ratio is therefore prothrombotic and

adhesive, and in keeping with the characteristic thrombotic microangiopathy of HUS. Our

observations are similar to those made in other disorders of endothelial cell function, including

sepsis, cerebral malaria, and streptococcal toxic shock.81, 90, 352

In each of these infectious

syndromes, an association has been described between the degree of angiopoietin dysregulation

and the severity of illness, and likely reflects a progressive degree of endothelial dysfunction.

Taken together, these findings suggest that angiopoietin dysregulation may represent a final

common pathway through which various infectious insults act to promote endothelial cell

activation.

In HUS, angiopoietin dysregulation likely occurs either directly or indirectly through the action

of the Shiga toxin. Angiopoietin-2 is stored in the Weibel-Palade bodies of endothelial cells for

rapid release upon detection of an inflammatory stimulus.7 Binding of the Shiga toxin to its

receptor, globotriaosylceramide (Gb3), concentrated in the renal microvasculature, may directly

60

stimulate Weibel-Palade body exocytosis, a hypothesis in keeping with the finding that other

constituents of Weibel-Palade bodies, such as von Willebrand Factor (vWF) are released after

Shiga toxin exposure in vitro.19

Alternatively, Shiga toxin may act indirectly through a mediator

such as VEGF, known to stimulate Weibel-Palade body release.362

In contrast, the decline in serum Ang-1 levels may reflect Shiga toxin-induced injury to the

pericytes, or frank thrombocytopenia in the case of children with HUS. However, the Ang-1

concentration declines in the pre-HUS phase of illness, before the onset of thrombocytopenia,

making the latter an unlikely mechanism.

Previous work by Chandler et al has shown that children in the pre-HUS phase of illness have

normal blood cell counts, yet show evidence of early vascular injury with levels of prothrombin

fragment 1+2, tissue plasminogen activator (tPA) antigen, tPA-plasminogen-activator inhibitor

type 1 complex, and D-dimer, that are significantly higher than those in children with

uncomplicated infection.18

Our work supports the concept that subclinical endothelial

dysfunction precedes the development of overt HUS and can be detected before the classically

recognized markers of disease.

Serum or plasma levels of Ang-1/2 have also been proposed as clinically informative

biomarkers in several other important clinical syndromes, including: sepsis, cerebral malaria,

severe bacterial infection in children.127, 129, 130, 305

Although previous studies have investigated

the role of markers of inflammation, such as procalcitonin and interleukin-6, in established HUS,

our study is unique in that it demonstrates the potential utility of a biomarker to detect impending

HUS in the setting of E. coli O157:H7 infection, a stage of illness for which no diagnostic tests

currently exist.363, 364

Although the degree of overlap between the groups precludes the use of

Ang-1 as a definitive single predictor of HUS, one or both of the angiopoietins might be

combined with other early markers of endothelial cell dysfunction, such as those described by

Chandler et al, to derive useful discriminatory or predictive classification rules. Our study

therefore provides a rationale for prospective validation of circulating angiopoietins as outcome-

specific biomarkers for children with E. coli O157:H7 infection. If the utility of one or more

biomarkers were to be confirmed clinically, it might provide the opportunity to initiate early

aggressive intervention, particularly fluid resuscitation, which has been shown to limit morbidity

in children with E. coli O157:H7 infection who ultimately progress to HUS.365

61

There are some important limitations to our study. We chose to investigate E. coli O157:H7-

associated HUS because it remains the most common serotype, and prior to the European

outbreak, was associated with the most severe disease and sequelae.359

Although the clinical

syndrome of HUS is largely the same regardless of serotype, we cannot conclusively state that

our findings apply to illness caused by non-O157 serotypes. In addition, we cannot exclude the

impact of renal failure itself on angiopoietin dysregulation in HUS. However, renal function was

intact in children in the pre-HUS phase of illness, and therefore, this limitation should not detract

from the finding that vascular injury precedes the overt clinical syndrome of HUS, nor from the

utility of Ang-1 to discriminate between children with uncomplicated infection and those in the

pre-HUS phase of illness. Finally, although we cannot exclude a contribution of platelet-derived

Ang-1, due to ex vivo activation of platelets after venipuncture, to the high serum Ang-1

concentrations found in individuals with uncomplicated infection, we would expect the degree of

ex vivo activation to be the same between the children with uncomplicated infection and those in

the pre-HUS phase of illness. Therefore, it is unlikely that this particular limitation influenced

the relative difference in serum Ang-1 concentration between these two groups.

In conclusion, this study is the first to demonstrate angiopoietin dysregulation in E. coli O157:H7

infection, in which the degree of dysregulation is associated with disease severity and predictive

of disease progression upon presentation with bloody diarrhea. If these observations are

confirmed in prospective clinical studies, peripheral blood levels of Ang-1/2 may be clinically

useful to identify children with E. coli O157:H7 infection at high risk of progression to HUS.

62

Chapter 5 Prognostic Biomarkers for Sepsis in a Predominantly HIV-infected

Population in a Resource-limited Setting

5.1 Introduction

Sepsis is a significant cause of infection-related mortality in high-income countries, and,

although data are limited, an emerging and increasingly recognized issue in low- to middle-

income countries.366

In both settings, careful and directed triage of limited resources and

personnel is required, yet at present there is no well-defined, rapid, and specific test to identify

those patients at highest risk of mortality, and therefore most in need of aggressive therapeutic

intervention.

The clinical manifestations of sepsis occur because of an exaggerated response - both

immunologic and vascular - to an infection, typically bacterial. Angiopoietin-1 and -2, sTie-2,

procalcitonin, C-reactive protein, sTREM-1, Chi3L1, IP-10, PF4, vWF, ICAM-1, VEGF, and

sFlt-1 have each been proposed as individual prognostic biomarkers in sepsis. However, as

described in Chapter 1, no single biomarker has been demonstrated to have adequate sensitivity

and specificity in all populations tested, and controversy persists in the literature surrounding

each candidate molecule. There are few studies in which the prognostic utility of these markers

has been explored in combination, and even fewer in resource-limited settings.129, 305

Therefore, the aim of this study is twofold: 1) to determine whether a combination of

biomarkers, representing activation of both the immunologic and vascular pathways, can predict

mortality from an episode of sepsis in a predominantly HIV-infected population in a resource-

limited setting, and 2) to explore the utility of each of these markers, or a combination thereof, to

predict the etiology of an episode of sepsis in the same setting.

5.2 Methods

Study population and enrolment criteria

Beginning in 2009, consecutive patients who presented on weekdays to the Emergency

Departments of two large referral hospitals in Uganda (Mulago Hospital in Kampala and Masaka

Regional Referral Hospital in Masaka) were evaluated for trial enrolment. Inclusion criteria were

63

defined as 2 of 4 criteria for SIRS or thermal dysregulation alone, in combination with systolic

blood pressure ≤ 100 mmHg, suspected infection as per the admitting physician, and an elevated

plasma lactate level (>2.5 mmol/L) or low Karnofsky Performance score (≤ 40). Use of the

Karnofsky Performance Score, a well-described and commonly cited morbidity assessment tool,

for this indication had been investigated previously and found to correlate well with mortality in

Ugandan patients with sepsis.367

Exclusion criteria included: age < 18 years, pregnancy,

gynecologic or surgical presenting illness, acute stroke, and gastrointestinal hemorrhage.

Informed consent was obtained from the patient and her/his relative(s), and demographic,

clinical, biochemical, microbiologic, and outcome data were collected. Resources available in

each hospital for the diagnosis and treatment of infection were previously outlined by

Jacob et al. 367

In total, 336 patients were included in this study, of whom 170 (50.6%) were female. The median

age at enrolment was 34 years (range: 18 - 73 years). 282 patients (84%) had HIV disease, and

amongst those, the median CD4+ T cell count was 45 cells/μL (range 1 - 746 cells/μL). Previous

studies in this population indicated that only 12% of the patients infected with HIV were actively

receiving antiretroviral therapy.367

Fourteen percent of patients had positive malaria smears. At

enrolment, the median lactate level was 3.9 mmol/L (range 1 - 16.6 mmol/L) and the median

Karnofsky Performance score was 40 (range 10 - 80). The median white blood cell count was 4.8

x 109 cells/L (range 0.1 - 82 x 10

9 cells/L).

In-hospital mortality data was available for 332 of 336 patients, of whom 89 of 332 (27%) died.

Twenty-eight day mortality data was available for 306 of 336 patients, of whom 119 (39%) died.

In a univariate analysis of positive microbiologic results (any mycobacterial culture, aerobic

blood culture, or malaria smear), only a positive mycobacterial culture was significantly

associated with a risk of death (data not shown).

The candidate was not involved in this aspect of the study.

Biomarker Assays

Plasma was obtained at study enrolment, divided into aliquots, and stored at -70oC until use.

Care was taken to maintain the frozen samples during prolonged transportation and shipping.

Plasma concentrations of all biomarkers were measured by ELISA (R&D Systems, Minneapolis

64

MN except where indicated) as per the manufacturer‟s instructions. The technical upper limits of

detection for the assays were 250 ng/mL for vWF, 20 000 pg/mL for Ang-1, sTie-2, and sFlt-1,

7000 pg/mL for Ang-2, 5000 pg/mL for PF4, and 4000 pg/mL for C-reactive protein, sTREM-1,

Chi3L1, IP-10, ICAM-1, and VEGF, yielding effective upper limits of detection of 500 μg/mL

for vWF, 80 000 pg/mL for Ang-1, 400 000 pg/mL for sTie-2, 40 000 pg/mL for sFlt-1, 28 000

pg/mL for Ang-2, 50 ng/mL for PF4, 400 g/mL for CRP, 4000 pg/mL for sTREM-1, 2 ng/mL

for Chi3L1, 8000 pg/mL for IP-10, 4 ng/mL for ICAM-1, and 8000 pg/mL for VEGF, for the

dilutions employed in the analysis. Lower limits of detection were as follows: 244 pg/mL for

vWF, 19.5 pg/mL for Ang-1, sTie-2, and sFlt-1, 6.84 pg/mL for Ang-2, 4.88 pg/mL for PF4, and

3.91 pg/mL for CRP, sTREM-1, Chi3L1, IP-10, ICAM-1, and VEGF. Plasma concentrations of

procalcitonin were measured by ELISA (RayBiotech Inc, Norcross GA) as per the

manufacturer‟s instructions, with a technical upper limit of detection of 60 000 pg/mL, an

effective upper limit of detection of 300 000 pg/mL, and a lower limit of detection of 58.6

pg/mL.

All assays were performed by an investigator blinded to clinical and microbiologic data, as well

as outcome, and each sample was tested in duplicate.

Statistical Analyses

Statistical analysis was performed using GraphPad Prism v.4.03 (San Diego CA) and SPSS

v.16.0 (IBM, Chicago IL). Median plasma levels of all biomarkers upon enrolment in survivors

and non-survivors were compared using the Mann-Whitney test. Receiver operating

characteristic (ROC) curves were generated for each biomarker for the prediction of in-hospital

and 28-day mortality, as well as positive mycobacterial or aerobic bacterial cultures, with the

null hypothesis that the area under the curve (AUC) equals 0.5. Classification and Regression

Tree (CaRT) analysis was performed to predict the outcome (survival versus death) of an episode

of sepsis or the likelihood of a laboratory-confirmed microbiologic diagnosis by a combination

of clinical factors (Karnofsky Performance score) and host biomarkers measured upon first

presentation to hospital. CaRT analysis has been used extensively in biomarker studies in cancer,

cardiology, and infectious diseases, among other disciplines, and uses recursive partitioning to

derive cut-points for the independent variables that are then combined to create an algorithm that

can be used to classify patients according to the dependent variable.305, 368, 369

All biomarkers that

65

discriminated between survivors and non-survivors for each of the primary (in-hospital) and

secondary (28-day mortality, positive aerobic culture, and positive mycobacterial culture)

endpoints on univariate analysis were entered into each combinatorial analysis. The cost of

misclassifying an eventual death or positive culture was considered to be 5 times that of

misclassifying survival or a negative culture. CaRT analysis was chosen for this study because it

yields a clinically relevant and easily followed algorithm, and because it makes no assumptions

about the distribution or completeness of the data.370

Because over-fitting, the creation of an

excessively complex tree that does not generalize well, is a potential limitation of CaRT analysis,

the following criteria for tree construction were specified: minimum 10 cases for a parent node

and 5 for a child node, customized prior probabilities based on a prospective observational study

undertaken at the same sites,367

maximum depth of 3 nodes, and pruning to reduce overfitting.

The cut-points as selected by the analysis for each biomarker are indicated between parent and

child nodes, while the predicted classification of patients in each terminal node (no further

branching) is highlighted.

5.3 Results

In this large cohort of Ugandan patients presenting to the Emergency Department with sepsis,

multiple biomarkers were significantly associated with in-hospital mortality (Figure 6). In

particular, angiopoietin dysregulation (an increase in Ang-2 and the Ang-2:Ang-1 ratio) was

more severe in those patients who later died during hospital admission. The median plasma Ang-

2 level was 1927 pg/mL (IQR: 1073 - 5494 pg/mL) in non-survivors and significantly lower at

1366 pg/mL (IQR: 748 - 3122 pg/mL) in survivors, P < 0.01, while the median Ang-2:Ang-1

ratio was 0.40 (IQR: 0.17 - 1.9) in non-survivors and only 0.23 (IQR: 0.074 - 0.84) in survivors,

P < 0.01. In addition, biomarkers representing both the inflammatory (procalcitonin, sTREM-1,

Chi3L1, and IP-10) and vascular (vWF, ICAM-1, and sFlt-1) pathways were each significantly

different in those patients who survived to hospital discharge versus those who did not. The

median plasma procalcitonin level was significantly elevated in non-survivors (11 104 pg/mL,

IQR: 2666 - 34 475 pg/mL) as compared to survivors (6425 pg/mL, IQR: 815 - 23 411 pg/mL),

P < 0.05, as were the median concentrations of sTREM-1 (247 pg/mL, [IQR: 107 - 446 pg/mL]

versus 124 pg/mL, [IQR: 83 - 223 pg/mL], P < 0.001), Chi3L1 (813 ng/mL, [IQR: 253 - 1589

ng/mL] versus 359 ng/mL, [IQR: 178 - 827 ng/mL], P < 0.001), and IP-10 (771 pg/mL, [IQR:

309 - 1273 pg/mL] versus 540 pg/mL, [IQR: 256 - 942 pg/mL], P < 0.01), all in non-survivors

66

versus survivors, respectively. Amongst markers of endothelial or vascular activation, the

median plasma concentration of vWF was significantly higher in patients who died prior to

hospital discharge than in those who survived (265 ng/mL, [IQR: 188 - 394 ng/mL] versus 186

ng/mL, [IQR: 98 - 297 ng/mL], P < 0.001), as was also true of the median plasma concentrations

of ICAM-1 (564 ng/mL, [IQR: 357 - 929 ng/mL] versus 439 ng/mL, [IQR: 303 - 608 ng/mL], P

< 0.001), and sFlt-1 (1133 pg/mL, [IQR: 39 - 3051 pg/mL] versus 319 pg/mL, [IQR: 39 - 1596

pg/mL], P < 0.01).

ROC curves constructed for each of these biomarkers indicated that all nine indices were

predictive of in-hospital morality, albeit to differing degrees: Ang-2 (AUC: 0.597, 95% CI: 0.529

- 0.665; P < 0.01), the Ang-2:Ang-1 ratio (AUC: 0.611, 95% CI: 0.544 - 0.677; P < 0.01),

procalcitonin (AUC: 0.571, 95% CI: 0.503 - 0.639, P < 0.05), sTREM-1 (AUC: 0.668, 95% CI:

0.599 - 0.738, P < 0.001), Chi3L1 (AUC: 0.634, 95% CI: 0.564 - 0.705, P < 0.001), IP-10

(AUC: 0.610, 95% CI: 0.541 - 0.680, P < 0.01), vWF (AUC: 0.645, 95% CI: 0.580 - 0.710, P <

0.001), sICAM-1 (AUC: 0.634, 95% CI: 0.564 - 0.705, P < 0.001), and sFlt-1 (AUC: 0.600, 95%

CI: 0.529 - 0.671, P < 0.01) (data not shown). ROC curves for the other biomarkers tested did

not meet statistical significance (data not shown). Although the ROC curve for sTREM-1

demonstrated the greatest AUC, a plasma sTREM-1 cut-off value of 164.7 pg/mL yielded only

62% sensitivity and specificity for the prediction of in-hospital mortality at the onset of an

episode of sepsis in this population.

Both plasma lactate and the clinical Karnofsky Performance score were significantly different

between those patients who died prior to hospital discharge and those who survived: 4.4 mmol/L

(IQR: 3.3 - 5.7 mmol/L) versus 3.8 mmol/L (IQR: 3.0 - 4.6 mmol/L), P < 0.001, and 30 (IQR: 30

- 40) versus 40 (IQR: 30 - 50), P < 0.001, respectively (data not shown). ROC curves for both

were significantly predictive of in-hospital mortality, however, plasma lactate performed slightly

worse (AUC: 0.609, 95% CI: 0.543 - 0.676, P < 0.001) than many of the newer biomarkers

discussed above, and Karnofsky Performance score (AUC: 0.668, 95% CI: 0.609 - 0.728, P <

0.001) approximately equally (data not shown).

67

68

Figure 6. Plasma Biomarkers Associated with In-hospital Mortality in Ugandan patients with

sepsis. Absolute and median concentrations of Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-2),

Ang-2:Ang-1 ratio, soluble Tie-2 receptor (sTie-2), C-reactive protein (CRP), Procalcitonin

(PCT), soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1), Chitinase-3-like

protein 1 (Chi3L1), Interferon-γ-inducible Protein-10 kDa (IP-10), Platelet Factor 4 (PF4), von

Willebrand Factor (vWF), soluble Intercellular Adhesion Molecule-1 (sICAM-1), Vascular

Endothelial Growth Factor (VEGF), and soluble Fms-like tyrosine kinase-1 (sFlt-1) at the time

of hospital admission in patients with sepsis, divided according to in-hospital mortality . * P <

0.05; ** P < 0.01; *** P < 0.001.

Multiple biomarkers were also significantly associated with 28-day mortality (Figure 7). With

one exception, these markers were identical to the subset that predicted in-hospital mortality.

Again, angiopoietin dysregulation at hospital admission was significantly greater in those

patients who died in the subsequent 28 days than in those who survived. The median plasma

concentration of Ang-2 was 1927 pg/mL (IQR: 871 - 5427 pg/mL) in non-survivors and

significantly lower at 1381 pg/mL (IQR: 758 - 3077 pg/mL) in survivors, P < 0.05, while the

median Ang-2:Ang-1 ratio was 0.49 (IQR: 0.17 - 1.8) in non-survivors and 0.21 (IQR: 0.072 -

0.68) in survivors, P < 0.001. Again mirroring the results found with in-hospital mortality,

biomarkers representing both the inflammatory (sTREM-1, Chi3L1, and IP-10) and vascular

(PF4, vWF, sICAM-1, and sFlt-1) pathways were each significantly different in patients who

survived for at least 28 days after hospital admission versus those who did not. The median

plasma concentrations of sTREM-1 (178 pg/mL, [IQR: 98 - 385 pg/mL] versus 127 pg/mL,

[IQR: 85 - 225 pg/mL], P < 0.01), Chi3L1 (749 ng/mL, [IQR: 228 - 1309 ng/mL] versus 352

ng/mL, [IQR: 154 - 723 ng/mL], P < 0.001), and IP-10 (771 pg/mL, [IQR: 300 - 1319 pg/mL]

versus 504 pg/mL, [IQR: 259 - 870 pg/mL], P < 0.001), all differed significantly amongst non-

survivors versus survivors at 28 days, respectively. Similarly, the median plasma concentrations

of vWF (260 ng/mL, [IQR: 175 - 387 ng/mL] versus 172 ng/mL, [IQR: 89 - 281 ng/mL], P <

0.001), ICAM-1 (536 ng/mL, [IQR: 357 - 880 ng/mL] versus 425 ng/mL, [IQR: 305 - 601

ng/mL], P < 0.001), and sFlt-1 (1159 pg/mL, [IQR: 39 - 3051 pg/mL] versus 125 pg/mL, [IQR:

39 - 1224 pg/mL], P < 0.001) were all significantly higher at hospital admission in those patients

69

who died within 28 days than in those who survived (non-survivor versus survivor, respectively,

for each biomarker). In contrast, the median plasma concentration of PF4 at hospital admission

demonstrated the opposite pattern and was significantly lower in non-survivors (944 ng/mL,

IQR: 342 - 2150 ng/mL) than survivors (1244 ng/mL, IQR: 942 - 3535 ng/mL), P < 0.05.

Notably, the median plasma procalcitonin level upon presentation to hospital did not differ

significantly amongst survivors and non-survivors at 28 days as it did amongst survivors and

non-survivors at hospital discharge, while the opposite was true of the median plasma PF4

concentration.

Each of these nine biomarkers discriminated between survivors and non-survivors at 28 days

according to the area under the ROC curve: Ang-2 (AUC: 0.579, 95% CI: 0.513 - 0.644; P <

0.05), the Ang-2:Ang-1 ratio (AUC: 0.619, 95% CI: 0.554 - 0.683; P < 0.001), sTREM-1 (AUC:

0.607, 95% CI: 0.540 - 0.675, P < 0.01), Chi3L1 (AUC: 0.639, 95% CI: 0.574 - 0.703, P <

0.001), IP-10 (AUC: 0.619, 95% CI: 0.554 - 0.685, P < 0.001), PF4 (AUC: 0.587, 95% CI: 0.521

- 0.652, P < 0.05), vWF (AUC: 0.663, 95% CI: 0.602 - 0.724, P < 0.001), sICAM-1 (AUC:

0.627, 95% CI: 0.562 - 0.693, P < 0.001), and sFlt-1 (AUC: 0.631, 95% CI: 0.564 - 0.698, P <

0.001) (data not shown). However, while statistically significant, none of these indices were

sufficiently sensitive and specific to be clinically useful as the definitive and sole prognostic

indicator in patients presenting to the Emergency Department with sepsis.

Again, both lactate (4.1 mmol/L [IQR: 3.3 - 5.6 mmol/L] versus 3.8 mmol/L [IQR: 3.0 - 4.6

mmol/L], P < 0.01 in non-survivors and survivors, respectively) and Karnofsky Performance

score (30 [IQR: 30 - 40] versus 40 [IQR: 30 - 50], P < 0.001 in non-survivors and survivors,

respectively) differed according to 28-day mortality. ROC curves indicated that both were

predictive of 28-day mortality - lactate AUC: 0.578, 95% CI: 0.518 - 0.639, P < 0.01, and

Karnofsky Performance score AUC: 0.656, 95% CI: 0.601 - 0.712, P < 0.001 - but neither was

sufficiently predictive for use as the sole predictive marker in a clinical setting.

70

71

Figure 7. Plasma Biomarkers Associated with 28-day Mortality in Ugandan patients with sepsis.

Absolute and median concentrations of Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-2), Ang-

2:Ang-1 ratio, soluble Tie-2 receptor (sTie-2), C-reactive protein (CRP), Procalcitonin (PCT),

soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1), Chitinase-3-like protein

1 (Chi3L1), Interferon-γ-inducible Protein-10 kDa (IP-10), Platelet Factor 4 (PF4), von

Willebrand Factor (vWF), soluble Intercellular Adhesion Molecule-1 (sICAM-1), Vascular

Endothelial Growth Factor (VEGF), and soluble Fms-like tyrosine kinase-1 (sFlt-1) at the time

of hospital admission in patients with sepsis, divided according to 28-day mortality . * P < 0.05;

** P < 0.01; *** P < 0.001.

CaRT analysis was undertaken to determine whether a combination of biomarkers might be more

clinically useful for the prediction of mortality at the time of presentation to the Emergency

Department. With in-hospital mortality defined as the dependent variable and those biomarkers

that had effectively discriminated between survivors and non-survivors in univariate analyses

defined as independent variables, CaRT analysis identified a model incorporating sTREM-1 (cut-

off value > 246 pg/mL), the Karnofsky Performance score (cut-off value ≤ 30), and Ang-2 (cut-

off value > 1404 pg/mL) that predicted in-hospital mortality with a sensitivity of 97% (Figure

8A). Sensitivity was not appreciably increased when the model was expanded to incorporate

additional biomarkers, nor when either of the standard predictive measures (Karnofsky

Performance score or lactate) were specified a priori as the first variable in the model (data not

shown). Furthermore, a model incorporating only the Karnofsky Performance score and lactate

yielded a higher misclassification rate (data not shown). For the prediction of 28-day mortality,

CaRT analysis yielded a model that incorporated vWF (cut-off value 194 ng/mL), the Ang-

2:Ang-1 ratio (cut-off value > 0.23), and Chi3L1 (cut-off value > 84 ng/mL), and identified non-

survivors with 93% sensitivity (Figure 8B).

72

Figure 8. Classification and Regression Tree (CaRT) analysis of biomarker combinations to predict mortality following an episode of

sepsis in a Ugandan population. A. sTREM-1, Ang-2, and the Karnofsky Performance score predict in-hospital mortality: sensitivity

97%, specificity 38%, misclassification rate 26%, standard error 2.4%. B. vWF, Ang-2:Ang-1, and Chi3L1 predict 28-day mortality:

sensitivity 93%, specificity 38%, misclassification rate 20%, standard error 2.3%.

A. B.

73

In addition to prognostic indicators, we sought to identify those biomarkers that might be

associated with specific microbial etiologies. Of 226 patients who had blood cultures drawn, 37

(16%) were positive. This is in keeping with the incidence of positive cultures in previous studies

in this population.367

The median plasma concentrations for 5 of 12 biomarkers tested were

significantly different at hospital admission in those patients with positive aerobic blood cultures

than in those patients whose blood cultures remained negative: CRP (178 μg/mL, [IQR: 98 - 385

μg/mL] versus 127 μg/mL, [IQR: 85 - 225 μg/mL], P < 0.01), PCT (749 pg/mL, [IQR: 228 -

1309 pg/mL] versus 352 pg/mL, [IQR: 154 - 723 pg/mL], P < 0.001), sTREM-1 (771 pg/mL,

[IQR: 300 - 1319 pg/mL] versus 504 pg/mL, [IQR: 259 - 870 pg/mL], P < 0.001), Chi3L1 (851

ng/mL, [IQR: 441 - 1992 ng/mL] versus 295 ng/mL, [IQR: 151 - 827 ng/mL], P < 0.001), and

sICAM-1 (500 ng/mL, [IQR: 386 ng/mL - 695 ng/mL], versus 407 ng/mL, [IQR: 287 - 603

ng/mL], P < 0.05), in patients with positive and negative aerobic blood cultures, respectively

(Figure 9A).

Each of these five biomarkers discriminated between those patients with positive blood cultures

and those without as per the area under the ROC curves for each (Figure 9B): CRP (AUC: 0.614,

95% CI: 0.514 - 0.714, P < 0.05), PCT (AUC: 0.757, 95% CI: 0.667 - 0.847, P < 0.001),

sTREM-1 (AUC: 0.623, 95% CI: 0.523 - 0.723, P < 0.05), Chi3L1 (AUC: 0.730, 95% CI: 0.650

- 0.809, P < 0.001), and sICAM-1 (AUC: 0.609, 95% CI: 0.518 - 0.700, P < 0.05). PCT and

Chi3L1 were the best predictors of positive aerobic blood cultures from amongst the five

significant biomarkers, and this was confirmed in a CaRT analysis, in which these two markers,

with the addition of sTREM-1, predicted positive blood cultures with a specificity of 86%

(Figure 9C).

Neither lactate nor Karnofsky performance score differed amongst patients with and without

positive aerobic blood cultures (data not shown).

74

A. C. B.

75

Figure 9. Plasma biomarkers associated with positive aerobic blood cultures. A. Individual and

median plasma biomarker concentrations in patients with positive and negative aerobic blood

cultures. * P < 0.05; ** P < 0.01; *** P < 0.001. C-reactive protein (CRP), Procalcitonin (PCT),

soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1), Chitinase-3-like protein

1 (Chi3L1), and soluble Intercellular Adhesion Molecule-1 (sICAM-1). B. ROC curves for each

biomarker that significantly predicted a positive aerobic blood culture. C. CaRT analysis of a

biomarker combination to predict positive aerobic blood cultures at the time of hospital

admission.

A similar analysis was performed to examine biomarkers predictive of mycobacterial infection.

Results were positive in 71 of 243 patients (29%) from whom mycobacterial cultures were taken,

consistent with the findings of previous studies in the same population.367

In patients presenting

to the Emergency Department with sepsis, all biomarkers tested were significantly different

between those with and without mycobacterial infection except for sTie-2, sFlt-1, and VEGF

(Table 2). Furthermore, ROC curves confirmed that each of these markers were predictive of

positive mycobacterial culture results, albeit to varying degrees. sTREM-1 and IP-10 had the

highest and second highest AUCs by ROC curve analysis, respectively, and CaRT analysis

included both in a model that predicted positive mycobacterial culture results with 79%

sensitivity and 76% specificity (Figure 10). Although plasma lactate was significantly different

between those with and without mycobacterial infection, the associated AUC from the ROC

curve analysis was lower than that of most of the other biomarkers tested (data not shown).

Karnofsky performance score was not significantly different between patients with and without

mycobacterial infection (data not shown).

Interestingly, the most commonly used indicator of infection of any type, leukocyte count, was

not significantly different at hospital admission in those patients with and without either

bacteremia or mycobacterial infection (data not shown). Furthermore, although the white blood

cell count was slightly lower in non-survivors than survivors, and although this difference did

reach statistical significance (P < 0.05), admission white blood cell count was very poorly

76

discriminative of both in-hospital and 28-day mortality, and was outperformed by all other

statistically significant biomarkers (data not shown).

77

Table 2. Significant predictors of mycobacterial infection in Ugandan adults with sepsis. Angiopoietin-1 (Ang-1), Angiopoietin-2

(Ang-2), Ang-2:Ang-1 ratio, C-reactive protein (CRP), Procalcitonin (PCT), soluble Triggering Receptor Expressed on Myeloid cells-1

(sTREM-1), Chitinase-3-like protein 1 (Chi3L1), Interferon-γ-inducible Protein-10 kDa (IP-10), Platelet Factor 4 (PF4), von Willebrand

Factor (vWF), and soluble Intercellular Adhesion Molecule-1 (sICAM-1).

Median Plasma Concentration (IQR) Area under the ROC curve (95% CI)

Biomarker Positive

Mycobacterial culture

Negative

Mycobacterial culture

P P

Ang-1 (pg/mL) 3835 (1425 - 11 032) 5598 (2529 - 13 549) < 0.05 0.588 (0.522 - 0.665) < 0.05

Ang-2 (pg/mL) 2043 (1099 - 5078) 1382 (778 - 3113) < 0.01 0.606 (0.533 - 0.680) < 0.01

Ang-2:Ang-1 0.660 (0.152 - 1.92) 0.225 (0.078 - 0.770) < 0.001 0.641 (0.569 - 0.713) < 0.001

CRP (μg/mL) 134 (78.3 - 221) 84.9 (35.2 - 169) < 0.001 0.631 (0.561 - 0.701) < 0.001

PCT (pg/mL) 13 709 (4858 - 44 738) 5605 (440 - 20 800) < 0.001 0.635 (0.563 - 0.707) < 0.001

sTREM-1 (pg/mL) 294 (160 - 458) 117 (82 - 223) < 0.001 0.777 (0.720 - 0.834) < 0.001

Chi3L1 (ng/mL) 726 (343 - 1309) 348 (172 - 932) < 0.001 0.657 (0.588 - 0.725) < 0.001

IP-10 (pg/mL) 901 (670 - 1342) 444 (216 - 884) < 0.001 0.754 (0.697 - 0.811) < 0.001

PF4 (ng/mL) 794 (317 - 2139) 1180 (509 - 3271) < 0.05 0.587 (0.512 - 0.662) < 0.05

vWF (μg/mL) 249 (177 - 367) 192 (91 - 302) < 0.01 0.625 (0.559 - 0.691) < 0.01

sICAM-1 (ng/mL) 592 (386 - 815) 432 (294 - 631) < 0.001 0.647 (0.577 - 0.716) < 0.001

78

Figure 10. Classification and Regression Tree (CaRT) analysis for the prediction of positive

mycobacterial cultures in Ugandan adults with sepsis. Interferon-γ-inducible protein-10 kDa (IP-

10), soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1). Sensitivity 79%,

specificity 76%.

79

5.4 Discussion/Conclusion

The current study of 336 Ugandan patients with sepsis is, to our knowledge, the largest study

published to date that has investigated the role of prognostic biomarkers in a predominantly HIV-

infected population in a resource-limited setting. As a result, a number of unique conclusions can

be drawn.

First, angiopoietin dysregulation is a prominent feature of sepsis in this population, and both the

plasma Ang-2 concentration and the Ang-2:Ang-1 ratio at the time of hospital admission predict

in-hospital mortality and mortality at 28 days. Angiopoietin dysregulation, manifested primarily

by increased Ang-2 resulting in an increased Ang-2:Ang-1 ratio, has been found in septic

patients in high-income countries, and in these settings has been found to correlate with severity

of illness, pulmonary edema, ALI/ARDS, multi-organ dysfunction, and mortality.9, 120-123, 127

The

increase in Ang-2 in sepsis may reflect Weibel-Palade body exocytosis in response to reduced

endothelial nitric oxide bioavailability, increased sphingosine-1-phosphate, or perturbation of

another as yet unidentified factor.5, 53, 123

Since many proposed biomarkers of infection have shown wide variation in performance

characteristics when studied in different patient populations, our study is significant in that it

confirms that the association between angiopoietin dysregulation and mortality in sepsis is

present in a predominantly HIV-infected, significantly immunosuppressed population in which

the median CD4+ T cell count was 45 cells/L. Mankhambo et al conducted the only other

investigation of angiopoietin dysregulation in a similar population: a study of 293 Malawian

children with meningitis or pneumonia of whom 53% were HIV-infected and in whom a low

Ang-1 concentration was associated with mortality in a multivariate analysis.129

Our study is also unique in that, in addition to Ang-2 and the Ang-2:Ang-1 ratio, sTREM-1,

Chi3L1, IP-10, vWF, sICAM-1, and sFlt-1 were all predictive both of in-hospital and 28-day

mortality in this septic population, while PCT and PF4 were predictive of in-hospital and 28-day

mortality, respectively. This is the first study to document an association between mortality in

sepsis and Chi3L1 or PF4, and the second to document the same association with IP-10. Chi3L1

has only been examined once before in sepsis, in a small study of 45 patients in which it was

associated with illness severity but not mortality,302

although elevated Chi3L1 has been

associated with mortality following pneumococcal bacteremia.300

PF4 had previously been

80

studied in 10 patients with sepsis, but no attempt was made to correlate increased circulating

levels of PF4 with mortality.314

Finally, IP-10 had previously been studied only in 16 patients

with sepsis, but despite the small sample size elevated levels were nonetheless found to correlate

with mortality.303

Interestingly, all of the biomarkers studied here were higher in non-survivors

than survivors, with the exception of PF4, which demonstrated the opposite pattern. PF4 is

known to facilitate generation of the anticoagulant molecule activated protein C, and, in keeping

with our findings, increased levels of PF4 have been shown to correlate with survival in animal

models of sepsis.312, 313

While the literature is generally in agreement that sTREM-1 can predict prognosis in sepsis,

there is some controversy regard the direction of change. Our results coincide with those of

Zhang et al and Dimopoulou et al, both of whom described higher levels of sTREM-1 in non-

survivors, but conflict with those of Gibot et al, who describe lower sTREM-1 in non-

survivors.260, 289, 290

It is possible that the discrepancy reflects the inherent paradox of sepsis,

namely that the inflammatory response can be both helpful and harmful. sTREM-1 amplifies the

innate immune response induced by pathogens detected by pattern recognition factors such as

TLR4.371

If this response is controlled, carefully regulated, and directed, the end result is

pathogen clearance and minimal host morbidity. If, however, this response is exaggerated,

excessive, and indiscriminate, the end result is sepsis and potentially significant morbidity.

Therefore, the discrepancy may simply reflect the fact that an intermediate level of sTREM-1 is

required for an appropriate response, and levels that are either too high or too low are

detrimental.

PCT, vWF, sICAM-1, and sFlt-1 have all been associated with mortality and/or disease severity

in other studies of septic patients, and in this regard our results confirm previous findings.34, 120,

254-263, 267, 268, 273-275, 318, 326, 331

Of note, we did not detect an association between Ang-1, sTie-2, VEGF, or CRP and mortality in

this population of adult patients with sepsis. Ang-1 has not been identified as a significant

predictor of mortality in sepsis as consistently as have Ang-2 or the Ang-2:Ang-1 ratio.127, 128

Furthermore, Ang-1 has been shown to fluctuate widely in individual patients over a 24-hour

period,118

and therefore may have limited clinical reproducibility and utility. With respect to the

other candidate biomarkers, our results are in keeping with the published literature in which

81

neither sTie-2, nor VEGF, nor CRP have been consistently and reliably associated with mortality

in sepsis.128, 131, 266, 272, 334

Despite the number of biomarkers that were positively associated with mortality in this cohort,

none demonstrated sufficient sensitivity or specificity to be employed clinically as the sole

indicator of prognosis. Therefore, we exploited Classification and Regression Tree (CaRT)

analysis to determine whether combinations of biomarkers were more predictive of mortality in

this population than any single marker alone. Combining biomarkers has previously been shown

to enhance their prognostic utility in both sepsis and malaria,127, 305

and likewise the

combinations illustrated in Figures 6 and 7 predict death with better sensitivity than any

individual marker alone. The predictive models for both in-hospital and 28-day mortality are

representative of the underlying pathophysiology of sepsis and incorporate markers from both

the inflammatory and vascular pathways. To maximize clinical utility and feasibility in our

patient population and resource-limited setting, this analysis was performed with the prespecified

goals of prioritizing sensitivity over specificity and limiting the number of markers included in

each model to three. By identifying patients at highest risk of mortality, this approach may be

useful for triage and directed allocation of limited resources. A recent pan-African survey of

anesthesiologists found that < 2% of African hospitals regularly had access to the resources,

including medications, disposable and permanent equipment, and monitoring and diagnostic

tools, required to implement the Surviving Sepsis Campaign guidelines, now considered the

standard of care for patients with sepsis.372, 373

A convenient, inexpensive, point-of-care test that

could be used at the time of hospital admission to identify those patients at highest risk of

mortality and therefore most likely to benefit from intensive therapy would be of value, and

should be studied prospectively. Rapid tests are currently available for the diagnosis of HIV,

syphilis, and malaria, and new microfluidics-based technology allows all of the steps of a

traditional ELISA to be mimicked on a point-of-care chip, making implementation of biomarker

analysis feasible even in resource-limited settings.374, 375

Rapid identification of the specific microbial etiology responsible for an episode of sepsis would

likewise be a key component of any algorithm designed to maximize appropriate resource

allocation. To that end, we confirmed that procalcitonin has reasonable specificity for bacterial

infection despite the high prevalence of co-infection with HIV, malaria, and mycobacteria in this

population. Interestingly, Chi3L1 had equivalent specificity for bacterial infection in this

82

population, and has not previously been studied for its discriminative ability in this regard.

Although the known association of Chi3L1 with the specific granules of activated neutrophils is

in keeping with the finding of elevated Chi3L1 in bacterial infection, elevated levels of Chi3L1

have also been reported in association with HIV encephalitis and hepatitis C-induced liver

fibrosis.299, 376, 377

Therefore, further study of Chi3L1 is warranted to confirm its specificity for

bacterial infection in septic patients.

In contrast, IP-10 has been well-studied in association with tuberculosis and was also most

predictive of mycobacterial infection in our cohort of Ugandan patients with sepsis. IP-10 acts to

amplify the adaptive immune response: it can be produced by monocytes, macrophages, and

vascular endothelial cells and act as a specific chemotactic signal to draw activated Th1 cells that

express its cognate receptor, CXCR3, to a site of infection, and in turn, expression of IP-10 is

further upregulated by interferon-γ produced by the infiltrating Th1 cells.378

Consistent with our

findings and its role in the Th1 inflammatory response, elevated levels of IP-10 have been

demonstrated in areas of granulomatous inflammation and pneumonitis in the lungs of patients

with tuberculosis.379

However, it should be noted that elevated IP-10 levels are not

pathognomonic of mycobacterial infection. IP-10 has also been shown to be elevated in, and

crucial to host control of, Toxoplasma gondii and influenza infection.380, 381

If combined in a point-of-care test panel, these biomarkers might facilitate the initiation of rapid,

appropriate, empiric antimicrobial therapy in the Emergency Department, without the time delay

required for standard diagnostic methods. A recent study compared blood culture, bacterial

polymerase chain reaction (PCR), and procalcitonin for the diagnosis of bacteremia, and found

that the average time from testing to the receipt of results was 33 hours for blood cultures, 10

hours for PCR, and 45 minutes for procalcitonin, a significant difference that is not

inconsequential given the well-documented relationship between delay of appropriate antibiotics

and mortality in sepsis.231, 382

Procalcitonin is not a substitute for blood culture since the latter

has a clear advantage in terms of antimicrobial identification and susceptibility testing. However,

procalcitonin and similar markers may be particularly useful in resource-limited settings to

prioritize those patients most likely to benefit from broad-spectrum antibiotics, or, in the case of

IP-10, to identify and appropriately treat those patients likely to have a non-bacterial cause of

sepsis.

83

In summary, our study has demonstrated that a variety of biomarkers from the inflammatory and

vascular pathways are predictive of both in-hospital and 28-day mortality when measured at the

time of hospital presentation in predominantly HIV-infected Ugandan patients with sepsis.

Biomarker combinations warrant further study with the goal of developing rapid, point-of-care

tests to assist Emergency Department physicians in the appropriate allocation of limited

resources, particularly as sepsis becomes increasingly recognized as an important contributor to

morbidity and mortality in low- and middle- income countries.

84

Chapter 6 Discussion

Angiopoietin dysregulation (decreased Ang-1 and increased Ang-2) is present in streptococcal

toxic shock syndrome, E. coli O157:H7-induced hemolytic-uremic syndrome, and sepsis. This

work is the first to identify angiopoietin dysregulation in both invasive group A streptococcal

infection and E. coli O157:H7 infection, and adds both syndromes to the published literature on

infectious diseases, including sepsis, severe bacterial infection, and malaria, in which

angiopoietin dysregulation has been described.119, 128, 133

No association between angiopoietin

dysregulation and exotoxin-mediated diseases has previously been reported, and therefore, these

findings may provide the rationale for future experimental work in this area, particularly as

regards the direct effects of exotoxins on the endothelium in general and on Weibel-Palade body

exocytosis in particular. The impact of streptococcal exotoxins on the release of WPBs or their

contents has not been studied in vitro, and while Shiga toxins have been linked indirectly to

WPB exocytosis (via rapid toxin-mediated stimulation of vWF release), the mechanism for

decreased Ang-1, as found in our study, is not known.383

In patients with sepsis, our findings of angiopoietin dysregulation, and in particular of increased

plasma Ang-2 and an elevated Ang-2:Ang-1 ratio, were in accordance with those reported in

other cohorts of septic patients.9, 120-123, 127

Furthermore, angiopoietin dysregulation performed

better than lactate, a standard marker of tissue hypoperfusion, for the prediction of mortality in

sepsis, confirming results reported by others.124

Our study, however, was unique in that it was

conducted in a resource-limited setting amongst adult patients with sepsis, of whom 84% were

HIV-infected and 29% had mycobacteremia, and as such, is the only study of its kind to

document the association between angiopoietins and sepsis in this patient population. The study

of biomarkers for sepsis in resource-limited settings is crucial prior to clinical implementation, as

patient populations in these settings differ markedly from those enrolled in usual clinical trials.

Mortality rates are often higher than in settings in which protocolized sepsis bundles are

considered standard of care.367

As demonstrated in the cohort under study here, patients in

resource-limited areas of Africa who present with sepsis have a high rate of co-infections,

including HIV, malaria, and mycobacteremia, not seen in the usual sepsis studies conducted in

North America, Australia, and Europe. Co-infections that influence basal levels of circulating

85

inflammatory and vascular mediators are particularly significant as they may accentuate or mask

between-group differences in the biomarker under study, and as a consequence, render those

biomarkers either more or less useful than they appeared in studies conducted in different target

populations. For example, healthy patients from malaria endemic areas are known to exhibit a

different basal cytokine expression profile than that found in patients from resource-limited but

malaria non-endemic areas, 384, 385

and active malaria infection has been shown to abrogate the

clinical utility of both procalcitonin and C-reactive protein to discriminate between bacterial and

viral pneumonia.283

Similarly, HIV infection is known to induce the production of pro-

inflammatory cytokines that with treatment decline towards levels found in healthy

individuals.386

Our study is therefore significant in that it demonstrates the utility of angiopoietin

dysregulation for the prediction of mortality in a patient population with a high rate of untreated,

severe HIV disease in which the median CD4+ T cell count was 45 cells/μL.

In addition, angiopoietin dysregulation was present despite a high rate of mycobacterial infection

in this Ugandan population. Although plasma levels of the angiopoietins have not been studied

previously in patients with mycobacterial disease, pleural fluid Ang-2 levels were found to be

lower in tuberculous pleural effusions than in other types of exudative effusions.387

Furthermore,

angiopoietin dysregulation has not been found in animal models of pulmonary tuberculosis. In

fact, caseating granuloma formation in these animals has been attributed in part to impaired

neovascularization and subsequent hypoxia as a result of relatively low angiopoietin levels.388

Given these findings, it was important that angiopoietin dysregulation be confirmed specifically

in septic patients with a high rate of mycobacterial disease prior to clinical use or further study in

this population.

If added to the initial panel of diagnostic tests in patients with suspected severe bacterial

infection, the angiopoietins might be useful adjuncts to the clinical assessment of prognosis. The

presence of angiopoietin dysregulation can differentiate between those patients with invasive

streptococcal infection and those with streptococcal toxic shock, between those patients with E.

coli O157:H7 infection who will have an uncomplicated course and those who will develop the

hemolytic-uremic syndrome, and between those who will and will not survive after an episode of

sepsis at the time of presentation to the Emergency Department. By providing information about

the functional status of the endothelium, angiopoietin dysregulation acts as a marker of the extent

of the pathophysiologic abnormalities induced by the disease process, and, at least in the case of

86

E. coli O157:H7 infection, is a more sensitive indicator of impending complications than are

traditional laboratory investigations alone. In patients who present to the Emergency Department

with sepsis, angiopoietin dysregulation is associated with an increased risk of both early and late

mortality. The identification of those patients with the greatest degree of angiopoietin

dysregulation, and presumably endothelial dysfunction, may be an important signal for the need

for more aggressive therapeutic intervention. Early, appropriate, supportive therapy is known to

improve outcome in both HUS and sepsis, and the ability to accurately triage or prioritize

patients for the receipt of these interventions would enhance any Emergency Department critical

illness protocol and would be especially crucial in resource-limited settings.231, 232, 365, 366

The

clinical utility of the angiopoietins in multiple different infectious syndromes is a major

advantage over biomarkers with limited or specific indications and argues for their use in the

undifferentiated patient with suspected infection.

In sepsis, the prognostic utility of the angiopoietins, representative of vascular and endothelial

activation, can be enhanced by their use together with other biomarkers representative of

inflammation and immunologic activation. For the prediction of in-hospital mortality, Ang-2 can

be combined with the Karnofsky Performance score (representing clinical assessment) and

sTREM-1 (representing the inflammatory pathway), while for the prediction of 28-day mortality,

the Ang-2:Ang-1 ratio can be combined with vWF and Chi3L1 (the former representing the

vascular pathway and the latter the inflammatory pathway), to create in both cases clinical

prediction pathways capable of identifying those patients at greatest risk of mortality with

sensitivities in excess of 90%. The resulting algorithms are composed of biomarkers from the

two primary pathways perturbed during sepsis and therefore may more accurately represent the

underlying pathophysiology of the sepsis syndrome than might angiopoietin dysregulation alone.

Biomarker combinations have likewise been proposed for use in risk assessment and outcome

prediction during acute coronary syndromes (ACS). As reviewed by Morrow and Braunwald,

ACS results from the maladaptive activation of multiple systems (in a manner strikingly similar

to sepsis), including the inflammatory, thrombotic, and vascular systems, and therefore,

prognostication is best done using an algorithm that incorporates information from each of the

potentially abnormal pathways.389

Testing this theory in the Emergency Department, Birkhahn et

al found that a point-of-care test (POCT) incorporating multiple cardiac biomarkers was equally

accurate for the diagnosis of non-ST elevation myocardial infarction as standard laboratory-

87

based single marker testing, but was significantly faster.390

With regards to infectious syndromes,

biomarker combinations have been proposed for the prediction of mortality amongst Ugandan

children with severe malaria and amongst patients with sepsis in a high-income country,127, 305

and represent the most likely form in which these prognostic biomarkers will be introduced into

clinical settings.

The finding of angiopoietin dysregulation in three seemingly disparate infectious disease

syndromes also suggests the possibility of a shared underlying final common pathway of

endothelial injury or activation. As such, the angiopoietin system becomes an attractive target for

therapeutic intervention, particularly in patients who present with non-specific hypotension and

major organ (such as renal) dysfunction of unclear etiology. In such cases, empiric antibiotic

therapy requires time to take effect, may be initially inappropriate for the causative organism

(particularly in the era of increasingly common multi-drug resistant organisms), or may be

ineffective in the case of toxin-mediated disease. This is perhaps best illustrated by recent

outcome studies in streptococcal toxic shock, in which even with prompt antimicrobial therapy

and surgical debridement where indicated for necrotizing fasciitis, mortality rates remain in

excess of 40%.14

Furthermore, even with protocolized sepsis management guidelines mandating

early goal-directed therapy in the Emergency Department under optimal clinical trial conditions,

in-hospital mortality for patients presenting with sepsis was 30%.391

Amongst those who did not

receive early goal-directed therapy, mortality was 46.5%. Ten years after the advent of early

goal-directed therapy and the subsequent widespread implementation of severe sepsis

resuscitation bundles and the Surviving Sepsis Campaign recommendations, mortality rates from

sepsis remain at 30% and efforts to further improve outcomes have stalled.370, 392, 393

Angiopoietin-based therapies might be useful adjuncts to the current best-practice interventions

to support such patients until the etiology of the presentation is determined and therapy

appropriately targeted.

There are limitations to the body of work presented here. These studies document an association

between angiopoietin dysregulation and each of the syndromes under investigation, and therefore

suggest, but do not prove, that the angiopoietins are involved in the underlying pathogenesis.

Furthermore, the impact of angiopoietin testing on physician decision-making and patient

outcome should be explored in a prospective manner before firm conclusions can be drawn

regarding their true clinical utility. Regardless, the three studies presented here add to the

88

growing literature on angiopoietin dysregulation by a) identifying two syndromes (STSS and

HUS) not previously known to be associated with angiopoietin dysregulation and demonstrating

that dysregulation correlates with disease severity in each and is predictive of clinical outcome in

E. coli O157:H7 infection, and b) confirming that angiopoietin dysregulation is associated with

mortality in sepsis in a predominantly HIV-infected patient population and illustrating that a

combination of biomarkers, including the angiopoietins as markers of vascular dysfunction and

either sTREM-1 or Chi3L1 as markers of inflammation, predicts mortality better than any

individual marker alone.

89

Chapter 7 Conclusion

Angiopoietin dysregulation is present in invasive group A streptococcal infection, E. coli

O157:H7 infection, and sepsis. It is associated with disease severity in streptococcal toxic shock

and is a prognostic biomarker in undifferentiated E. coli O157:H7 infection and sepsis,

suggesting that measurement of the angiopoietins upon patient presentation with a suspected

severe bacterial infection may offer clinically useful information beyond that which can be

obtained from clinical judgement and traditional diagnostic tests.

However, in recent years many biomarkers have been proposed, studied, and ultimately

associated with a variety of different diseases and syndromes, yet have failed to make the

transition to clinical practice. This has prompted some authors to raise concerns regarding the

true clinical utility of biomarkers identified in association studies such as those presented here.394

Vasan has elegantly outlined the ideal characteristics for biomarkers, referring specifically to

biomarkers of cardiovascular disease but with principles equally applicable to biomarkers of

infectious diseases.395

To be useful in clinical practice, biomarkers should be specific,

reproducible, representative in a single measure but correlated with disease over time, applicable

to and studied in multiple different populations, able to provide information not available from

clinical assessment, and amenable to accurate laboratory testing that can be automated to achieve

high throughput and short turnaround time. Based on these criteria, the angiopoietins appear

well-positioned to evolve from experimental to clinical use, and will likely do so first in sepsis,

in which they have been most often studied.

When compared to other biomarkers recently proposed as prognostic indicators in patients with

sepsis, the angiopoietins have several advantages. First, unlike procalcitonin and C-reactive

protein, the angiopoietins have been consistently associated with disease severity in sepsis in

multiple different studies and wide-ranging patient populations, from undifferentiated critically

ill patients in the ICU to immunocompromised patients with febrile neutropenia to

predominantly HIV-infected patients as studied here. Our study, in particular, is one of the

largest undertaken, with an enrolment of 336 patients, and confirms the results found in earlier,

smaller studies, a major criterion for clinical use not always fulfilled by other proposed

biomarkers.391

While a meta-analysis of studies examining the association between angiopoietin

90

dysregulation and disease severity in sepsis might be useful to obtain a more accurate estimate of

effect size, such an analysis is unlikely to overturn the basic finding of the association itself

given the almost uniformly positive results in studies published to date.

Second, in their use as prognostic biomarkers in sepsis, the angiopoietins have the added benefit

of directly reflecting a component of the underlying pathophysiology, namely endothelial cell

activation. Unlike procalcitonin or Chi3L1, whose precise roles in sepsis and inflammation are

still unclear, angiopoietin dysregulation has a very well-defined and well-studied effect on the

vascular endothelium,9 and therefore makes physiologic sense as a marker of both endothelial

dysfunction and the resultant mortality in sepsis.

Third, the magnitude of angiopoietin dysregulation approximates disease severity both early and

late in sepsis. Angiopoietin dysregulation has been shown to correlate with the evolution of

endothelial activation and its clinical manifestations, including hypotension and ARDS, over

time, such that angiopoietin concentrations remain abnormal during ongoing illness and

normalize with convalescence.9, 127

The persistence of angiopoietin dysregulation during

prolonged illness is a distinct advantage over other proposed biomarkers of sepsis, such as the

pro-inflammatory cytokines IL-6 and TNF, which may be elevated only transiently and therefore

missed during one-time sampling and simply not useful for serial assessments over the course of

a patient‟s hospital or ICU admission.119

Although angiopoietin dysregulation can herald

ongoing endothelial cell activation in prolonged critical illness, it has been studied most often for

its prognostic value upon patient presentation, and has been found to be an equally effective

marker of disease severity in this setting. Ang-2 concentrations rise within two hours of

experimental human endotoxemia, indicating that angiopoietin dysregulation occurs, and

presumably can be detected, in the earliest stages of severe systemic bacterial infection.119

The

phenomenon of early angiopoietin dysregulation is supported by our findings in E. coli O157:H7

infection, in which angiopoietin dysregulation preceded development of the overt clinical

syndrome of HUS, and could distinguish between two clinically indistinguishable patient

populations: those who would experience an uncomplicated course of illness, and those who

would develop HUS.

Finally, angiopoietin dysregulation is widely applicable to a variety of infectious diseases and

syndromes associated with endothelial cell activation or dysfunction. In contrast to a biomarker

91

such as IP-10, which is relatively specific for mycobacterial infection, angiopoietin dysregulation

is sufficiently generalizable to be of potential use in the estimation of disease severity in

undifferentiated patients with suspected severe bacterial infection in the Emergency Department.

At the same time, angiopoietin dysregulation is specific enough to distinguish septic patients

from those with non-infectious critical illness.121, 122

Nonetheless, there are some aspects of the use of angiopoietin dysregulation as a prognostic

biomarker in sepsis that need to be clarified or resolved prior to implementation in clinical

practice. First, the ideal cut-off value for Ang-1, Ang-2, or the Ang-2:Ang-1 ratio needs to be

established. Since this may differ according to whether disease severity or mortality is being

predicted, both values will have to be made available in a format easily interpretable for

clinicians. Furthermore, any biomarker proposed for clinical use must be sufficiently sensitive

and specific to correctly classify patients. To achieve these performance characteristics, it is

likely that angiopoietin dysregulation would need to be combined with other biomarkers in a

prognostic algorithm or clinical decision pathway, and once again, the format of such a pathway

would need to facilitate ease of clinical use. Second, a rapid test, whether laboratory-based or

point-of-care, will need to be established and validated, as the ELISA method used in the studies

presented above will not yield a sufficiently rapid turnaround time to enable use of angiopoietin

dysregulation in clinical decision-making. This test will also need to be relatively inexpensive, or

be associated with a demonstrable net cost savings, in order to justify its use in preference to the

standard (lactate) and free (Karnofsky Performance score) means of prognostication traditionally

used in sepsis. Finally, the ability of angiopoietin monitoring to influence physician management

and patient outcome will need to be proven in prospective studies prior to widespread clinical

implementation. Although there are clearly aspects on which additional research is required,

angiopoietin dysregulation as a prognostic indicator in sepsis fulfills many of Vasan‟s criteria for

the ideal biomarker, and therefore, deserves further study.

Even if the angiopoietins are never employed clinically as biomarkers of infection, the studies

presented here have still identified the angiopoietins as potential mediators of the endothelial

dysfunction present in streptococcal toxic shock syndrome, E. coli O157:H7-induced hemolytic-

uremic syndrome, and sepsis. In that regard, these studies have provided a rationale for further

investigation into the pathophysiology of each syndrome, with a focus on the regulation of

angiopoietin production, release, and impact on the endothelium in each case. Furthermore, these

92

studies provide the basis for pre-clinical and clinical studies to explore the angiopoietins as

potential therapeutic targets in undifferentiated patients with severe bacterial infections. In this

way, studies such as these that document an association between angiopoietin dysregulation and

various infectious diseases characterized by prominent endothelial cell dysfunction can serve to

advance research in the field and potentially impact clinical outcome by means other than simply

through the identification of potential prognostic biomarkers.

93

Chapter 8 Future Directions

The three studies presented above provide the necessary background for conceptualizing future

research on the angiopoietin system. By suggesting that the angiopoietins play a role in the

pathogenesis of toxin-mediated diseases such as streptococcal toxic shock syndrome and the

hemolytic-uremic syndrome, these studies provide a rationale for in vitro experimental work

investigating the interaction between exotoxins and the endothelium. In STSS, we reported

elevated levels of Ang-2 in proportion to disease severity, however, our association study was

not able to determine whether the increased circulating Ang-2 was due to a direct effect of the

streptococcal exotoxin on the endothelium, or whether an intermediary molecule was required.

To determine whether streptococcal exotoxins could induce WPB exocytosis, cultures of human

umbilical vein endothelial cells and human dermal microvascular endothelial cells could be

exposed to streptococcal exotoxins prior to live-cell imaging for the detection of WPB

exocytosis and subsequent measurement of the angiopoietin-2 concentration in the cell culture

supernatant by ELISA. Abrogation of the effect of the streptococcal exotoxin in an experiment

repeated after treatment of the cultured cells with a specific inhibitor of WPB exocytosis would

confirm the mechanism of Ang-2 release. In contrast, in E. coli O157:H7 infection, we reported

decreased circulating Ang-1 levels as the predominant contributor to angiopoietin dysregulation.

Although Ang-1 is produced in part by platelets, decreased Ang-1 levels were present in children

in the pre-HUS phase of illness, prior to the onset of thrombocytopenia, and therefore cannot be

attributed to a decline in platelet numbers. Since pericytes are the primary source of Ang-1, a co-

culture of pericytes and endothelial cells could be exposed to Shiga toxin, with measurement of

Ang-1 levels in the cell culture supernatant both before and after toxin treatment. Cell cultures of

pericytes alone could also be observed and assayed for apoptosis and necrosis after direct

exposure to Shiga toxin. In this way, in vitro studies could explain the observed clinical results.

By suggesting that the angiopoietins may be useful prognostic indicators in E. coli O157:H7

infection and sepsis, these studies provide the rationale for the prospective exploration of the

impact of an angiopoietin-based algorithm on clinical decision-making in the management of

patients with suspected severe bacterial infection. However, as outlined in Chapter 7, appropriate

94

cut-off values need to be defined, and a rapid and perhaps point-of-care test developed before

such studies could be planned or implemented.

Finally, as a continuation of the studies presented above, angiopoietin dysregulation could be

investigated in a variety of clinical conditions with known endothelial dysfunction in which a

demonstrated association may shed light on the underlying, and heretofore unknown,

pathogenesis. Additional work could explore the potential utility of the angiopoietins and other

biomarkers to guide clinical decision-making in areas in which standard diagnostic tools are

either absent or lacking sufficient sensitivity and/or specificity.

8.1 Effect of Ventilatory Strategy on Angiopoietin levels in ARDS

ARDS is characterized by the diffuse leak of intravascular fluid into the alveolar space, resulting

from disruption of the alveolar-capillary membrane barrier. Pulmonary microvascular endothelial

cell dysfunction is a necessary pre-requisite. Consistent with the known effects of the

angiopoietins - endothelial cell quiescence induced by Ang-1 and endothelial cell activation

induced by Ang-2 - a relative excess of Ang-2 has been shown to play a role in the pathogenesis

of ARDS. In vitro, Ang-2 sensitizes human pulmonary microvascular endothelial cells to

thrombin-induced contractility and inter-cellular gap formation.95

In a murine model of LPS-

induced acute lung injury, an increase in alveolar Ang-2 relative to Ang-1 was associated with an

increase in the neutrophil and protein content of bronchoalveolar lavage fluid, suggestive of

pulmonary capillary leak.396

In the same model, overexpression or administration of exogenous

Ang-1 produced improvements in all indices of endothelial function, lung injury, and

inflammation.216, 397, 398

In humans, polymorphisms of the Ang-2 gene have been associated with

increased susceptibility to ALI and ARDS.399, 400

Circulating plasma Ang-2 levels can

differentiate between critically ill patients with ALI and those without.401

Serum from critically

ill patients has been shown to contain high levels of Ang-2 in proportion to the degree of

hypoxemia, and to be capable of disrupting normal barrier function when added to human

microvascular endothelial cells in vitro.9 This effect was reversed by Ang-1 and resolved as

circulating Ang-2 levels declined with convalescence. High Ang2 levels also predicted mortality

in these patients, as has since been confirmed in several other studies of ALI/ARDS in critically

ill patient populations.320, 357, 402

Taken together, these data indicate that angiopoietin

dysregulation is present in, and likely contributes to, the pathogenesis of ALI/ARDS.

95

Mechanical ventilation is the mainstay of supportive therapy for patients with ARDS, however,

ventilator-associated lung injury (also known as „ventilator-induced lung injury‟ [VILI]) remains

a major contributor to morbidity and mortality. High frequency oscillation (HFO) is a means of

delivering small tidal volume ventilation that may limit VILI and the resulting inflammatory

response. When compared to lung-protective conventional ventilation providing similar mean

airway pressures, HFO has been shown to reduce inflammatory cell infiltrate and pro-

inflammatory cytokine production in the lungs in large animal models.403, 404

Human data come

primarily from studies in neonates, which have demonstrated a reduction in the circulating levels

of some, but not all, pro-inflammatory cytokines with the use of HFO.405

Given these findings,

we hypothesize that angiopoietin dysregulation will be present in patients with ARDS receiving

conventional low tidal volume ventilation and will be attenuated in those patients transitioned to

HFO. The only currently existing data on the effect of ventilatory strategy on levels of

circulating angiopoietins are contradictory and derive from two murine models. In the first,

hyperoxia-induced downregulation of Ang-1 was reversed in the setting of permissive

hypercapnia, a model of low tidal volume ventilation.406

In the second, the administration of

exogenous Ang-1 failed to prevent VILI in the setting of high tidal volume ventilation.229

There

are no studies exploring the effect of HFO on circulating levels of the angiopoietins.

There, we propose to measure serum or plasma Ang-1 and Ang-2 levels at study enrolment and

serially thereafter in patients with ALI/ARDS receiving either conventional, lung-protective, low

tidal volume ventilation or high frequency oscillatory ventilation, to test the following

hypotheses:

1. Circulating levels of Ang-1 and Ang-2 will correlate with the severity of illness and will

predict clinical outcome in ventilated patients with ARDS

2. Lower levels of circulating Ang-2 and/or higher levels of circulating Ang-1 (and/or more

rapid correction of angiopoietin dysregulation) will be found in patients receiving HFO

versus those receiving conventional mechanical ventilation

Ultimately, the angiopoietins may be useful biomarkers to guide clinical management decisions

concerning the choice of ventilatory strategies in patients with ARDS.

96

8.2 Angiopoietins in HIV

Untreated HIV disease is associated with endothelial activation and dysfunction, attenuated in

part by the initiation of antiretroviral therapy (ART).407, 408

IL-6, sICAM-1, sVCAM-1, P-

selectin, and vWF are all elevated in patients with HIV disease as compared to healthy

controls.409

In patients receiving ART, both CRP and IL-6 were found to be predictive of the

development of an opportunistic disease.410

Taken together, these results indicate that there is a

baseline level of inflammation and endothelial dysfunction that occurs in chronic HIV infection

and which subsequently diminishes (but does not normalize) with treatment and increases again

in the setting of intercurrent illness or disease.

Although the advent of truly effective ART has been associated with clear survival benefits in

HIV-infected individuals, it has also exposed patients to the consequences and adverse effects of

longterm ART use. Both endothelial activation and an increase in carotid intima-media thickness

(a marker of subclinical atherosclerosis) are well-documented amongst treated patients with HIV

infection.411

Furthermore, certain antiretroviral medications appear to be associated with a

particular increased risk of cardiovascular disease, namely abacavir (ABC) and certain protease

inhibitors (indinavir or lopinavir-ritonavir).412

This risk is not fully accounted for by changes in

lipid profile or other metabolic disorders. Data conflict on the degree of endothelial activation

specifically associated with ABC. VEGF, vWF, IL-6, CRP, and other markers of thrombosis,

inflammation, and endothelial activation have been studied but not found to be different amongst

patients receiving ABC and those receiving non-ABC-containing regimens.413

In contrast, in

vitro studies demonstrated ICAM-1 upregulation and a resultant increase in rolling and adhesion

of neutrophils and peripheral blood mononuclear cells along an endothelial monolayer in the

presence of ABC.414

At this time, the pathogenesis of ABC-induced cardiovascular risk remains

unknown.

Given the known association between angiopoietin dysregulation, cardiovascular disease, and the

traditional risk factors for cardiovascular disease (reviewed in Chapter 1.7.3) in the general

population, I hypothesize that angiopoietin dysregulation might also play a role in ABC-induced

cardiovascular disease in HIV. This hypothesis is further supported by in vitro data that describe

the attenuation of ABC-induced endothelial dysfunction by specific inhibitors of the Erk

97

signalling pathway, one of the major downstream regulators of Tie-2 activation and associated

with endothelial cell proliferation.415

Although the majority of patients in the sepsis study reported here (Chapter 5) were HIV-

infected, most were not receiving ART, and all had angiopoietin levels measured in the setting of

acute illness. To my knowledge, there are no studies that explore the angiopoietins in treated

HIV disease, nor specifically in association with ABC-induced cardiovascular disease.

Therefore, I propose to investigate Ang-1 and Ang-2 levels in a cohort of patients receiving

ABC, and to compare these levels to that of a control group of patients (matched for age,

duration and stage of HIV infection, and cardiovascular risk factors and disease) receiving non-

ABC-containing ARV regimens. I hypothesize that angiopoietin dysregulation (decreased Ang-1

and/or increased Ang-2) will be greater in those patients receiving ABC-containing ARV

regimens than in the control group. I further hypothesize that amongst those patients receiving

non-ABC-containing regimens, angiopoietin dysregulation will be an independent predictor of

cardiovascular disease.

8.3 Angiopoietins in Obstructive Sleep Apnea

Obstructive sleep apnea (OSA) is defined by frequent and recurrent complete or partial collapse

of the upper airway during sleep, resulting in apneas or hypopneas, intermittent asphyxia, and

periodic awakenings leading to daytime somnolence.416

OSA is common, and depending on the

precise disease definition used, has been documented in as many as 24% of middle-aged men.417

OSA is also serious; it is an independent predictor of future drug-resistant hypertension,

myocardial infarction, stroke, and death from cardiovascular disease.416

As recently reviewed by

Drager et al, multiple mechanisms have been proposed to link atherosclerosis and OSA,

including dyslipidemia, the generation of reactive oxygen species and subsequent lipid

peroxidation, and inflammation.416

Each of these risk factors for atherosclerosis can be induced

in a mouse model by intermittent hypoxia, a defining characteristic of OSA in humans.416

As the site of atherosclerosis, the endothelium is clearly implicated as a mediator of many of the

adverse consequences of OSA. However, impaired endothelial function occurs even in the

absence of overt cardiovascular disease, as evidenced by diminished endothelium-dependent

vasodilatation, enhanced expression of cell-surface adhesion molecules, and increased

circulating levels of tissue factor and vWF.418-422

98

Since hypoxia is a known regulator of Weibel-Palade body exocytosis, and since a

vasoconstrictive, adhesive, prothrombotic endothelial cell phenotype such as that seen in OSA is

the predicted result of a net increase in circulating Ang-2, I hypothesize that angiopoietin

dysregulation might occur in obstructive sleep apnea and might provide an additional link

between intermittent hypoxia and the endothelium. Although angiopoietin-like protein 4 has

been associated with chronic intermittent hypoxia through its role in lipid metabolism

(specifically as an inhibitor of the enzyme lipoprotein lipase needed to clear triglyceride-rich

lipoproteins from the circulation), Angiopoietin-1 and -2 have not been studied in OSA.423

To investigate this hypothesis, I propose a study comparing plasma levels of Ang-1 and Ang-2 in

three groups of patients: a) untreated patients with newly diagnosed OSA, b) patients with

treated OSA adherent to the prescribed regimen of nasal continuous positive airway pressure

(CPAP), and c) untreated healthy controls without OSA. Based on the above data, I expect to

find angiopoietin dysregulation, independent of other cardiovascular disease risk factors, present

in untreated patients with OSA. Furthermore, treated patients with OSA who are adherent to

CPAP should manifest an angiopoietin profile akin to that of healthy controls without sleep-

disordered breathing, provided that the prescribed CPAP therapy has been documented with a

titration sleep study to reduce or eliminate apneic/hypopneic episodes. As a substudy, I propose

to follow the newly diagnosed patients with serial measurements of plasma Ang-1 and Ang-2

both before and after the initiation of CPAP with the intent of documenting reduced angiopoietin

dysregulation in individual patients in whom CPAP is successful in eliminating or reducing

intermittent nocturnal hypoxia.

8.4 Procalcitonin in patients with Hematologic Malignancy

The duration of appropriate broad-spectrum antibiotic therapy in patients with prolonged febrile

neutropenia without obvious source is a common and contentious issue.424

Early discontinuation

in the setting of unrecognized infection could lead to morbidity or mortality, while delayed

discontinuation in the absence of infection exposes patients to unnecessary antibiotics and the

inherent risk therein, and may increase antimicrobial resistance. Since microbiologic cultures and

other diagnostic tests are often negative in this patient population, an alternative marker of

infection with sufficient negative predictive value to guide the safe discontinuation of antibiotics

(although not necessarily antifungals) would be of benefit.

99

Procalcitonin has been proposed as a marker to guide antibiotic therapy, and in this role was the

subject of a recent systematic review by Schuetz et al.425

Fourteen randomized controlled trials,

enrolling a total of 4467 patients with respiratory tract infections or sepsis in primary care,

emergency department or ICU settings, were included in the analysis. Use of a procalcitonin cut-

off value of < 0.25 ng/mL in primary care or the ED, or < 0.25 - 1 ng/mL in the ICU to

determine initiation or duration of antibiotics resulted in a significant decrease in antibiotic use

and duration, without a change in mortality. A second systematic review by Agarwal and

Schwartz focused on sepsis in the ICU and confirmed that procalcitonin-based algorithms for

antibiotic discontinuation lead to reduced antibiotic exposure without compromising survival.426

Importantly, this finding was sufficiently robust to withstand relatively high rates of physician

non-adherence to the algorithm (failure to discontinue antibiotics despite a low or declining

procalcitonin level). A third systematic review focused on only those studies exploring PCT-

guided antibiotic therapy in patients with respiratory tract infections, and found results identical

to those of Schuetz and Agarwal and Schwartz.427

However, since the publication of these meta-analyses, a large study of 1200 critically ill patients

randomized to either standard or PCT-guided antibiotic therapy found that PCT-guided therapy

did not improve survival, and in fact prolonged ICU stay and increased use of broad-spectrum

antimicrobial agents.428

In this study, a PCT measurement of ≥ 1 ng/mL triggered escalation and

expansion of the spectrum of the patient‟s current antimicrobial regimen. Nonetheless, PCT may

still be useful to guide antimicrobial discontinuation, as opposed to escalation.

Before such a study is undertaken, the sensitivity and specificity of PCT for the diagnosis of

bacterial infection in this population must be clarified. In a mixed population of

immunocompromised patients (HIV or hematologic/solid tumour malignancies), procalcitonin

levels greater than 0.5 ng/mL displayed excellent sensitivity for the diagnosis of bacterial

infection, and remained an independent predictor of such in a multivariate analysis.429

However,

a separate study found that while all patients with bacteremia in the setting of febrile neutropenia

ultimately manifested a procalcitonin level above 0.5 ng/mL (the lower limit of detection for first

generation assays), a substantial portion of patients with febrile neutropenia without source also

manifested PCT values above this limit.430

While some studies have found that PCT can

effectively identify those patients with bacteremia from amongst the total population of patients

with febrile neutropenia, others have found reduced sensitivity and specificity of PCT in febrile

100

neutropenic as compared to febrile non-neutropenic patients for the diagnosis of bacterial

infection.431, 432

In a study of 90 patients undergoing chemotherapy for a hematologic malignancy

and who presented with febrile neutropenia, admission plasma procalcitonin levels, using a

sufficiently sensitive assay, were unable to discriminate between those with microbiologically or

clinically documented infections and those with fever without source.433

However, if the peak

procalcitonin level occurred more than three days after the onset of febrile neutropenia, it was

predictive of an invasive fungal infection, perhaps indicating that elevated procalcitonin levels

have a different significance or should be interpreted differently in immunocompromised

patients, and certainly indicating that the utility of PCT in patients with febrile neutropenia

requires further exploration.

Consequently, I would propose two sequential studies:

1. Cohort study of the performance characteristics of PCT to discriminate

microbiologically-confirmed bacterial from non-bacterial infection in non-neutropenic

patients with hematologic malignancy and in neutropenic patients with fever of short

duration (< 5 days).

2. Randomized controlled trial of PCT-guided antibiotic discontinuation in prolonged

febrile neutropenia (> 14 days) without source, provided the findings from study 1

confirm a sufficient negative predictive value for PCT in this setting.

8.5 Systematic review of procalcitonin for the prediction of mortality in sepsis and severe bacterial infection

As can be seen from the literature review in Chapter 1.9.1 and from Table 1, there are a

substantial number of publications, many with conflicting results, pertaining to the use of PCT to

predict outcome in sepsis or severe bacterial infection. Distilling and organizing this information

in a systematic fashion may yield a better understanding of the potential benefit, or lack thereof,

to the use of PCT for this indication. To my knowledge, no such review is ongoing, and searches

of both the Cochrane Database of Systematic Reviews and the newly launched Prospero

International Registry of Systematic Reviews yielded no registered reviews.

101

References

1. Aird WC. Endothelium as an organ system. Crit Care Med. 2004; 32: S271-S279.

2. Levi M, van der Poll T. Inflammation and coagulation. Crit Care Med. 2010; 38: S26-S34.

3. Schouten M, Wiersinga WJ, Levi M, et al. Inflammation, endothelium, and coagulation in

sepsis. J Leukoc Biol. 2008; 83: 536-545.

4. Aird WC. Phenotypic heterogeneity of the endothelium. I. Structure, Function, and

Mechanisms. Circ Res. 2007; 100: 158-173.

5. Lowenstein CJ, Morrell CN, Yamakuchi M. Regulation of Weibel-Palade body exocytosis.

Trends Cardiovasc Med. 2005; 15: 302-308.

6. Matsushita K, Morrell CN, Cambien B, et al. Nitric oxide regulates exocytosis by S-

nitrosylation of N-ethylmaleimide-sensitive factor. Cell. 2003; 115: 139-150.

7. Fiedler U, Scharpfenecker M, Koidl S, et al. The Tie-2 ligand angiopoietin-2 is stored in

and rapidly released upon stimulation from endothelial cell Weibel-Palade bodies. Blood.

2004; 103: 4150-4156.

8. Aghajanian A, Wittchen ES, Allingham MJ, et al. Endothelial cell junctions and the

regulation of vascular permeability and leukocyte transmigration. J Thromb Haemost.

2008; 6: 1453-1460.

9. Parikh SM, Mammoto T, Schultz A, et al. Excess circulating angiopoietin-2 may

contribute to pulmonary vascular leak in sepsis in humans. PLoS Med. 2006; 3: e46.

10. The Working Group on Severe Streptococcal Infections: Breiman RF, Davis JP, Facklam

RR, et al. Defining the Group A Streptococcal Toxic Shock Syndrome: Rationale and

consensus definition. JAMA. 1993; 269: 390-391.

11. Norrby-Teglund A, Chatellier S, Low DE, et al. Host variation in cytokine responses to

superantigens determine the severity of invasive group A streptococcal infection. Eur J

Immunol. 2000; 30: 3247-3255.

102

12. Johansson L, Thulin P, Low DE, Norrby-Teglund A. Getting under the skin: the

immunopathogenesis of Streptococcus pyogenes deep tissue infections. Clin Infect Dis.

2010; 51: 58-65.

13. Lappin E, Ferguson AJ. Gram-positive toxic shock syndromes. Lancet Infect Dis. 2009; 9:

281-290.

14. Lepoutre A, Doloy A, Bidet P, et al. Epidemiology of invasive Streptococcus pyogenes

infections in France, 2007. J Clin Microbiol. 2011; Oct 5. [Epub ahead of print].

15. Rasko DA, Webster DR, Sahl JW, et al. Origins of the E. coli strain causing an outbreak of

hemolytic-uremic syndrome in Germany. New Engl J Med. 2011; 365: 709-717.

16. Tarr PI, Gordon CA, Chandler WL. Shiga-toxin-producing Escherichia coli and

haemolytic uraemic syndrome. Lancet. 2005; 365: 1073-1086.

17. Richardson SE, Karmali MA, Becker LE, Smith CR. The histopathology of the hemolytic

uremic syndrome associated with verocytotoxin-producing Escherichia coli infections.

Hum Pathol. 1988; 19: 1102-1108.

18. Chandler WL, Jelacic S, Boster DR, et al. Prothrombotic coagulation abnormalities

preceding the haemolytic-uremic syndrome. New Engl J Med. 2002; 346: 23-32.

19. Nolasco LH, Turner NA, Bernardo A, et al. Hemolytic uremic syndrome-associated Shiga

toxins promote endothelial-cell secretion and impair ADAMTS13 cleavage of unusually

large von Willebrand factor multimers. Blood. 2005; 106: 4199-4209.

20. Zoja C, Angioletti S, Donadelli R, et al. Shiga toxin-2 triggers endothelial leukocyte

adhesion and transmigration via NF-κB dependent up-regulation of IL-8 and MCP-1.

Kidney Int. 2002; 62: 846-856.

21. Morigi M, Micheletti G, Figliuzzi M, et al. Verotoxin-1 promotes leukocyte adhesion to

cultured endothelial cells under physiologic flow conditions. Blood. 1995; 86: 4553-4558.

22. van de Kar NCAJ, Monnens LAH, Karmali MA, van Hinsbergh VWM. Tumor Necrosis

Factor and Interleukin-1 induce expression of the verocytotoxin receptor

103

Globotriaosylceramide on human endothelial cells: implications for the pathogenesis of the

hemolytic uremic syndrome. Blood. 1992; 80: 2755-2764.

23. Matussek A, Lauber J, Bergau A, et al. Molecular and functional analysis of Shiga toxin-

induced response patterns in human vascular endothelial cells. Blood. 2003; 102: 1323-

1332.

24. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines

for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference

Committee. American College of Chest Physicians/Society of Critical Care Medicine.

Chest. 1992; 101: 1644-1655.

25. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS

International Sepsis Definitions Conference. Crit Care Med. 2003; 31: 1250-1256.

26. Canadian Institute for Health Information, In Focus: A National Look at Sepsis. (Ottawa,

Ont.: CIHI, 2009).

27. Martin CM, Priestap F, Fisher H, et al. A prospective, observational registry of patients

with severe sepsis: the Canadian Sepsis Treatment and Response Registry. Crit Care Med.

2009; 37: 81-88.

28. Angus DC, Linde-Zwirble WT, Lidicker J, et al. Epidemiology of severe sepsis in the

United States: Analysis of incidence, outcome, and associated costs of care. Crit Care Med.

2001; 29: 1303-1310.

29. Engel C, Brunkhorst FM, Bone HG, et al. Epidemiology of sepsis in Germany: results

from a national prospective multicenter study. Intensive Care Med. 2007; 33: 606-618.

30. Adrie C, Alberti C, Chaix-Couturier C, et al. Epidemiology and economic evaluation of

severe sepsis in France: Age, severity, infection site, and place of acquisition (community,

hospital, or intensive care unit) as determinants of workload and cost. J Crit Care. 2005;

20: 46-58.

31. Lee WL, Liles WC. Endothelial activation, dysfunction and permeability during severe

infections. Curr Opin Hematol. 2011; 18: 191-196.

104

32. Ye X, Ding J, Zhou X, et al. Divergent roles of endothelial NF-kappaB in multiple organ

injury and bacterial clearance in mouse models of sepsis. J Exp Med. 2008; 205: 1303-

1315.

33. Davis JS, Yeo TW, Thomas JH, et al. Sepsis-associated microvascular dysfunction

measured by peripheral arterial tonometry: an observational study. Crit Care. 2009; 13:

R155.

34. Shapiro NI, Schuetz P, Yano K, et al. The association of endothelial cell signalling,

severity of illness, and organ dysfunction in sepsis. Crit Care. 2010; 14: R182.

35. Andonegui G, Bonder CS, Green F, et al. Endothelium-derived Toll-like receptor-4 is the

key molecule in LPS-induced neutrophil sequestration into lungs. J Clin Invest. 2003; 111:

1011-1020.

36. Zanotti G, Casiraghi M, Abano JB, et al. Novel critical role of Toll-like receptor 4 in lung

ischemia-reperfusion injury and edema. Am J Physiol Lung Cell Mol Physiol. 2009; 297:

L52-L63.

37. Magder S, Neculcea J, Neculcea V, Sladek R. Lipopolysaccharide and TNF-alpha produce

very similar changes in gene expression in human endothelial cells. J Vasc Res. 2006; 43:

447-461.

38. Faure E, Thomas L, Xu H, et al. Bacterial lipopolysaccharide and IFN-gamma induce Toll-

like receptor 2 and Toll-like receptor 4 expression in human endothelial cells: role of NF-

kappa B activation. J Immunol. 2001; 166: 2018-2024.

39. Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis.

Nature. 2011; 473: 298-307.

40. Zhang J, Fukuhara S, Sako K, et al. Angiopoietin-1/Tie2 signal augments basal Notch

signal controlling vascular quiescence by inducing delta-like 4 expression through AKT-

mediated activation of β-catenin. J Biol Chem. 2011; 286: 8055-8066.

41. Mandriota SJ, Pepper MS. Regulation of angiopoietin-2 mRNA levels in bovine

microvascular endothelial cells by cytokines and hypoxia. Circ Res. 1998; 83: 852-859.

105

42. Rychli K, Kaun C, Hohensinner PJ, et al. The inflammatory mediator oncostatin M

induces angiopoietin 2 expression in endothelial cells in vitro and in vivo. J Thromb

Haemost. 2010; 8: 596-604.

43. Huang Y, Song N, Ding Y, et al. Pulmonary vascular destabilization in the premetastatic

phase facilitates lung metastasis. Cancer Res. 2009; 69: 7529-7537.

44. Suchting S, Freitas C, le Noble F, et al. The Notch ligand Delta-like 4 negatively regulates

endothelial tip cell formation and vessel branching. Proc Natl Acad Sci USA. 2007; 104:

3225-3230.

45. Hayes AJ, Huang WQ, Mallah J, et al. Angiopoietin-1 and its receptor Tie-2 participate in

the regulation of capillary-like tubule formation and survival of endothelial cells.

Microvasc Res. 1999; 58: 224-237.

46. Davis S, Papadopoulos N, Aldrich TH, et al. Angiopoietins have distinct modular domains

essential for receptor binding, dimerization and superclustering. Nat Struct Biol. 2003; 10:

38-44.

47. Kim KT, Choi HH, Steinmetz MO, et al. Oligomerization and multimerization are critical

for angiopoietin-1 to bind and phosphorylate Tie2. J Biol Chem. 2005; 280: 20126-20131.

48. Ribatti D, Nico B, Crivellato E. The role of pericytes in angiogenesis. Int J Dev Biol. 2011;

55: 261-268.

49. Nadar SK, Blann AD, Lip GYH. Plasma and platelet-derived vascular endothelial growth

factor and angiopoietin-1 in hypertension: effects of antihypertensive therapy. J Intern

Med. 2004; 256: 331-337.

50. Neagoe PE, Brkovic A, Hajjar F, Sirois MG. Expression and release of angiopoietin-1

from human neutrophils: intracellular mechanisms. Growth Factors. 2009; 27: 335-344.

51. Sundberg C, Kowanetz M, Brown LF, et al. Stable expression of angiopoietin-1 and other

markers by cultured pericytes: phenotypic similarities to a subpopulation of cells in

maturing vessels during later stages of angiogenesis in vivo. Lab Invest. 2002; 82: 387-

401.

106

52. Fujii T, Kuwano H. Regulation of the expression balance of angiopoietin-1 and

angiopoietin-2 by Shh and FGF-2. In vitro Cell Dev Biol. 2010; 46: 487-491.

53. Matsushita K, Morrell CN, Lowenstein CJ. Sphingosine-1-phosphate activates Weibel-

Palade body exocytosis. Proc Natl Acad Sci USA. 2004; 101: 11483-11487.

54. Jang C, Koh YJ, Lim NK, et al. Angiopoietin-2 exocytosis is stimulated by sphingosine-1-

phosphate in human blood and lymphatic endothelial cells. Arterioscler Thromb Vasc Biol.

2009; 29: 401-407.

55. Ventkataraman K, Lee YM, Michaud J, et al. Vascular endothelium as a contributor of

plasma sphingosine 1-phosphate. Circ Res. 2008; 102: 669-676.

56. Li X, Stankovic M, Bonder CS, et al. Basal and angiopoietin-1-mediated endothelial

permeability is regulated by sphingosine kinase-1. Blood. 2008; 11: 3489-3497.

57. Sato TN, Qin Y, Kozak CA, Audus KL. Tie-1 and tie-2 define another class of putative

receptor tyrosine kinase genes expressed in early embryonic vascular system. Proc Natl

Acad Sci USA. 1993; 90: 9355-9358.

58. Barton WA, Tzvetkova-Robev D, Miranda EP, et al. Crystal structures of the Tie2 receptor

ectodomain and the angiopoietin-2-Tie2 complex. Nat Struct Mol Biol. 2006; 13: 524-532.

59. Fiedler U, Krissl T, Koidl S, et al. Angiopoietin-1 and angiopoietin-2 share the same

binding domains in the Tie-2 receptor involving the first Ig-like loop and the Epidermal

Growth Factor-like repeats. J Biol Chem. 2003; 278: 1721-1727.

60. Yuan HT, Khankin EV, Karumanchi SA, Parikh SM. Angiopoietin 2 is a partial

agonist/antagonist of Tie2 signalling in the endothelium. Mol Cell Biol. 2009; 29: 2011-

2022.

61. Bogdanovic E, Nguyen VPKH, Dumont DJ. Activation of Tie2 by angiopoietin-1 and

angiopoietin-2 results in their release and receptor internalization. J Cell Sci. 2006; 119:

3551-3560.

107

62. Bogdanovic E, Coombs N, Dumont DJ. Oligomerized Tie2 localizes to clathrin-coated pits

in response to angiopoietin-1. Histochem Cell Biol. 2009; 132: 225-237.

63. Saharinen P, Eklund L, Miettinen J, et al. Angiopoietins assemble distinct Tie2 signalling

complexes in endothelial cell-cell and cell-matrix contacts. Nat Cell Biol. 2008; 10: 527-

537.

64. Wherle C, Van Slyke P, Dumont DJ. Angiopoietin-1 induced ubiquitylation of Tie2 by c-

CBL is required for internalization and degradation. Biochem J. 2009; 423: 375-380.

65. Findley CM, Cudmore MJ, Ahmed A, Kontos CD. VEGF induces Tie2 shedding via a

phosphoinositide 3-kinase/Akt-dependent pathway to modulate Tie2 signalling.

Arterioscler Thromb Vasc Biol. 2007; 27: 2619-2626.

66. Yuan HT, Venkatesha S, Chan B, et al. Activation of the orphan endothelial receptor Tie1

modifies Tie2-mediated intracellular signalling and cell survival. FASEB J. 2007; 21:

3171-3183.

67. Saharinen P, Kerkela K, Ekman N, et al. Multiple angiopoietin recombinant proteins

activate the Tie1 receptor tyrosine kinase and promote its interaction with Tie2. J Cell

Biol. 2005; 169: 239-243.

68. Hansen TM, Singh H, Tahir TA, Brindle NPJ. Effects of angiopoietins-1 and -2 on the

receptor tyrosine kinase Tie2 are differentially regulated at the endothelial cell surface.

Cell Signal. 2010; 22: 527-532.

69. Seegar TCM, Eller B, Tzvetkova-Robev D, et al. Tie1-Tie2 interactions mediate functional

differences between angiopoietin ligands. Mol Cell. 2010; 37: 643-655.

70. Singh H, Milner CS, Aguilar Hernandez MM, et al. Vascular endothelial growth factor

activates the Tie family of receptor tyrosine kinases. Cell Signal. 2009; 21: 1346-1350.

71. Marron MB, Singh H, Tahir TA, et al. Regulated proteolytic processing of Tie1 modulates

ligand responsiveness of the receptor-tyrosine kinase Tie2. J Biol Chem. 2007; 282:

30509-30517.

108

72. Yabkowitz R, Meyer S, Black T, et al. Inflammatory cytokines and vascular endothelial

growth factor stimulate the release of soluble tie receptor from human endothelial cells via

metalloproteinase activation. Blood. 1999; 93: 1969-1979.

73. Chen-Konak L, Guetta-Shubin Y, Yahav H, et al. Transcriptional and post-translation

regulation of the Tie1 receptor by fluid shear stress changes in vascular endothelial cells.

FASEB J. 2003; 17: 2121-2123.

74. Sato TN, Tozawa Y, Deutsch U, et al. Distinct roles of the receptor tyrosine kinases Tie-1

and Tie-2 in blood vessel formation. Nature. 1995; 376: 70-74.

75. D‟Amico G, Korhonen EA, Waltari M, et al. Loss of the endothelial Tie1 receptor impairs

lymphatic vessel development - brief report. Arterioscler Thromb Vasc Biol. 2010; 30:

207-209.

76. Puri MC, Partanen J, Rossant J, Bernstein A. Interaction of the TEK and TIE receptor

tyrosine kinases during cardiovascular development. Development. 1999; 126: 4569-4580.

77. Woo KV, Qu X, Babaev VR, et al. Tie1 attenuation reduces murine atherosclerosis in a

dose-dependent and shear stress-specific manner. J Clin Invest. 2011; 121: 1624-1635.

78. Chan B, Yuan HT, Karumanchi SA, Sukhatme VP. Receptor tyrosine kinase Tie-1

overexpression in endothelial cells upregulates adhesion molecules. Biochem Biophys Res

Commun. 2008; 371: 475-479.

79. Chan B, Sukhatme VP. Suppression of Tie-1 in endothelial cells in vitro induces a change

in the genome-wide expression profile reflecting an inflammatory function. FEBS Lett.

2009; 583: 1023-1028.

80. Milner CS, Hansen TM, Singh H, Brindle NPJ. Roles of the receptor tyrosine kinases Tie1

and Tie2 in mediating the effects of angiopoietin-1 on endothelial permeability and

apoptosis. Microvasc Res. 2009; 77: 187-191.

81. Kim I, Kim HG, So JN, et al. Angiopoietin-1 regulates endothelial cell survival through

the phosphatidylinositol 3‟-kinase/Akt signal transduction pathway. Circ Res. 2000; 86:

24-29.

109

82. Fukuhara S, Sako K, Minami T, et al. Differential function of Tie2 at cell-cell contacts and

cell-substratum contacts regulated by angiopoietin-1. Nat Cell Biol. 2008; 10: 513-526.

83. Katoh SY, Kamimoto T, Yamakawa D, Takakura N. Lipid rafts serve as signalling

platforms for Tie2 receptor tyrosine kinase in vascular endothelial cells. Exp Cell Res.

2009; 315: 2818-2823.

84. Sako K, Fukuhara S, Minami T, et al. Angiopoietin-1 induces Krüppel-like factor 2

expression through a phosphoinositide 3-kinase/AKT-dependent activation of myocyte

enhancer factor 2. J Biol Chem. 2009; 284: 5592-5601.

85. Abdel-Malak NA, Mofarrahi M, Mayaki D, et al. Early Growth Response-1 regulates

angiopoietin-1 induced endothelial cell proliferation, migration, and differentiation.

Arterioscler Thromb Vasc Biol. 2009; 29: 209-216.

86. Abdel-Malak NA, Harfouche R, Hussain SNA. Transcriptome of angiopoietin 1-activated

human umbilical vein endothelial cells. Endothelium. 2007; 14: 285-302.

87. Salmon AHJ, Neal CR, Sage LM, et al. Angiopoietin-1 alters microvascular permeability

coefficients in vivo via modification of endothelial glycocalyx. Cardiovasc Res. 2009; 83:

24-33.

88. Mammoto T, Parikh SM, Mammoto A, et al. Angiopoietin-1 requires p190 RhoGAP to

protect against vascular leakage in vivo. J Biol Chem. 2007; 282: 23910-23918.

89. Choudhury A, Freestone B, Patel J, Lip GYH. Relationship of soluble CD40 ligand to

vascular endothelial growth factor, angiopoietins, and tissue factor in atrial fibrillation.

Chest. 2007; 132: 1913-1919.

90. Kim I, Oh JL, Ryu YS, et al. Angiopoietin-1 negatively regulates expression and activity

of tissue factor in endothelial cells. FASEB J. 2002; 16: 126-128.

91. McCarter SD, Lai PFH, Suen RS, Stewart DJ. Regulation of Endothelin-1 by angiopoietin-

1: Implications for inflammation. Exp Biol Med. 2006; 231: 985-991.

110

92. Kim I, Moon SO, Park SK, et al. Angiopoietin-1 reduces VEGF-stimulated leukocyte

adhesion to endothelial cells by reducing ICAM-1, VCAM-1, and E-selectin expression.

Circ Res. 2001; 89: 477-479.

93. Kim DH, Jung YJ, Lee AS, et al. Comp-Angiopoietin-1 decreases lipopolysaccharide-

induced acute kidney injury. Kidney Int. 2009; 76: 1180-1191.

94. Niu Q, Perruzzi C, Voskas D, et al. Inhibition of Tie-2 signalling induces endothelial cell

apoptosis, decreases Akt signalling and induces endothelial cell expression of the

endogenous anti-angiogenic molecule, thrombospondin-1. Cancer Biol Ther. 2004; 3: 402-

405.

95. van der Heijden M, van Nieuw Amerongen GP, van Bezu J, et al. Opposing effects of the

angiopoietins on the thrombin-induced permeability of human pulmonary microvascular

endothelial cells. PLoS ONE. 2011; 6: e23448.

96. Thomas M, Felcht M, Kruse K, et al. Angiopoietin-2 stimulation of endothelial cells

induces αvβ3 integrin internalization and degradation. J Biol Chem. 2010; 285: 23942-

23849.

97. Scharpfenecker M, Fiedler U, Reiss Y, Augustin HG. The Tie-2 ligand angiopoietin-2

destabilizes quiescent endothelium through an internal autocrine loop mechanism. J Cell

Sci. 2004; 118: 771-780.

98. Scholz A, Lang V, Henschler R, et al. Angiopoietin-2 promotes myeloid cell infiltration in

a β2 integrin-dependent manner. Blood. 2011; 118: 5050-5059.

99. Suri C, Jones PF, Patan S, et al. Requisite role of angiopoietin-1, a ligand for the Tie2

receptor, during embryonic angiogenesis. Cell. 1996; 87: 1171-1180.

100. Maisonpierre PC, Suri C, Jones PF, et al. Angiopoietin-2, a natural antagonist for Tie2 that

disrupts in vivo angiogenesis. Science. 1997; 277: 55-60.

101. Gale NW, Thurston G, Hackett SF, et al. Angiopoietin-2 is required for postnatal

angiogenesis and lymphatic patterning, and only the latter role is rescued by angiopoietin-

1. Dev Cell. 2002; 3: 411-423.

111

102. Lobov IB, Brooks PC, Lang RA. Angiopoietin-2 displays VEGF-dependent modulation of

capillary structure and endothelial cell survival in vivo. Proc Natl Acad Sci USA. 2002; 99:

11205-11210.

103. Nguyen VPKH, Chen SH, Trinh J, et al. Differential response of lymphatic, venous, and

arterial endothelial cells to angiopoietin-1 and angiopoietin-2. BMC Cell Biol. 2007; 8: 10.

104. Fiedler U, Reiss Y, Scharpfenecker M, et al. Angiopoietin-2 sensitizes endothelial cells to

TNF- and has a crucial role in the induction of inflammation. Nat Med. 2006; 12: 235-

239.

105. Thurston G, Suri C, Smith K, et al. Leakage-resistant blood vessels in mice transgenically

overexpressing angiopoietin-1. Science. 1999; 286: 2511-2514.

106. Thurston G, Rudge JS, Ioffe E, et al. Angiopoietin-1 protects the adult vasculature against

plasma leakage. Nat Med. 2000; 6: 460-463.

107. Jeansson M, Gawlik A, Anderson G, et al. Angiopoietin-1 is essential in mouse

vasculature during development and in response to injury. J Clin Invest. 2001; 121: 2278-

2289.

108. Lee HJ, Cho CH, Hwang SJ, et al. Biological characterization of angiopoietin-3 and

angiopoietin-4. FASEB J. 2004; 18: 1200-1208.

109. Valenzuela DM, Griffiths JA, Rojas R, et al. Angiopoietins 3 and 4: Diverging gene

counterparts in mice and humans. Proc Natl Acad Sci USA. 1999; 96: 1904-1909.

110. Olsen MWB, Ley CD, Junker N, et al. Angiopoietin-4 inhibits angiogenesis and reduces

interstitial fluid pressure. Neoplasia. 2006; 8: 364-372.

111. Brunckhorst MK, Wang H, Lu R, Yu Q. Angiopoietin-4 promotes glioblastoma

progression by enhancing tumor cell viability and angiogenesis. Cancer Res. 2010; 70:

7283-7293.

112

112. Chomel C, Cazes A, Faye C, et al. Interaction of the coiled-coil domain with

glycosaminoglycans protects angiopoietin-like 4 from proteolysis and regulates its

antiangiogenic activity. FASEB J. 2009; 23: 940-949.

113. Huang RL, Teo Z, Chong HC, et al. ANGPLT4 modulates vascular junction integrity by

integrin signalling and disruption of intercellular VE-cadherin and claudin-5 clusters.

Blood. 2011; 118: 3990-4002.

114. Koliwad SK, Kuo T, Shipp LE, et al. Angiopoietin-like 4 (ANGPLT4/FIAF) is a direct

glucocorticoid receptor target and participates in glucocorticoid-regulated triglyceride

metabolism. J Biol Chem. 2009; 284: 25593-25601.

115. Tabata M, Kadomatsu T, Fukuhara S, et al. Angiopoietin-like protein 2 promotes chronic

adipose tissue inflammation and obesity-related systemic insulin resistance. Cell Metab.

2009; 10: 178-188.

116. Musunuru K, Pirruccello JP, Do R, et al. Exome sequencing, ANGPTL3 mutations, and

familial combined hypolipidemia. New Engl J Med. 2010; 363: 2220-2227.

117. Kadomatsu T, Tabata M, Oike Y. Angiopoietin-like proteins: emerging targets for

treatment of obesity and related metabolic diseases. FEBS J. 2011; 278: 559-564.

118. Lukasz A, Hellpap J, Horn R, et al. Circulating angiopoietin-1 and angiopoietin-2 in

critically ill patients: development and clinical application of two new immunoassays. Crit

Care. 2008; 12: R94.

119. Kümpers P, van Meurs M, David S, et al. Time course of angiopoietin-2 release during

experimental human endotoxemia and sepsis. Crit Care. 2009; 13: R64.

120. van der Heijden M, Pickkers P, van Nieuw Amerongen GP, et al. Circulating angiopoietin-

2 levels in the course of septic shock: relation with fluid balance, pulmonary dysfunction

and mortality. Intensive Care Med. 2009; 35: 1567-1574.

121. Orfanos SE, Kotanidou A, Glynos C, et al. Angiopoietin-2 is increased in severe sepsis:

Correlation with inflammatory mediators. Crit Care Med. 2007; 35: 199-206.

113

122. Siner JM, Bhandari V, Engle KM, et al. Elevated serum angiopoietin 2 levels are

associated with increased mortality in sepsis. Shock. 2009; 31: 348-353.

123. Davis JS, Yeo TW, Piera KA, et al. Angiopoietin-2 is increased in sepsis and inversely

associated with nitric oxide-dependent microvascular reactivity. Crit Care. 2010; 14: R89.

124. Kümpers P, Lukasz A, David S, et al. Excess circulating angiopoietin-2 is a strong

predictor of mortality in critically ill medical patients. Crit Care. 2008; 12: R147.

125. Kümpers P, Hafer C, David S, et al. Angiopoietin-2 in patients requiring renal replacement

therapy in the ICU: relation to acute kidney injury, multiple organ dysfunction syndrome

and outcome. Intensive Care Med. 2010; 36: 462-470.

126. Giuliano JS, Lahni PM, Harmon K, et al. Admission angiopoietin levels in children with

septic shock. Shock. 2007; 28: 650-654.

127. Ricciuto DR, dos Santos CC, Hawkes M, et al. Angiopoietin-1 and angiopoietin-2 as

clinically informative prognostic biomarkers of morbidity and mortality in severe sepsis.

Crit Care Med. 2011; 39: 702-710.

128. Mankhambo LA, Banda DL, the IPD Study Group, et al. The role of angiogenic factors in

predicting clinical outcome in severe bacterial infection in Malawian children. Crit Care.

2010; 14: R91.

129. Minhas N, Xue M, Fukudome K, Jackson CJ. Activated protein C utilizes the

angiopoietin/Tie2 axis to promote endothelial barrier function. FASEB J. 2010; 24: 873-

881.

130. Alves BE, Montalvao SAL, Aranha FJP, et al. Imbalances in serum angiopoietin

concentrations are early predictors of septic shock development in patients with post

chemotherapy febrile neutropenia. BMC Infect Dis. 2010; 10: 143.

131. van der Heijden M, van Nieuw Amerongen GP, van Hinsbergh VWM, Groeneweld ABJ.

The interaction of soluble Tie2 with angiopoietins and pulmonary vascular permeability in

septic and nonseptic critically ill patients. Shock. 2010; 33: 263-268.

114

132. Yeo TW, Lampah DA, Gitawati R, et al. Angiopoietin-2 is associated with decreased

endothelial nitric oxide and poor clinical outcome in severe falciparum malaria. Proc Natl

Acad Sci USA. 2008; 105: 17097-17102.

133. Lovegrove FE, Tangpukdee N, Opoka RO, et al. Serum angiopoietin-1 and -2 levels

discriminate cerebral malaria from uncomplicated malaria and predict clinical outcome in

African children. PLoS ONE. 2009; 4: e4912.

134. Conroy AL, Lafferty EI, Lovegrove FE, et al. Whole blood angiopoietin-1 and -2 levels

discriminate cerebral and severe (non-cerebral) malaria from uncomplicated malaria. Malar

J. 2009; 8: 295.

135. Conroy AL, Phiri H, Hawkes M, et al. Endothelium-based biomarkers are associated with

cerebral malaria in Malawian children: a retrospective case-control study. PLoS ONE.

2010; 5: e15291.

136. Silver KL, Zhong K, Leke RGF, et al. Dysregulation of angiopoietins is associated with

placental malaria and low birth weight. PLoS ONE. 2010; 5: e9481.

137. Lieb W, Zachariah JP, Xanthakis V, et al. Clinical and genetic correlates of circulating

angiopoietin-2 and soluble Tie-2 in the community. Circ Cardiovasc Genet. 2010; 3: 300-

306.

138. David S, Kümpers P, Lukasz A, et al. Circulating angiopoietin-2 in essential hypertension:

relation to atherosclerosis, vascular inflammation, and treatment with

olmesartan/pravastatin. J Hypertens. 2009; 27: 1641-1647.

139. Nadar SK, Blann A, Beevers DG, Lip GYH. Abnormal angiopoietins 1&2, angiopoietin

receptor Tie-2 and vascular endothelial growth factor levels in hypertension: relationship

to target organ damage [a sub-study of the Anglo-Scandinavian Cardiac Outcomes Trial

(ASCOT)]. J Intern Med. 2005; 258: 336-343.

140. Singh H, Brindle NJP, Zammit VA. High glucose and elevated fatty acids suppress

signalling by the endothelium protective ligand angiopoietin-1. Microvasc Res. 2010; 79:

121-127.

115

141. Lim HS, Blann AD, Chong AY, et al. Plasma vascular endothelial growth factor,

angiopoietin-1, and angiopoietin-2 in diabetes. Diabetes Care. 2004; 27: 2918-2924.

142. Lip PL, Chatterjee S, Caine GJ, et al. Plasma vascular endothelial growth factor,

angiopoietin-2, and soluble angiopoietin receptor tie-2 in diabetic retinopathy: effects of

laser photocoagulation and angiotensin receptor blockade. Br J Ophthalmol. 2004; 88:

1543-1546.

143. Rasul S, Reiter MH, Ilhan A, et al. Circulating angiopoietin-2 and soluble Tie-2 in type 2

diabetes mellitus: a cross-sectional study. Cardiovasc Diabetol. 2011; 10: 55.

144. Lee KW, Lip GYH, Blann AD. Plasma angiopoietin-1, angiopoietin-2, angiopoietin

receptor Tie-2, and vascular endothelial growth factor levels in acute coronary syndromes.

Circulation. 2004; 110: 2355-2360.

145. Fam NP, Arab S, Billia F, et al. Increased myocardial expression of angiopoietin-2 in

patients undergoing urgent surgical revascularization for acute coronary syndromes. Can J

Cardiol. 2010; 26: 365-370.

146. Chong AY, Caine GJ, Freestone B, et al. Plasma angiopoietin-1, angiopoietin-2, and

angiopoietin receptor Tie-2 levels in congestive heart failure. J Am Coll Cardiol. 2004; 43:

423-428.

147. Iribarren C, Phelps BH, Darbinian JA, et al. Circulating angiopoietins-1 and -2,

angiopoietin receptor Tie2 and vascular endothelial growth factor-A as biomarkers of acute

myocardial infarction: a prospective nested case-control study. BMC Cardiovasc Dis.

2011; 11: 31.

148. Zhou HJ, Tang T, Guo C, et al. Expression of angiopoietin-1 and the receptor Tie-2

mRNA in rat brains following intracerebral hemorrhage. Acta Neurobiol Exp. 2008; 68:

147-154.

149. Sobrino T, Arias S, Rodriguez-Gonzalez R, et al. High serum levels of growth factors are

associated with good outcome in intracerebral hemorrhage. J Cereb Blood Flow Metab.

2009; 29: 1968-1974.

116

150. Fischer M, Broessner G, Dietmann A, et al. Angiopoietin-1 is associated with cerebral

vasospasm and delayed cerebral ischemia in subarachnoid hemorrhage. BMC Neurol.

2011; 11: 59.

151. Sobrino T, Millan M, Castellanos M, et al. Association of growth factors with arterial

recanalization and clinical outcome in patients with ischemic stroke treated with tPA. J

Thromb Haemost. 2010; 8: 1567-1574.

152. Chen J, Yu H, Sun K, et al. Promoter variant of angiopoietin-2 and plasma angiopoietin-2

are associated with risk of stroke recurrence in lacunar infarct patients. Biochem Biophys

Res Commun. 2010; 23: 212-216.

153. Nag S, Manias JL, Stewart DJ. Pathology and new players in the pathogenesis of brain

edema. Acta Neuropathol. 2009; 118: 197-217.

154. Davis B, Dei Cas A, Long DA, et al. Podocyte-specific expression of angiopoietin-2

causes proteinuria and apoptosis of glomerular endothelia. J Am Soc Nephrol. 2007; 18:

2320-2329.

155. Jung YJ, Kim DH, Lin AS, et al. Peritubular capillary preservation with COMP-

Angiopoietin-1 decreases ischemia-reperfusion-induced acute kidney injury. Am J Physiol

Renal Physiol. 2009; 297: F952-F960.

156. Lee S, Kim W, Kim DH, et al. Protective effect of COMP-Angiopoietin-1 on

cyclosporine-induced renal injury in mice. Nephrol Dial Transplant. 2008; 23: 2784-2794.

157. David S, Kümpers P, Hellpap J, et al. Angiopoietin 2 and cardiovascular disease in dialysis

and kidney transplantation. Am J Kidney Dis. 2009; 53: 770-778.

158. David S, Kümpers P, Lukasz A, et al. Circulating angiopoietin-2 levels increase with

progress of chronic kidney disease. Nephrol Dial Transplant. 2010; 25: 2571-2576.

159. David S, John SG, Jefferies HJ, et al. Angiopoietin-2 levels predict mortality in CKD

patients. Nephrol Dial Transplant. 2011; Oct 4. [Epub ahead of print].

117

160. Hirokoshi K, Maeshima Y, Kobayashi K, et al. Increase of serum angiopoietin-2 during

pregnancy is suppressed in women with preeclampsia. Am J Hypertens. 2005; 18: 1181-

1188.

161. Nadar SK, Karalis I, Al Yemeni E, et al. Plasma markers of angiogenesis in pregnancy

induced hypertension. Thromb Haemost. 2005; 94: 1071-1076.

162. Akolekar R, Casagrandi D, Skyfta E, et al. Maternal serum angiopoietin-2 at 11 to 13

weeks of gestation in hypertensive disorders of pregnancy. Prenat Diagn. 2009; 29: 847-

851.

163. Leinonen E, Wathen KA, Alfthan H, et al. Maternal serum angiopoietin-1 and -2 and Tie-2

in early pregnancy ending in preeclampsia or intrauterine growth retardation. J Clin

Endocrinol Metab. 2010; 95: 126-133.

164. Bolin M, Wiberg-Itzel E, Wikstrom AK, et al. Angiopoietin-1/Angiopoietin-2 ratio for

prediction of preeclampsia. Am J Hypertens. 2009; 22: 891-895.

165. Kamal M, El-Khayat W. Do serum angiopoietin-1, angiopoietin-2, and their receptor Tie-2

and 4G/5G variant of PAI-1 gene have a role in the pathogenesis of preeclampsia? J

Investig Med. 2011; 59: 1147-1150.

166. Gotsch F, Romero R, Kusanovic JP, et al. Preeclampsia and small-for-gestational age are

associated with decreased concentrations of a factor involved in angiogenesis: Soluble Tie-

2. J Matern Fetal Neonatal Med. 2008; 21: 389-402.

167. Pacora P, Romero R, Chaiworapongsa T, et al. Amniotic fluid angiopoietin-2 in term and

preterm parturition, and intra-amniotic infection/inflammation. J Perinat Med. 2009; 37:

503-511.

168. Coffelt SB, Tal AO, Scholz A, et al. Angiopoietin-2 regulates gene expression in Tie2-

expressing monocytes and augments their inherent proangiogenic functions. Cancer Res.

2010; 70: 5270-5280.

118

169. Maffei R, Martinelli S, Santachiara R, et al. Angiopoietin-2 plasma dosage predicts time to

first treatment and overall survival in chronic lymphocytic leukemia. Blood. 2010; 116:

584-592.

170. Nakayama T, Inaba M, Naito S, et al. Expression of angiopoietin-1, 2 and 4 and Tie-1 and

2 in gastrointestinal stromal tumor, leiomyoma and schwannoma. World J Gastroenterol.

2007; 13: 4473-4479.

171. Nakayama T, Hatachi G, Wen CY, et al. Expression and significance of Tie-1 and Tie-2

receptors, and antiopoietins-1, 2 and 4 in colorectal adenocarcinoma:

Immunohistochemical analysis and correlation with clinicopathological factors. World J

Gastroenterol. 2005; 11: 964-969.

172. Brown LF, Dezube BJ, Tognazzi K, et al. Expression of Tie1, Tie2, and angiopoietins 1, 2,

and 4 in Kaposi‟s sarcoma and cutaneous angiosarcoma. Am J Pathol. 2000; 156: 2179-

2183.

173. Schliemann C, Bieker R, Thoennissen N, et al. Circulating angiopoietin-2 is a strong

prognostic factor in acute myeloid leukemia. Leukemia. 2007; 21: 1901-1906.

174. Quartarone E, Alonci A, Allegra A, et al. Differential levels of soluble angiopoietin-2 and

Tie-2 in patients with haematological malignancies. Eur J Hematol. 2006; 77: 480-485.

175. Kastritis E, Roussou M, Michael M, et al. High levels of serum angiogenic growth factors

in patients with AL amyloidosis: comparisons with normal individuals and multiple

myeloma patients. Br J Haematol. 2010; 150: 587-591.

176. Jo MJ, Lee JH, Nam BH, et al. Preoperative serum angiopoietin-2 levels correlate with

lymph node status in patients with early gastric cancer. Ann Surg Oncol. 2009; 16: 2052-

2057.

177. Tomic TT, Gustavsson H, Wang W, et al. Castration resistant prostate cancer is associated

with increased blood vessel stabilization and elevated levels of VEGF and Ang-2. Prostate.

2011; Aug 1. [Epub ahead of print].

119

178. Schulz P, Fischer C, Detjen KM, et al. Angiopoietin-2 drives lymphatic metastasis of

pancreatic cancer. FASEB J. 2011; 25: 3325-3335.

179. Sallinen H, Heikura T, Laidinen S, et al. Preoperative angiopoietin-2 serum levels. A

marker of malignant potential in ovarian neoplasms and poor prognosis in epithelial

ovarian cancer. Int J Gynecol Cancer. 2010; 20: 1498-1505.

180. Volkova E, Willis JA, Wells JE, et al. Association of angiopoietin-2, C-reactive protein

and markers of obesity and insulin resistance with survival outcome in colorectal cancer.

Br J Cancer. 2011; 104: 51-59.

181. Srirajaskanthan R, Dancey G, Hackshaw A, et al. Circulating angiopoietin-2 is elevated in

patients with neuroendocrine tumours and correlates with disease burden and prognosis.

Endocr Relat Cancer. 2009; 16: 967-976.

182. Figueroa-Vega N, Diaz A, Adrados M, et al. The association of the angiopoietin/Tie-2

system with the development of metastasis and leukocyte migration in neuroendocrine

tumors. Endocr Relat Cancer. 2010; 17: 897-908.

183. Park JH, Park KJ, Kim YS, et al. Serum angiopoietin-2 as a clinical marker for lung

cancer. Chest. 2007; 132: 200-206.

184. Helfrich I, Edler L, Sucker A, et al. Angiopoietin-2 levels are associated with disease

progression in metastatic malignant melanoma. Clin Cancer Res. 2009; 15: 1384-1392.

185. Scholz A, Rehm VA, Rieke S, et al. Angiopoietin-2 serum levels are elevated in patients

with liver cirrhosis and hepatocellular carcinoma. Am J Gastroenterol. 2007; 102: 2471-

2481.

186. DuBois S, Demetri G. Markers of angiogenesis and clinical features in patients with

sarcoma. Cancer. 2007; 109: 813-819.

187. Caine GJ, Blann AD, Stonelake PS, et al. Plasma angiopoietin-1, angiopoietin-2 and Tie-2

breast and prostate cancer: a comparison with VEGF and Flt-1. Eur J Clin Invest. 2003; 33:

883-890.

120

188. van der Veldt AAM, Vroling L, de Haas RR, et al. Sunitinib-induced changes in

circulating endothelial cell-related proteins in patients with metastatic renal cell cancer. Int

J Cancer. 2011; Sept 22. [Epub ahead of print].

189. Cheng CL, Hou HA, Jhuang JY, et al. High bone marrow angiopoietin-1 expression is an

independent poor prognostic factor for survival in patients with myelodysplastic

syndromes. Br J Cancer. 2011; 105: 975-982.

190. Gallagher DC, Bhatt RS, Parikh SM, et al. Angiopoietin 2 is a potential mediator of high-

dose interleukin 2-induced vascular leak. Clin Cancer Res. 2007; 13: 2115-2120.

191. Huang J, Bae JO, Tsai JP, et al. Angiopoietin-1/Tie-2 activation contributes to vascular

survival and tumor growth during VEGF blockade. Int J Oncol. 2009; 34: 79-87.

192. Zanini A, Chetta A, Imperatori AS, et al. The role of the bronchial microvasculature in the

airway remodelling in asthma and COPD. Respir Res. 2010; 11: 132.

193. Makinde TO, Agrawal DK. Increased expression of angiopoietins and Tie2 in the lungs of

chronic asthmatic mice. Am J Respir Cell Mol Biol. 2011; 44: 384-393.

194. Kanazawa H, Asai K, Tochino Y, et al. Increased levels of angiopoietin-2 in induced

sputum from smoking asthmatic patients. Clin Exp Allergy. 2009; 39: 1330-1337.

195. Cho YJ, Ma JE, Yun EY, et al. Serum angiopoietin-2 levels are elevated during acute

exacerbations of COPD. Respirology. 2011; 16: 284-290.

196. Lee KY, Lee KS, Park SJ, et al. Clinical significance of plasma and serum vascular

endothelial growth factor in asthma. J Asthma. 2008; 45: 735-739.

197. Kugathasan L, Ray JB, Deng Y, et al. The angiopoietin-1-Tie2 pathway prevents rather

than promotes pulmonary arterial hypertension in transgenic mice. J Exp Med. 2009; 206:

2221-2234.

198. Kugathasan L, Dutly AE, Zhao YD, et al. Role of angiopoietin-1 in experimental and

human pulmonary arterial hypertension. Chest. 2005; 128: 633S-642S.

121

199. Hiremath J, Thanikachalam S, Parikh K, et al. Exercise improvement and plasma

biomarker changes with intravenous treprostinil therapy for pulmonary arterial

hypertension: A placebo-controlled trial. J Heart Lung Transplant. 2010; 29: 137-149.

200. Kümpers P, Nickel N, Lukasz A, et al. Circulating angiopoietins in idiopathic pulmonary

arterial hypertension. Eur Heart J. 2010; 31: 2291-2300.

201. Westra J, de Groot L, Plaxton SL, et al. Angiopoietin-2 is highly correlated with

inflammation and disease activity in recent-onset rheumatoid arthritis and could be

predictive for cardiovascular disease. Rheumatology. 2011; 50: 665-673.

202. Kümpers P, David S, Haubitz M, et al. The Tie2 receptor antagonist angiopoietin 2

facilitates vascular inflammation in systemic lupus erythematosus. Ann Rheum Dis. 2009;

68: 1638-1643.

203. Breunis WB, Davila S, Shimizu C, et al. Disruption of vascular homeostasis in Kawasaki

disease patients: Involvement of vascular endothelial growth factor and angiopoietins.

Arthritis Rheum. 2011; Sep 8. [Epub ahead of print].

204. Kümpers P, Hellpap J, David S, et al. Circulating angiopoietin-2 is a marker and potential

mediator of endothelial cell detachment in ANCA-associated vasculitis with renal

involvement. Nephrol Dial Transplant. 2009; 24: 1845-1850.

205. Figueroa-Vega N, Alfonso-Perez M, Cuesta-Mateos C, et al. Tie-2 is overexpressed by

monocytes in autoimmune thyroid disorders and participates in their recruitment to the

thyroid gland. J Clin Endocrinol Metab. 2009; 94: 2626-2633.

206. Ishikawa A, Okada J, Nishi K, Hirohata S. Efficacy of serum angiopoietin-1 measurement

in the diagnosis of early rheumatoid arthritis. Clin Exp Rheumatol. 2011; 29: 604-608.

207. Choe JY, Park SH, Kim SK. Serum angiopoietin-1 level is increased in patients with

Behçet‟s disease. Joint Bone Spine. 2010; 77: 340-344.

208. Ganter MT, Cohen MJ, Brohi K, et al. Angiopoietin-2, marker and mediator of endothelial

activation with prognostic significance early after trauma? Ann Surg. 2008; 247: 320-326.

122

209. Mohan JS, Lip PL, Blann AD, et al. The angiopoietin/Tie-2 system in proliferative sickle

retinopathy: relation to vascular endothelial growth factor, its soluble receptor Flt-1 and

von Willebrand factor, and to the effects of laser treatment. Br J Ophthalmol. 2005; 89:

815-819.

210. Duits AJ, Rodriquez T, Schnog JJB for the CURAMA study group. Serum levels of

angiogenic factors indicate a pro-angiogenic state in adults with sickle cell disease. Br J

Haematol. 2006; 134: 116-119.

211. Koutroubakis IE, Xidakis C, Karmiris K, et al. Potential role of soluble angiopoietin-2 and

Tie-2 in patients with inflammatory bowel disease. Eur J Clin Invest. 2006; 36: 127-132.

212. Yoshizaki A, Nakayama N, Naito S, Sekine I. Expression patterns of angiopoietin-1, -2,

and Tie-2 receptor in ulcerative colitis support involvement of the angiopoietin/Tie

pathway in the progression of ulcerative colitis. Dig Dis Sci. 2009; 54: 2094-2099.

213. Whitcomb DC, Muddana V, Langmead CJ, et al. Angiopoietin-2, a regulator of vascular

permeability in inflammation, is associated with persistent organ failure in patients with

acute pancreatitis from the United States and Germany. Am J Gastroenterol. 2010; 105:

2287-2292.

214. Vespasiani-Gentilucci U, Galati G, Mazzarelli C, et al. Angiogenic cytokines in patients

undergoing antiviral treatment for chronic hepatitis C virus infection. J Interferon Cytokine

Res. 2011; 31: 207-210.

215. Li Y, Chen J, Wu C, et al. Hepatitis B virus/hepatitis C virus upregulate angiopoietin-2

expression through mitogen-activated protein kinase pathway. Hepatol Res. 2010; 40:

1022-1033.

216. McCarter SD, Mei SHJ, Lai PFH, et al. Cell-based angiopoietin-1 gene therapy for acute

lung injury. Am J Respir Crit Care Med. 2007; 175: 1014-1026.

217. Zhao YD, Campbell AIM, Robb M, et al. Protective role of angiopoietin-1 in experimental

pulmonary hypertension. Circ Res. 2003; 92: 984-991.

123

218. Van Slyke P, Alami J, Martin D, et al. Acceleration of diabetic wound healing by an

angiopoietin peptide mimetic. Tissue Eng Part A. 2009; 15: 1269-1280.

219. Jeong BC, Kim HJ, Bae IH, et al. COMP-Ang1, a chimeric form of angiopoietin 1,

enhances BMP2-induced osteoblast differentiation and bone formation. Bone. 2010; 46:

479-486.

220. Klopper J, Lindenmaier W, Fiedler U, et al. High efficient adenoviral-mediated VEGF and

Ang-1 gene delivery into osteogenically differentiated human mesenchymal stem cells.

Microvasc Res. 2008; 75: 83-90.

221. Su H, Takagawa J, Huang Y, et al. Additive effect of AAV-mediated angiopoietin-1 and

VEGF expression on the therapy of infarcted heart. Int J Cardiol. 2009; 133: 191-197.

222. Park BH, Jang KY, Kim KH, et al. COMP-Angiopoietin-1 ameliorates surgery-induced

ischemic necrosis of the femoral head in rats. Bone. 2009; 44: 886-892.

223. Piao W, Wang H, Inoue M, et al. Transplantation of sendai viral angiopoietin-1-modified

mesenchymal stem cells for ischemic limb disease. Angiogenesis. 2010; 13: 203-210.

224. Shin HY, Lee YJ, Kim HJ, et al. Protective role of COMP-Ang1 in ischemic rat brain. J

Neurosci Res. 2010; 88: 1052-1063.

225. Cho CH, Kammerer RA, Lee HJ, et al. Designed angiopoietin-1 variant, COMP-Ang1,

protects against radiation-induced endothelial cell apoptosis. Proc Natl Acad Sci USA.

2004; 101: 5553-5558.

226. Han S, Arnold SA, Sithu SD, et al. Rescuing vasculature with intravenous angiopoietin-1

and αvβ3 integrin peptide is protective after spinal cord injury. Brain. 2010; 133: 1026-

1042.

227. Reikvam H, Hatfield KJ, Lassalle P, et al. Targeting the angiopoietin (Ang)/Tie-2 pathway

in the cross-talk between acute myeloid leukaemia and endothelial cells: Studies of Tie-2

blocking antibodies, exogenous Ang-2 and inhibition of constitutive agonistic Ang-1

release. Expert Opin Investig Drugs. 2010; 19: 169-183.

124

228. Falcon BL, Hashizume H, Koumoutsakos P, et al. Contrasting actions of selective

inhibitors of angiopoietin-1 and angiopoietin-2 on the normalization of tumor blood

vessels. Am J Pathol. 2009; 175: 2159-2170.

229. Hegeman MA, Hennus MP, van Meurs M, et al. Angiopoietin-1 treatment reduces

inflammation but does not prevent ventilator-induced lung injury. PLoS ONE. 2010; 5:

e15653.

230. Mita AC, Takimoto CH, Mita M, et al. Phase 1 study of AMG 386, a selective

angiopoietin 1/2-neutralizing peptibody, in combination with chemotherapy in adults with

advanced solid tumors. Clin Cancer Res. 2010; 16: 3044-3056.

231. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of

effective antimicrobial therapy is the critical determinant of survival in human septic

shock. Crit Care Med. 2006; 34: 1589-1596.

232. Gaieski DF, Mikkelsen ME, Band RA, et al. Impact of time to antibiotics on survival in

patients with severe sepsis or septic shock in whom early goal-directed therapy was

initiated in the emergency department. Crit Care Med. 2010; 38: 1045-1053.

233. Gilbert DN. Procalcitonin as a biomarker in respiratory tract infection. Clin Infect Dis.

2011; 52: S346-S356.

234. Delèvaux I, André M, Colombier M, et al. Can procalcitonin measurement help in

differentiating between bacterial infection and other kinds of inflammatory processes? Ann

Rheum Dis. 2003; 62: 337-340.

235. Thia KT, Chan ES, Ling KL, et al. Role of procalcitonin in infectious gastroenteritis and

inflammatory bowel disease. Dig Dis Sci. 2008; 53: 2960-2968.

236. Linscheid P, Seboek D, Schaer DJ, et al. Expression and secretion of procalcitonin and

calcitonin gene-related peptide by adherent monocytes and by macrophage-activated

adipocytes. Crit Care Med. 2004; 32: 1715-1721.

125

237. Linscheid P, Seboek D, Nylen ES, et al. In vitro and in vivo calcitonin I gene expression in

parenchymal cells: a novel product of human adipose tissue. Endocrinology. 2003; 144:

5578-5584.

238. Wiedermann FJ, Kaneider N, Egger P, et al. Migration of human monocytes in response to

procalcitonin. Crit Care Med. 2002; 30: 1112-1117.

239. Liappis AP, Gibbs KW, Nylen ES, et al. Exogenous procalcitonin evokes a pro-

inflammatory cytokine response. Inflamm Res. 2011; 60: 203-207.

240. Nylen ES, Whang KT, Snider RH, et al. Mortality is increased by procalcitonin and

decreased by an antiserum reactive to procalcitonin in experimental sepsis. Crit Care Med.

1998; 26: 1001-1006.

241. Simon L, Gauvin F, Amre DK, et al. Serum procalcitonin and C-reactive protein levels as

markers of bacterial infection: A systematic review and meta-analysis. Clin Infect Dis.

2004; 39: 206-217.

242. Jones AE, Fiechtl JF, Brown MD, et al. Procalcitonin test in the diagnosis of bacteremia: a

meta-analysis. Ann Emerg Med. 2007; 50: 34-41.

243. Charles PE, Ladoire S, Aho S, et al. Serum procalcitonin elevation in critically ill patients

at the onset of bacteremia caused by either Gram negative or Gram positive bacteria. BMC

Infect Dis. 2008; 8: 38.

244. Charles PE, Tinel C, Barbar S, et al. Procalcitonin kinetics within the first days of sepsis:

relationship with the appropriateness of antibiotic therapy and the outcome. Crit Care.

2009; 13: R38.

245. Bafadhel M, Clark TW, Reid C, et al. Procalcitonin and C-reactive protein in hospitalized

adult patients with community-acquired pneumonia or exacerbation of asthma or COPD.

Chest. 2011; 139: 1410-1418.

246. Cuquemelle E, Soulis F, Villers D, et al. Can procalcitonin help identify associated

bacterial infection in patients with severe influenza pneumonia? A multicentre study.

Intensive Care Med. 2011; 37: 796-800.

126

247. Nakamura A, Wada H, Ikejiri M, et al. Efficacy of procalcitonin in the early diagnosis of

bacterial infections in a critical care unit. Shock. 2009; 31: 586-591.

248. Uzzan B, Cohen R, Nicolas P, et al. Procalcitonin as a diagnostic test for sepsis in

critically ill adults after surgery or trauma: A systematic review and meta-analysis. Crit

Care Med. 2006; 34: 1996-2003.

249. Tang BM, Eslick GD, Craig JC, McLean AS. Accuracy of procalcitonin for sepsis

diagnosis in critically ill patients: systematic review and meta-analysis. Lancet Infect Dis.

2007; 7: 210-217.

250. Uusitalo-Seppala R, Koskinen P, Leino A, et al. Early detection of severe sepsis in the

emergency room: Diagnostic value of plasma C-reactive protein, procalcitonin, and

interleukin-6. Scand J Infect Dis. 2011; Sep 6. [Epub ahead of print].

251. Nyamande N, Lalloo UG. Serum procalcitonin distinguishes CAP due to bacteria,

Mycobacterium tuberculosis and PJP. Int J Tuberc Lung Dis. 2006; 10: 510-515.

252. Schleicher GK, Herbert V, Brink A, et al. Procalcitonin and C-reactive protein levels in

HIV-positive subjects with tuberculosis and pneumonia. Eur Respir J. 2005; 25: 688-692.

253. Gerard Y, Hober D, Assicot M, et al. Procalcitonin as a marker of bacterial sepsis in

patients infected with HIV-1. J Infect. 1997; 35: 41-46.

254. Bloos F, Marshall JC, Dellinger RP, et al. Multinational, observational study of

procalcitonin in ICU patients with pneumonia requiring mechanical ventilation: a

multicenter observational study. Crit Care. 2011; 15: R88.

255. Boeck L, Eggimann P, Smyrnios N, et al. Midregional pro-atrial natriuretic peptide and

procalcitonin improve survival prediction in VAP. Eur Respir J. 2011; 37: 595-603.

256. Boussekey N, Leroy O, Georges H, et al. Diagnostic and prognostic values of admission

procalcitonin levels in community-acquired pneumonia in an intensive care unit. Infection.

2005; 33: 257-263.

127

257. Clec‟h C, Ferriere F, Karoubi P, et al. Diagnostic and prognostic value of procalcitonin in

patients with septic shock. Crit Care Med. 2004; 32: 1166-1169.

258. Dahaba AA, Hagara B, Fall A, et al. Procalcitonin for early prediction of survival outcome

in postoperative critically ill patients with severe sepsis. Br J Anaesth. 2006; 97: 503-508.

259. Giamarellos-Bourboulis EJ, Tsangaris I, Kanni T, et al. Procalcitonin as an early indicator

of outcome in sepsis: a prospective observational study. J Hosp Infect. 2011; 77: 58-63.

260. Gibot S, Cravoisy A, Kolopp-Sarda MN, et al. Time-course of sTREM (soluble triggering

receptor expressed on myeloid cells)-1, procalcitonin, and C-reactive protein plasma

concentrations during sepsis. Crit Care Med. 2005; 33: 792-796.

261. Guan J, Lin Z, Lue H. Dynamic change of procalcitonin, rather than concentration itself, is

predictive of survival in septic shock patients when beyond 10 ng/mL. Shock. 2011; Sep 7.

[Epub ahead of print].

262. Hillas G, Vassilakopoulos T, Plantza R, et al. C-reactive protein and procalcitonin as

predictors of survival and septic shock in ventilator-associated pneumonia. Eur Respir J.

2010; 35: 805-811.

263. Huang DT, Weissfeld LA, Kellum JA, et al. Risk prediction with procalcitonin and clinical

rules in community-acquired pneumonia. Ann Emerg Med. 2008; 52: 48-58.

264. Karlsson S, Heikkinen M, Pettilä V, et al. Predictive value of procalcitonin decrease in

patients with severe sepsis: a prospective observational study. Crit Care. 2010; 14: R205.

265. Lee CC, Chen SY, Tsai CL, et al. Prognostic value of Mortality in Emergency Department

Sepsis score, procalcitonin, and C-reactive protein in patients with sepsis at the Emergency

Department. Shock. 2008; 29: 322-327.

266. Menéndez R, Martínez R, Reyes S, et al. Biomarkers improve mortality prediction by

prognostic scales in community-acquired pneumonia. Thorax. 2009; 64: 587-591.

267. Meng FS, Su L, Tang YQ, et al. Serum procalcitonin at the time of admission to the ICU

as a predictor of short-term mortality. Clin Biochem. 2009; 42: 1025-1031.

128

268. Novotny A, Emmanuel K, Matevossian E, et al. Use of procalcitonin for early prediction

of lethal outcome of postoperative sepsis. Am J Surg. 2007; 194: 35-39.

269. Oberholzer A, Souza SM, Tschoeke SK, et al. Plasma cytokine measurements augment

prognostic scores as indicators of outcome in patients with severe sepsis. Shock. 2005; 23:

488-493.

270. Phua J, Koay ESC, Lee KH. Lactate, procalcitonin and amino-terminal pro-B-type

natriuretic peptide versus cytokine measurements and clinical severity scores for

prognostication in septic shock. Shock. 2008; 29: 328-333.

271. Rau BM, Frigerio I, Büchler MW, et al. Evaluation of procalcitonin for predicting septic

multiorgan failure and overall prognosis in secondary peritonitis: a prospective,

international multicenter study. Arch Surg. 2007; 142: 134-142.

272. Ruiz-Alvarez MJ, Garcia-Valdecasas S, De Pablo R, et al. Diagnostic efficacy and

prognostic value of serum procalcitonin concentration in patients with suspected sepsis. J

Intensive Care Med. 2009; 24: 63-71.

273. van Langevelde P, Joop K, van Loon J, et al. Endotoxin, cytokines, and procalcitonin in

febrile patients admitted to the hospital: Identification of subjects at high risk of mortality.

Clin Infect Dis. 2000; 31: 1343-1348.

274. Viallon A, Guyomarc‟h S, Marjollet O, et al. Can emergency physicians identify a high

mortality subgroup of patients with sepsis: role of procalcitonin. Eur J Emerg Med. 2008;

15: 26-33.

275. Wunder C, Eichelbronner O, Roewer N. Are IL-6, IL-10 and PCT plasma concentrations

reliable for outcome prediction in severe sepsis? A comparison with APACHE III and

SAPS II. Inflamm Res. 2004; 53: 158-163.

276. Jensen JU, Heslet L, Jensen TH, et al. Procalcitonin increase in early identification of

critically ill patients at high risk of mortality. Crit Care Med. 2006; 34: 2596-2602.

277. Schneider CP, Yilmaz Y, Kleespies A, et al. Accuracy of procalcitonin for outcome

prediction in unselected postoperative critically ill patients. Shock. 2009; 31: 568-573.

129

278. Carrol ED, Newland P, Thomson APJ, Hart CA. Prognostic value of procalcitonin in

children with meningococcal sepsis. Crit Care Med. 2005; 33: 224-225.

279. Sauer M, Tiede K, Fuchs D, et al. Procalcitonin, C-reactive protein, and endotoxin after

bone marrow transplantation: identification of children at high risk of morbidity and

mortality from sepsis. Bone Marrow Transplant. 2003; 31: 1137-1142.

280. Hatherill M, Tibby SM, Turner C, et al. Procalcitonin and cytokine levels; relationship to

organ failure and mortality in pediatric septic shock. Crit Care Med. 2000; 28: 2591-2594.

281. Vouloumanou EK, Plessa E, Karageorgopoulos DE, et al. Serum procalcitonin as a

diagnostic marker for neonatal sepsis: a systematic review and meta-analysis. Intensive

Care Med. 2011; 37: 747-762.

282. Carrol ED, Mankhambo LA, Jeffers G, et al. The diagnostic and prognostic accuracy of

five markers of serious bacterial infection in Malawian children with signs of severe

infection. PLoS ONE. 2009; 4: e6621.

283. Diez-Padrisa N, Bassat Q, Machevo S, et al. Procalcitonin and C-reactive protein for

invasive bacterial pneumonia diagnosis among children in Mozambique, a malaria-

endemic area. PLoS ONE. 2010; 5: e13226.

284. Bajwa EK, Khan UA, Januzzi JL, et al. Plasma C-reactive protein levels are associated

with improved outcome in ARDS. Chest. 2009; 136: 471-480.

285. Zanzinger K, Schellack C, Nausch N, Cerwenka A. Regulation of triggering receptor

expressed on myeloid cells 1 expression on mouse inflammatory monocytes. Immunology.

2009; 128: 185-915.

286. Lagler H, Sharif O, Haslinger I, et al. TREM-1 activation alters the dynamics of

pulmonary IRAK-M expression in vivo and improves host defense during pneumococcal

pneumonia. J Immunol. 2009; 183: 2027-2036.

287. Gibot S, Massin F, Marcou M, et al. TREM-1 promotes survival during septic shock in

mice. Eur J Immunol. 2007; 37: 456-466.

130

288. Bouchon A, Facchetti F, Weigand MA, Colonna M. TREM-1 amplifies inflammation and

is a crucial mediator of septic shock. Nature. 2001; 410: 1103-1107.

289. Zhang J, She D, Feng D, et al. Dynamic changes of serum soluble triggering receptor

expressed on myeloid cells-1 (sTREM-1) reflect sepsis severity and can predict prognosis:

a prospective study. BMC Infect Dis. 2011; 11: 53.

290. Dimopoulou I, Orfanos SE, Pelekanou A, et al. Serum of patients with septic shock

stimulates the expression of Trem-1 on U937 monocytes. Inflamm Res. 2009; 58: 127-132.

291. Bopp C, Hofer S, Bouchon A, et al. Soluble TREM-1 is not suitable for distinguishing

between systemic inflammatory response syndrome and sepsis survivors and nonsurvivors

in the early stage of acute inflammation. Eur J Anaesthesiol. 2009; 26: 504-507.

292. Gibot S, Kolopp-Sarda NM, Bene MC, et al. Plasma level of a triggering receptor

expressed on myeloid cells-1: Its diagnostic accuracy in patients with suspected sepsis.

Ann Intern Med. 2004; 141: 9-15.

293. Kwofie L, Rapoport BL, Fickl H, et al. Evaluation of circulating soluble triggering

receptor expressed on myeloid cells-1 (sTREM-1) to predict risk profile, response to

antimicrobial therapy, and development of complications in patients with chemotherapy-

associated febrile neutropenia: a pilot study. Ann Hematol. 2011; Oct 6. [Epub ahead of

print].

294. Lin CH, Yao M, Hsu SC, et al. Soluble triggering receptor expressed on myeloid cells-1 as

an infection marker for patients with neutropenic fever. Crit Care Med. 2011; 39: 993-999.

295. Porfyridis I, Plachouras D, Karagianni V, et al. Diagnostic value of triggering receptor

expressed on myeloid cells-1 and C-reactive protein for patients with lung infiltrates: an

observational study. BMC Infect Dis. 2010; 10: 286.

296. Ruiz-Gonzalez A, Esquerda A, Falguera M, et al. Triggering receptor (TREM-1) expressed

on myeloid cells predicts bacteremia better than clinical variables in community-acquired

pneumonia. Respirology. 2011; 16: 321-325.

131

297. Barati M, Bashar FR, Shahrami R, et al. Soluble triggering receptor expressed on myeloid

cells 1 and the diagnosis of sepsis. J Crit Care. 2010; 25: 362.e1-6.

298. Nielsen AR, Plomgaard P, Krabbe KS, et al. IL-6, but not TNF-α, increases plasma YKL-

40 in human subjects. Cytokine. 2011; 55: 152-155.

299. Nordenbaek C, Johansen JS, Junker P, et al. YKL-40, a matrix protein of specific granules

in neutrophils, is elevated in serum of patients with community-acquired pneumonia

requiring hospitalization. J Infect Dis. 1999; 180: 1722-1726.

300. Kronborg G, Ostergaard C, Weis N, et al. Serum level of YKL-40 is elevated in patients

with Streptococcus pneumoniae bacteremia and is associated with the outcome of the

disease. Scand J Infect Dis. 2002; 34: 323-326.

301. Johansen JS, Krabbe KS, Moller K, Pedersen BK. Circulating YKL-40 levels during

human endotoxaemia. Clin Exp Immunol. 2005; 140: 343-348.

302. Hattori N, Oda S, Sadahiro T, et al. YKL-40 identified by proteomic analysis as a

biomarker of sepsis. Shock. 2009; 32: 393-400.

303. Punyadeera C, Schneider EM, Schaffer D, et al. A biomarker panel to discriminate

between systemic inflammatory response syndrome and sepsis and sepsis severity. J

Emerg Trauma Shock. 2010; 3: 26-35.

304. Ng PC, Li K, Chui KM, et al. IP-10 is an early diagnostic marker for identification of late-

onset bacterial infection in preterm infants. Pediatr Res. 2007; 61: 93-98.

305. Erdman LK, Dhabangi A, Musoke C, et al. Combinations of host biomarkers predict

mortality among Ugandan children with severe malaria: a retrospective case-control study.

PLoS ONE. 2011; 6: e17440.

306. Nie CQ, Bernard NJ, Norman MU, et al. IP-10-mediated T cell homing promotes cerebral

inflammation over splenic immunity to malaria infection. PLoS Pathog. 2009; 5:

e1000369.

132

307. Chen DY, Shen GH, Chen YM, et al. Interferon-inducible protein-10 as a marker to detect

latent and active tuberculosis in rheumatoid arthritis. Int J Tuberc Lung Dis. 2011; 15: 192-

199.

308. Dheda K, Van-Zyl Smit RN, Sechi LA, et al. Clinical diagnostic utility of IP-10 and LAM

antigen levels for the diagnosis of tuberculous pleural effusions in a high burden setting.

PLoS ONE. 2009; 4: e4689.

309. Wang H, Yue J, Yang J, et al. Clinical diagnostic utility of adenosine deaminase,

Interferon-γ, Interferon-γ-induced protein of 10 kDa, and dipeptidyl peptidase 4 levels in

tuberculous pleural effusions. Heart Lung. 2011; Sep 13. [Epub ahead of print].

310. Ruhwald M, Bodmer T, Maier C, et al. Evaluating the potential of IP-10 and MCP-2 as

biomarkers for the diagnosis of tuberculosis. Eur Respir J. 2008; 32: 1607-1615.

311. Nakamura T, Ebihara I, Shoji H, et al. Treatment with polymyxin B-immobilized fiber

reduces platelet activation in septic shock patients: Decrease in plasma levels of soluble P-

selectin, platelet factor 4 and β-thromboglobulin. Inflamm Res. 1999; 48: 171-175.

312. Kowalska MA, Mahmud SA, Lambert MP, et al. Endogenous platelet factor 4 stimulates

activated protein C generation in vivo and improves survival after thrombin or

lipopolysaccharide challenge. Blood. 2007; 110: 1903-1905.

313. Slungaard A. Platelet factor 4 modulation of the thrombomodulin-protein C system. Crit

Care Med. 2004; 32: S331-S335.

314. Lorenz R, Brauer M. Platelet factor 4 (PF4) in septicaemia. Infection. 1988; 16: 273-276.

315. Patel KN, Soubra SH, Bellera RV, et al. Differential role of von Willebrand factor and P-

selectin on microvascular thrombosis in endotoxemia. Arterioscler Thromb Vasc Biol.

2008; 28: 2225-2230.

316. Patel KN, Soubra SH, Lam FW, et al. Polymicrobial sepsis and endotoxemia promote

microvascular thrombosis via distinct mechanisms. J Thromb Haemost. 2010; 8: 1403-

1409.

133

317. van Mourik JA, Boertjes R, Huisveld IA, et al. von Willebrand Factor propeptide in

vascular disorders: A tool to distinguish between acute and chronic endothelial cell

perturbation. Blood. 1999; 94: 179-185.

318. Claus RA, Bockmeyer CL, Budde U, et al. Variations in the ratio between von Willebrand

factor and its cleaving protease during systemic inflammation and association with severity

and prognosis of organ failure. Thromb Haemost. 2009; 101: 239-247.

319. Kremer Hovinga JA, Zeerleder S, Kessler P, et al. ADAMTS-13, von Willebrand factor

and related parameters in severe sepsis and septic shock. J Thromb Haemost. 2007; 5:

2284-2290.

320. van der Heijden M, van Nieuw Amerongen GP, Koolwijk P, et al. Angiopoietin-2,

permeability oedema, occurrence and severity of ALI/ARDS in septic and non-septic

critically ill patients. Thorax. 2008; 63: 903-909.

321. Bajaj MS, Tricomi SM. Plasma levels of the three endothelial-specific proteins von

Willebrand factor, tissue factor pathway inhibitor, and thrombomodulin do not predict the

development of acute respiratory distress syndrome. Intensive Care Med. 1999; 25: 1259-

1266.

322. Moss M, Ackerson L, Gillespie MK, et al. von Willebrand factor antigen levels are not

predictive for the Adult Respiratory Distress Syndrome. Am J Respir Crit Care Med. 1995;

151: 15-20.

323. Rubin DB, Wiener-Kronish JP, Murray JF, et al. Elevated von Willebrand Factor antigen

is an early plasma predictor of acute lung injury in nonpulmonary sepsis syndrome. J Clin

Invest. 1990; 86: 474-480.

324. Ware LB, Conner ER, Matthay MA. von Willebrand factor antigen is an independent

marker of poor outcome in patients with early acute lung injury. Crit Care Med. 2001; 29:

2325-2331.

325. DiStasi MR, Ley K. Opening the flood-gates: how neutrophil-endothelial interactions

regulate permeability. Trends Immunol. 2009; 30: 547-556.

134

326. Kayal S, Jaïs JP, Aguini N, et al. Elevated circulating E-selectin, intercellular adhesion

molecule 1, and von Willebrand factor in patients with severe infection. Am J Respir Crit

Care Med. 1998; 157: 776-784.

327. Sessler CN, Windsor AC, Schwartz M, et al. Circulating ICAM-1 is increased in septic

shock. Am J Respir Crit Care Med. 1995; 151: 1420-1427.

328. Shin JY, Yoon IH, Kim JS, et al. Vascular endothelial growth factor-induced chemotaxis

and IL-10 from T cells. Cell Immunol. 2009; 256: 72-78.

329. Yano K, Liaw PC, Mullington JM, et al. Vascular endothelial growth factor is an

important determinant of sepsis morbidity and mortality. J Exp Med. 2006; 203: 1447-

1458.

330. Pickkers P, Sprong T, van Eijk L, et al. Vascular endothelial growth factor is increased

during the first 48 hours of human septic shock and correlates with vascular permeability.

Shock. 2005; 24: 508-512.

331. Shapiro NI, Yano K, Okada H, et al. A prospective, observational study of soluble Flt-1

and vascular endothelial growth factor in sepsis. Shock. 2008; 29: 452-457.

332. Alves BE, Montalvao SA, Aranha FJ, et al. Time-course of sFlt-1 and VEGF-A release in

neutropenic patients with sepsis and septic shock: a prospective study. J Transl Med. 2011;

9: 23.

333. van der Flier M, van Leeuwen HJ, van Kessel KP, et al. Plasma Vascular endothelial

growth factor in severe sepsis. Shock. 2005; 23: 35-38.

334. Karlsson S, Pettila V, Tenhunen J, et al. Vascular endothelial growth factor in severe

sepsis and septic shock. Anesth Analg. 2008; 106: 1820-1826.

335. Harrington LS, Sainson RC, Williams CK, et al. Regulation of multiple angiogenic

pathways by Dll4 and Notch in human umbilical vein endothelial cells. Microvasc Res.

2008; 75: 144-154.

336. Pierrakos C, Vincent JL. Sepsis biomarkers: a review. Crit Care. 2010; 14: R15.

135

337. Bozza FA, Salluh JI, Japiassu AM, et al. Cytokine profiles as markers of disease severity

in sepsis: a multiplex analysis. Crit Care. 2007; 11: R49.

338. Kinasewitz GT, Yan SB, Basson B, et al. Universal changes in biomarkers of coagulation

and inflammation occur in patients with severe sepsis, regardless of causative micro-

organism. Crit Care. 2004; 8: R82-R90.

339. Reinhart K, Karzai W. Anti-tumor necrosis factor therapy in sepsis: Update on clinical

trials and lessons learned. Crit Care Med. 2001; 29: S121-S125.

340. Russell JA. Management of sepsis. New Engl J Med. 2006; 355: 1699-1713.

341. Cummings CJ, Sessler CN, Beall LD, et al. Soluble E-selectin levels in sepsis and critical

illness. Am J Respir Crit Care Med. 1997; 156: 431-437.

342. Megarbane B, Marchal P, Marfaing-Koka A, et al. Increased diffusion of soluble adhesion

molecules in meningitis, severe sepsis and systemic inflammatory response without

neurological infection is associated with intrathecal shedding in cases of meningitis.

Intensive Care Med. 2004; 30: 867-874.

343. Whalen MJ, Doughty LA, Carlos TM, et al. Intercellular adhesion molecule-1 and vascular

cell adhesion molecule-1 are increased in the plasma of children with sepsis-induced

multiple organ failure. Crit Care Med. 2000; 28: 2600-2607.

344. Froon AH, Bonten MJ, Gaillard CA, et al. Prediction of clinical severity and outcome of

ventilator-associated pneumonia. Comparison of simplified acute physiology score with

systemic inflammatory mediators. Am J Respir Crit Care Med. 1998; 158: 1026-1031.

345. Patschan SA, Patschan D, Temme J, et al. Endothelial progenitor cells (EPC) in sepsis

with acute renal dysfunction (ARD). Crit Care. 2011; 15: R94.

346. Rafat N, Hanusch C, Brinkkoetter PT, et al. Increased circulating endothelial progenitor

cells in septic patients: Correlation with survival. Crit Care Med. 2007; 35: 1677-1684.

347. Cribbs SK, Martin GS, Rojas M. Monitoring of endothelial dysfunction in critically ill

patients: the role of endothelial progenitor cells. Curr Opin Crit Care. 2008; 14: 354-360.

136

348. Agouni A, Lagrue-Lak-Hal AH, Ducluzeau PH, et al. Endothelial dysfunction caused by

circulating microparticles from patients with metabolic syndrome. Am J Pathol. 2008; 173:

1210-1219.

349. Mostefai HA, Meziani F, Mastronardi ML, et al. Circulating microparticles from patients

with septic shock exert protective role in vascular function. Am J Respir Crit Care Med.

2008; 178: 1148-1155.

350. Lorente L, Martin MM, Labarta L, et al. Matrix metalloproteinase-9, -10, and tissue

inhibitor of matrix metalloproteinases-1 blood levels as biomarkers of severity and

mortality in sepsis. Crit Care. 2009; 13: R158.

351. Nakamura T, Ebihara I, Shimada N, et al. Modulation of plasma metalloproteinase-9

concentrations and peripheral blood monocyte mRNA levels in patients with septic shock:

effect of fiber-immobilized polymyxin B treatment. Am J Med Sci. 1998; 316: 355-360.

352. Page AV, Kotb M, McGeer A, et al. Systemic dysregulation of angiopoietin-1/2 in

streptococcal toxic shock syndrome. Clin Infect Dis. 2011; 52: e157-e161.

353. Davies HD, McGeer A, Schwartz B, et al. Invasive Group A Streptococcal infections in

Ontario, Canada. New Engl J Med. 1996; 335: 547-554.

354. Bryant AE, Hayes-Schroer SM, Stevens DL. M type 1 and 3 group A streptococci

stimulate tissue factor-mediated procoagulant activity in human monocytes and endothelial

cells. Infect Immun. 2003; 71: 1903-1910.

355. Kim I, Kim JH, Ryu YS, et al. Tumor necrosis factor-alpha upregulates angiopoietin-2 in

human umbilical vein endothelial cells. Biochem Biophys Res Commun. 2000; 269: 361-

365.

356. van Meurs M, Kümpers P, Ligtenberg JJM, et al. Bench-to-bedside review: angiopoietin

signalling in critical illness - a future target? Crit Care. 2009; 13: 207.

357. Ong T, McClintock DE, Kallet RH, et al. Ratio of angiopoietin-2 to angiopoietin-1 as a

predictor of mortality in acute lung injury patients. Crit Care Med. 2010; 38: 1845-1851.

137

358. Morigi M, Galbusera M, Binda E, et al. Verotoxin-1-induced up-regulation of adhesive

molecules renders microvascular endothelial cells thrombogenic at high shear stress.

Blood. 2001; 98: 1828-1835.

359. Hadler JL, Clogher P, Hurd S, et al. Ten-year trends and risk factors for non-O157 Shiga

toxin-producing Escherichia coli found through Shiga toxin testing, Connecticut, 2000-

2009. Clin Infect Dis. 2011; 53: 269-276.

360. Sauter KAD, Melton-Celsa AR, Larkin K, et al. Mouse model of hemolytic-uremic

syndrome caused by endotoxin-free Shiga toxin 2 (Stx2) and protection from lethal

outcome by anti-Stx2 antibody. Infect Immun. 2008; 76: 4469-4478.

361. te Loo DM, Bosma N, van Hinsbergh V, et al. Elevated levels of vascular endothelial

growth factor in serum of patients with D+ HUS. Pediatr Nephrol. 2004; 19: 754-760.

362. Matsushita K, Yamakuchi M, Morrell CN, et al. Vascular endothelial growth factor

regulation of Weibel-Palade-body exocytosis. Blood. 2005; 105: 207-214.

363. Decaluwe H, Harrison LM, Mariscalco MM, et al. Procalcitonin in children with

Escherichia coli O157:H7 associated hemolytic uremic syndrome. Pediatr Res. 2006; 59:

579-583.

364. Proulx F, Turgeon JP, Litalien C, et al. Inflammatory mediators in Escherichia coli

O157:H7 hemorrhagic colitis and hemolytic-uremic syndrome. Pediatr Infect Dis J. 1998;

17: 899-904.

365. Ake JA, Jelacic S, Ciol MA, et al. Relative nephroprotection during Escherichia coli

O157:H7 Infections: Association with intravenous volume expansion. Pediatrics. 2005;

115: e673-e680.

366. Becker JU, Theodosis C, Jacob ST, et al. Surviving sepsis in low-income and middle-

income countries: new directions for care and research. Lancet Infect Dis. 2009; 9: 577-

582.

138

367. Jacob ST, Moore CC, Banura P, et al. Severe sepsis in two Ugandan hospitals: A

prospective observational study of management and outcomes in a predominantly HIV-1

infected population. PLoS ONE. 2009; 4: e7782.

368. Borgia JA, Basu S, Faber LP, et al. Establishment of a multi-analyte serum biomarker

panel to identify lymph node metastases in non-small cell lung cancer. J Thorac Oncol.

2009; 4: 338-347.

369. Glaser R, Peacock WF, Wu AH, et al. Placental growth factor and B-type natriuretic

peptide as independent predictors of risk from a multibiomarker panel in suspected acute

coronary syndrome (Acute Risk and Related Outcomes Assessed With Cardiac Biomarkers

[ARROW]) study. Am J Cardiol. 2011; 107: 821-826.

370. Ruberg SJ, Chen L, Wang Y. The mean does not mean as much anymore: finding sub-

groups for tailored therapeutics. Clin Trials. 2010; 7: 574-583.

371. Arts RJ, Joosten LA, Dinarello CA, et al. TREM-1 interaction with the LPS/TLR4 receptor

complex. Eur Cytokine Netw. 2011; 22: 11-14.

372. Baelani I, Jochberger S, Laimer T, et al. Availability of critical care resources to treat

patients with severe sepsis or septic shock in Africa: a self-reported, continent-wide survey

of anaesthesia providers. Crit Care. 2011; 15: R10.

373. Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Campaign: International

guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med. 2008;

36: 296-327.

374. Peeling RW, Mabey D. Point-of-care tests for diagnosing infections in the developing

world. Clin Microbiol Infect. 2010; 16: 1062-1069.

375. Chin CD, Laksanasopin T, Cheung YK, et al. Microfluidics-based diagnostics of infectious

diseases in the developing world. Nat Med. 2011; 17: 1015-1020.

376. Bonneh-Barkay D, Bissel SJ, Wang G, et al. YKL-40, a marker of Simian

Immunodeficiency Virus encephalitis, modulates the biological activity of basic fibroblast

growth factor. Am J Pathol. 2008; 173: 130-143.

139

377. Pungpapong S, Nunes DP, Krishna M, et al. Serum fibrosis markers can predict rapid

fibrosis progression after liver transplantation for hepatitis C. Liver Transpl. 2008; 14:

1294-1302.

378. Dixon AE, Mandac JB, Madtes DK, et al. Chemokine expression in Th1 cell-induced lung

injury: prominence of IFN-γ-inducible chemokines. Am J Physiol Lung Cell Mol Physiol.

2000; 279: L592-L599.

379. Bai X, Wilson SE, Chmura K, et al. Morphometric analysis of Th1 and Th2 cytokine

expression in human pulmonary tuberculosis. Tuberculosis (Edinb). 2004; 84: 375-385.

380. Khan IA, MacLean JA, Lee FS, et al. IP-10 is critical for effector T cell trafficking and

host survival in Toxoplasma gondii infection. Immunity. 2000; 12: 483-494.

381. Lee N, Wong CK, Chan PKS, et al. Cytokine response patterns in severe pandemic 2009

H1N1 and seasonal influenza among hospitalized adults. PLoS ONE. 2011; 6: e26050.

382. Rowther FB, Rodrigues CS, Deshmukh MS, et al. Prospective comparison of eubacterial

PCR and measurement of procalcitonin levels with blood culture for diagnosing septicemia

in Intensive Care Unit patients. J Clin Micro. 2009; 47: 2964-2969.

383. Huang J, Motto DG, Bundle DR, Sadler JE. Shiga toxin B subunits induce vWF secretion

by human endothelial cells and thrombotic microangiopathy in ADAMTS13-deficient

mice. Blood. 2010; 116: 3653-3659.

384. Rhee MS, Akanmori BD, Waterfall M, Riley EM. Changes in cytokine production associated

with acquired immunity to Plasmodium falciparum malaria. Clin Exp Immunol. 2001; 126:

503-510.

385. Sinha S, Qidwai T, Kanchan K, et al. Distinct cytokine profiles define clinical immune

response to falciparum malaria in regions of high or low disease transmission. Eur

Cytokine Netw. 2010; 21: 232-240.

140

386. Haissman JM, Vestergaard LS, Sembuche S, et al. Plasma cytokine levels in Tanzanian HIV-

1-infected adults and the effect of antiretroviral treatment. J Acquir Immune Defic Syndr.

2009; 52: 493-497.

387. Tomimoto H, Yano S, Muguruma H, et al. Levels of soluble vascular endothelial growth

factor receptor 1 are elevated in the exudative pleural effusions. J Med Invest. 2007; 54: 146-

153.

388. Aly S, Laskay T, Mages J, et al. Interferon-gamma-dependent mechanisms of mycobacteria-

induced pulmonary immunopathology: the role of angiostasis and CXCR3-targeted

chemokines for granuloma necrosis. J Pathol. 2007; 212: 295-305.

389. Morrow DA, Braunwald E. Future of biomarkers in acute coronary syndromes: moving

toward a multimarker strategy. Circulation. 2003; 108: 250-252.

390. Birkhahn RH, Haines E, Wen W, et al. Estimating the clinical impact of bringing a

multimarker cardiac panel to the bedside in the ED. Am J Emerg Med. 2011; 29: 304-308.

391. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe

sepsis and septic shock. N Engl J Med. 2001; 345: 1368-1377.

392. Nguyen HB, Van Ginkel C, Batech M, et al. Comparison of Predisposition,

Insult/Infection, Response, and Organ dysfunction, Acute Physiology And Chronic Health

Evaluation II, and Mortality in Emergency Department Sepsis in patients meeting criteria

for early goal-directed therapy and the severe sepsis resuscitation bundle. J Crit Care.

2011; Oct 25. [Epub ahead of print].

393. Levy MM, Dellinger RP, Townsend SR, et al. The Surviving Sepsis Campaign: results of an

international guideline-based performance improvement program targeting severe sepsis.

Intensive Care Med. 2010; 36: 222-231.

394. Ioannidis JPA, Panagiotou OA. Comparison of effect sizes associated with biomarkers

reported in highly cited individual articles and in subsequent meta-analyses. JAMA. 2011;

305: 2200-2210.

141

395. Vasan RS. Biomarkers of cardiovascular disease: Molecular basis and practical

considerations. Circulation. 2006; 113: 2335-2362.

396. Karmpaliotis D, Kosmidou I, Ingenito EP, et al. Angiogenic growth factors in the

pathophysiology of a murine model of acute lung injury. Am J Physiol Lung Cell Mol

Physiol. 2002; 283: L585-L595.

397. Huang YQ, Sauthoff H, Herscovici P, et al. Angiopoietin-1 increases survival and reduces

the development of lung edema induced by endotoxin administration in a murine model of

acute lung injury. Crit Care Med. 2008; 36: 262-267.

398. Mei SH, McCarter SD, Deng Y, et al. Prevention of LPS-induced acute lung injury in mice

by mesenchymal stem cells overexpressing angiopoietin 1. PLoS Med. 2007; 4: e269.

399. Meyer NJ, Li M, Feng R, et al. ANGPT2 genetic variant is associated with trauma-

associated acute lung injury and altered plasma angiopoietin-2 isoform ratio. Am J Respir

Crit Care Med. 2011; 183: 1344-1353.

400. Su L, Zhai R, Sheu CC, et al. Genetic variants in the angiopoietin-2 gene are associated

with increased risk of ARDS. Intensive Care Med. 2009; 35: 1024-1030.

401. Fremont RD, Koyama T, Calfee CS, et al. Acute lung injury in patients with traumatic

injuries: utility of a panel of biomarkers for diagnosis and pathogenesis. J Trauma. 2010;

68: 1121-1127.

402. Gallagher DC, Parikh SM, Balonov K, et al. Circulating angiopoietin 2 correlates with

mortality in a surgical population with acute lung injury/adult respiratory distress

syndrome. Shock. 2008; 29: 656-661.

403. Muellenbach RM, Kredel M, Said HM, et al. High-frequency oscillatory ventilation

reduces lung inflammation: a large-animal 24-h model of respiratory distress. Intensive

Care Med. 2007; 33: 1423-1433.

404. Yoder BA, Siler-Khodr T, Winter VT, Coalson JJ. High-frequency oscillatory ventilation.

Effects on lung function, mechanics, and airway cytokines in the immature baboon model

for neonatal chronic lung disease. Am J Respir Crit Care Med. 2000; 162: 1867-1876.

142

405. Vento G, Matassa PG, Ameglio F, et al. HFOV in premature neonates: effects on

pulmonary mechanics and epithelial lining fluid cytokines. A randomized controlled trial.

Intensive Care Med. 2005; 31: 463-470.

406. Masood A, Yi M, Lau M, et al. Therapeutic effects of hypercapnia on chronic lung injury

and vascular remodelling in neonatal rats. Am J Physiol Lung Cell Mol Physiol. 2009;

297: L920-L930.

407. Toriani FJ, Komarow L, Parker RA, et al. Endothelial function in human

immunodeficiency virus-infected antiretroviral-naïve subjects before and after starting

potent antiretroviral therapy. J Am Coll Cardiol. 2008; 52: 569-576.

408. Francisci D, Giannini S, Baldelli F, et al. HIV type 1 infection, and not short-term

HAART, induces endothelial dysfunction. AIDS. 2009; 23: 589-596.

409. Baker J, Ayenew W, Quick H, et al. High-density lipoprotein particles and markers of

inflammation and thrombotic activity in patients with untreated HIV infection. J Infect Dis.

2010; 201: 285-292.

410. Rodger AJ, Fox Z, Lundgren JD, et al. Activation and coagulation biomarkers are

independent predictors of the development of opportunistic disease in patients with HIV

infection. J Infect Dis. 2009; 200: 973-983.

411. Ross AC, Rizk N, O‟Riordan MA, et al. Relationship between inflammatory markers,

endothelial activation markers, and carotid intima-media thickness in HIV-infected patients

receiving antiretroviral therapy. Clin Infect Dis. 2009; 49: 1119-1127.

412. Worm SW, Sabin C, Weber R, et al. Risk of myocardial infarction in patients with HIV

infection exposed to specific individual antiretroviral drugs from the 3 major drug classes:

the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) Study. J Infect Dis.

2010; 201: 318-330.

413. Jong E, Meijers JC, van Gorp EC, et al. Markers of inflammation and coagulation indicate

a prothrombotic state in HIV-infected patients with long-term use of antiretroviral therapy

with or without abacavir. AIDS Res Ther. 2010; 7: 9.

143

414. De Pablo C, Orden S, Apostolova N, et al. Abacavir and didanosine induce the interaction

between human leukocytes and endothelial cells through Mac-1 upregulation. AIDS. 2010;

24: 1259-1266.

415. Wang X, Chai H, Lin PH, et al. Roles and mechanisms of human immunodeficiency virus

protease inhibitor ritonavir and other anti-human immunodeficiency virus drugs in

endothelial dysfunction of porcine pulmonary arteries and human pulmonary artery

endothelial cells. Am J Pathol. 2009; 174: 771-781.

416. Drager LF, Polotsky VY, Lorenzi-Filho G. Obstructive sleep apnea: An emerging risk

factor for atherosclerosis. Chest. 2011; 140: 534-542.

417. Young T, Palta M, Dempsey J, et al. The occurrence of sleep-disordered breathing among

middle-aged adults. N Engl J Med. 1993; 328: 1230-1235.

418. Kato M, Roberts-Thomson P, Phillips BG, et al. Impairment of endothelium-dependent

vasodilation of resistance vessels in patients with obstructive sleep apnea. Circulation.

2000; 102: 2607-2610.

419. Patt BT, Jarjoura D, Haddad DN, et al. Endothelial dysfunction in the microcirculation of

patients with obstructive sleep apnea. Am J Crit Care Med. 2010; 182: 1540-1545.

420. Sert Kuniyoshi FH, Singh P, Gami AS, et al. Patients with obstructive sleep apnea exhibit

impaired endothelial function after myocardial infarction. Chest. 2011; 140: 62-67.

421. Dyugovskaya L, Lavie P, Lavie L. Increased adhesion molecules expression and

production of reactive oxygen species in leukocytes of sleep apnea patients. Am J Respir

Crit Care Med. 2002; 165: 934-939.

422. El Solh AA, Akinnusi ME, Berim IG, et al. Hemostatic implications of endothelial cell

apoptosis in obstructive sleep apnea. Sleep Breath. 2008; 12: 331-337.

423. Drager LF, Li J, Shin MK, et al. Intermittent hypoxia inhibits clearance of triglyceride-rich

lipoproteins and inactivates adipose lipoprotein lipase in a mouse model of sleep apnoea.

Eur Heart J. 2011; Apr 9. [Epub ahead of print].

144

424. Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of

antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious

Diseases Society of America. Clin Infect Dis. 2011; 52: e56-e93.

425. Schuetz P, Chiappa V, Briel M, Greenwald JL. Procalcitonin algorithms for antibiotic

therapy decisions. A systematic review of randomized controlled trials and

recommendations for clinical algorithms. Arch Intern Med. 2011; 171: 1322-1331.

426. Agarwal R, Schwartz DN. Procalcitonin to guide duration of antimicrobial therapy in

intensive care units: A systematic review. Clin Infect Dis. 2011; 53: 379-387.

427. Li H, Luo YF, Blackwell TS, Xie CM. Procalcitonin-guided therapy in respiratory tract

infections: a meta-analysis and systematic review. Antimicrob Agents Chemother. 2011;

Sep 26. [Epub ahead of print].

428. Jensen JU, Hein L, Lundgren B, et al. Procalcitonin-guided interventions against infections

to increase early appropriate antibiotics and improve survival in the intensive care unit: a

randomized trial. Crit Care Med. 2011; 39: 2048-2058.

429. Bele N, Darmon M, Coquet I, et al. Diagnostic accuracy of procalcitonin in critically ill

immunocompromised patients. BMC Infect Dis. 2011; 11: 224.

430. Gac AC, Parienti JJ, Chantepie S, et al. Dynamics of procalcitonin and bacteremia in

neutropenic adults with acute myeloid leukemia. Leuk Res. 2011; 35: 1294-1296.

431. Juutilainen A, Hamalainen S, Pulkki K, et al. Biomarkers for bacteremia and severe sepsis

in hematological patients with neutropenic fever: multivariate logistic regression analysis

and factor analysis. Leuk Lymphoma. 2011; Jul 14. [Epub ahead of print].

432. Mori Y, Miyawaki K, Kato K, et al. Diagnostic value of serum procalcitonin and C-

reactive protein for infections after allogeneic hematopoietic stem cell transplantation

versus nontransplant setting. Intern Med. 2011; 50: 2149-2155.

433. Robinson JO, Lamoth F, Bally F, et al. Monitoring procalcitonin in febrile neutropenia:

what is its utility for initial diagnosis of infection and reassessment in persistent fever?

PLoS ONE. 2011; 6: e18886.