Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic determinants of the host response to infection in critically ill adults with systemic inflammatory… Sutherland, Ainsley M. 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-319321.pdf [ 11.54MB ]
Metadata
JSON: 831-1.0100762.json
JSON-LD: 831-1.0100762-ld.json
RDF/XML (Pretty): 831-1.0100762-rdf.xml
RDF/JSON: 831-1.0100762-rdf.json
Turtle: 831-1.0100762-turtle.txt
N-Triples: 831-1.0100762-rdf-ntriples.txt
Original Record: 831-1.0100762-source.json
Full Text
831-1.0100762-fulltext.txt
Citation
831-1.0100762.ris

Full Text

GENETIC DETERMINANTS OF THE HOST RESPONSE TO INFECTION IN CRITICALLY ILL ADULTS WITH SYSTEMIC INFLAMMATORY RESPONSE SYNDROME by A1NSLEY M. SUTHERLAND B.Sc.H., Queen's University at Kingston, 2001 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Experimental Medicine) THE UNIVERSITY OF BRITISH COLUMBIA August 2007 © Ainsley M . Sutherland, 2007 ABSTRACT Background: Activation of a systemic inflammatory response to infection varies significantly between individuals with important clinical implications. Genotype contributes substantially to outcome of infectious disease. Hypothesis: Allelic variants of inflammatory and innate immunity genes affect protein levels and function and are predictive of outcome in critically ill adults who have systemic inflammatory response syndrome (SIRS). Methods: Clinical data for derivation and validation cohorts of critically ill Caucasian patients with SIRS were collected for 28 days after admission to ICU and also for cardiac surgery patients. A subgroup of cardiac surgery patients had blood drawn post-operatively for cytokine measurements. Haplotypes of candidate genes were inferred from publicly available data using PHASE and cladistic structure determined using MEGA2. A set of "haplotype tag" single nucleotide polymorphisms (htSNPs) that defined major haplotype clades of the candidate genes and previously examined SNPs were chosen for genotyping in the three cohorts. Main Results: CD14 -159TT was associated with Gram-negative cultures in the derivation cohort and with mortality in the validation cohort. Mannose-binding lectin (MBL) haplotype pairs XO/O and O/O were associated with positive bacterial cultures and TLR2 -16933AA was associated with sepsis and with Gram-positive cultures in the derivation cohort. The C/T/A clade of IRAK4 was associated with Gram-positive cultures in a large derivation cohort and with decreased B-lymphocyte response to CpG and a trend to decreased fibroblast response to LPS. Haplotype clades of IL-6 were associated with 28-day mortality and organ dysfunction in the derivation cohort, but not in the validation cohort. Haplotype clades of IL-6 were associated with post-surgical vasodilation in cardiac surgery patients, but not with altered serum levels of cytokines after cardiac surgery. Conclusions: Variation in the key inflammatory and innate immunity genes IL-6, CD14, M B L and TLR2 contribute to variation in individuals' responses to inflammatory stimuli. ii TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES vii LIST OF FIGURES x LIST OF SYMBOLS, N O M E N C L A T U R E AND ABBREVIATIONS xiii GLOSSARY rxv ACKNOWLEDGEMENTS xvii CO-AUTHORSHIP STATEMENT xviii CHAPTER 1: INTRODUCTION 1.1 The Systemic Inflammatory Response Syndrome and Sepsis 1 1.2 Genetic Variability of the Host Immune and Inflammatory Responses 2 1.3 Measures of Genetic Variation 5 1.4 Hypothesis 14 1.5 Objectives and Experimental Approach 15 1.6 Tables and Figures 24 1.7 References 30 iii CHAPTER 2: POLYMORPHISMS IN CD14, MANNOSE-BINDING LECTIN, A N D TOLL-LIKE RECEPTOR-2 A R E ASSOCIATED WITH INCREASED P R E V A L E N C E OF INFECTION IN CRITICALLY ILL ADULTS 2.1 Introduction 35 2.2 Methods 37 2.3 Results .' 43 2.4 Discussion 47 2.5 Tables and Figures 53 2.6 References 72 CHAPTER 3: A HAPLOTYPE C L A D E OF INTERLEUKIN-1 RECEPTOR ASSOCIATED KINASE 4 (IRAK4) CONTAINING A NON-SYNONYMOUS POLYMORPHISM (ALA428THR) IS ASSOCIATED WITH A L T E R E D INNATE IMMUNE INFLAMMATORY RESPONSES EX VIVO AND BACTERIAL INFECTION IN CRITICALLY ILL PATIENTS 3.1 Introduction : 76 3.2 Methods 78 3.3 Results 87 3.4 Discussion 90 3.5 Tables and Figures 95 3.6 References 109 CHAPTER 4: THE ASSOCIATION OF INTERLEUKIN 6 HAPLOTYPE CLADES WITH O U T C O M E FROM THE SYSTEMIC INFLAMMATORY RESPONSE 4.1 Introduction 113 iv 4.2 Methods 117 4.3 Results :. 125 4.4 Discussion 130 4.5 Tables and Figures 139 4.6 References 161 CHAPTER 5: CONCLUSIONS 5.1 Summary of Main Findings 165 5.2 Strengths and Weaknesses 168 5.3 Significance and Relevance to Current Knowledge 174 5.4 Future research 176 5.5 References 183 APPENDIX 1: RECRUITMENT AND PHENOTYPING OF PATIENT COHORTS FOR GENETIC ASSOCIATION STUDIES A l . 1 Recruitment of Critically 111 Patients for Genetic Association Studies 187 A1.2 Collection of Clinical Data for the Cohort of Critically 111 Patients 188 A l .3 Clinical Outcomes in the Cohort of Critically 111 Patients 190 A l .4 Collection and Storage of Critically 111 Patients' DNA 191 A1.5 Recruitment of Patients Undergoing Coronary Artery Bypass Graft with Cardiopulmonary Bypass for Genetic Association Studies 191 A1.6 Collection of Clinical Data for the Cohort of Cardiopulmonary Bypass Patients 193 v A1.7 Clinical Outcomes in the Cohort of C A B G with Cardiopulmonary Bypass Patients 193 A1.8 Intermediate Phenotypes in the Cohort of C A B G with Cardiopulmonary Bypass Patients 194 A 1.9 Collection and Storage of Cardiopulmonary Bypass Patients' DNA and serum 195 ALIO Tables 197 A l . l l References 203 APPENDIX 2: ETHICS CERTIFICATES OF APPROVAL 204 vi LIST OF TABLES Table 1-1: Comparison of SNP-based and haplotype-based genetic association studies 23 Table 1-2. Clinical phenotypes used as primary and secondary outcome variables in the derivation and validation cohorts of critically ill Caucasians 24 Table 1-3. Human Toll-like receptors (TLRs) and their ligands 25 Table 2-1: M B L haplotypes associated with high, moderate, or low levels of serum M B L 53 Table 2-2: Primer and probe sequences used for genotyping CD14 C-159T in the validation cohort using the ABI Prism 7900HT Sequence Detection System 54 Table 2-3: M B L Haplotype Frequencies in the Derivation Cohort of Critically 111 Adults with SIRS 55 Table 2-3: CD14, MBL, and TLR2 genotype/hap lo type frequencies and derivation cohort patient baseline characteristics by genotype/haplotype 56 Table 2-4: CD14 genotype frequencies and validation cohort patient baseline characteristics by genotype 57 Table 2-5: Frequency of types and sources of positive microbiological cultures by CD 14 C-159T genotype in the derivation and validation cohorts of critically ill patients 58 Table 2-6: Comparison of derivation and validation cohort baseline characteristics by CD14 C-159T genotype 59 Table 2-7: Frequency of types and sources of positive microbiological cultures by M B L Haplotype pair in the derivation cohort of critically ill patients 60 Table 2-8: Frequency of types and sources of positive microbiological cultures by TLR2 T-16933 A genotype in the derivation cohort of critically ill patients 61 Table 3-1: Primer and probe sequences used for genotyping IRAK4 htSNPs in the ABI Prism 7900HT Sequence Detection System 95 vii Table 3-2: Synthetic bacterial DNA sequences (CpG) used for stimulation of B-lymphocytes 96 Table 3-3: Primers used for site-directed mutagenesis (SDM) and sequencing of the IRAK4 expression plasmid 97 Table 3-4: IRAK4 haplotype frequencies and patient baseline characteristics by haplotype 98 Table 3-5: Prevalence of sepsis and septic shock at admission to ICU, and 28-day survival by IRAK4 haplotype clade 99 Appendix 3-1: Coriell Cell Repository IDs, SeattleSNP IDs and IRAK4 G29429A genotypes of B-lymphocyte cell lines used in mechanistic studies 112 Table 4-1: Brussels organ dysfunction scoring criteria 139 Table 4-2: Baseline characteristics of the derivation cohort of 228 critically ill adults with SIRS 140 Table 4-3: Cox Proportional Hazard Analysis - Hazard ratios for mortality by IL-6 haplotype clade in derivation cohort of critically ill patients 141 Table 4-4: Genotype frequencies and allele frequencies for IL-6 htSNPs -174G/C, 1753C/G, and 2954G/C in a derivation cohort of 228 critically ill adults with SIRS 142 Table 4-5: Haplotype clade frequencies in the derivation cohort versus the validation cohort of critically ill patients 143 Table 4-6: Genotype frequencies and allele frequencies for IL-6 htSNPs 614G/A, 1753C/G, and 2954G/C in validation cohort of 441 critically ill adults with SIRS 144 Table 4-7: Baseline characteristics of validation cohort of 441 critically ill adults with SIRS 145 Table 4-8: Cox Proportional Hazard Analysis - Hazard ratios for mortality by IL-6 haplotype clade in validation cohort of critically ill patients 146 viii Table 4-9: Frequency of IL-6 haplotype clades in a cohort of 603 cardiopulmonary bypass patients 147 Table 4-10: Baseline demographics of cardiopulmonary bypass patients by IL-6 haplotype clade 148 Table 4-11: Surgical details of cardiopulmonary bypass patients by IL-6 haplotype clade 149 Table 4-12: Post-operative serum concentrations of IL-6, MCP-1, G-CSF, IL-8 and IL-lra in cardiac surgery patients 150 Table A l . l : Chart Review Form for collection of critically ill patients' data 197 Table A1.2: Brussels Organ Dysfunction Scoring System used for assessing organ dysfunction in critically ill patients 200 Table A1.3: Chart Review Form for cohort of C A B G with cardiopulmonary bypass patients 201 ix LIST OF FIGURES Figure 1-1: Innate immunity, inflammation and coagulation in systemic inflammatory response syndrome 26 Figure 1-2: Activation of Innate Immunity Receptors by Pathogens 27 Figure 1 -3: Activation of Innate Immunity Signalling Pathways through Toll-like Receptors 2 and 4 (TLR2 and TLR4) 28 Figure 2-1: Haplotype structure of the Toll-like receptor 2 (TLR2) gene 62 Figure 2-2: Prevalence of positive bacterial cultures at ICU admission by Cluster of differentiation 14 (CD14) C-159T genotype 63 Figure 2-3: Prevalence of Gram-negative bacterial cultures at ICU admission by CD14 C-159T genotype 64 Figure 2-4: Kaplan-Meier survival analysis by CD14 genotype in the derivation cohort 65 Figure 2-5: Kaplan-Meier survival analysis by CD14 genotype in the validation cohort 66 Figure 2-6: Prevalence of positive bacterial cultures at ICU admission by Mannose-binding lectin (MBL) haplotype pair 67 Figure 2-7: Kaplan-Meier survival analysis by M B L haplotype pair 68 Figure 2-8: Prevalence of sepsis at ICU admission by TLR2 T-16933A genotype 69 Figure 2-9: Prevalence of Gram-positive bacterial cultures at ICU admission by TLR2 T-1693 3A genotype 70 Figure 2-10: Kaplan-Meier survival analysis by TLR2 T-16933A genotype 71 Figure 3-1: Haplotype structure of the Interleukine-1 receptor associated kinase 4 (IRAK4) gene .' 100 Figure 3-2: Prevalence of positive bacterial cultures at ICU admission by IRAK4 haplotype 101 x Figure 3-3: Prevalence of Gram-positive bacterial cultures at ICU admission by IRAK4 haplotype clade 102 Figure 3-4: Prevalence of positive endovascular bacterial cultures (ie positive blood cultures) at ICU admission by IRAK4 haplotype clade 103 Figure 3-5: Prevalence of Gram-positive endovascular bacterial cultures (ie positive blood cultures) at ICU admission by IRAK4 haplotype clade 104 Figure 3-6: B-lymphocyte immune response (as measured by IL-6 supernatant concentration) by IRAK4 haplotype clade 105 Figure 3-7: TLR-ligand stimulation of normal adult fibroblasts (NAFs) and IRAK4-deficient fibroblasts (IDFs) 106 Figure 3-8: BLAST result comparing sequences of wild-type IRAK4 expression plasmid containing the 29429G allele (query) with IRAK4 expression plasmid mutated to carry the 29429A allele (subject) 107 Figure 3-9: TLR-ligand stimulation of lRAK4-deficient fibroblasts (IDFs) transfected with IRAK4 expression plasmids containing either the 29429G (IDF-G) or A (IDF-A) alleles or with the empty vector (IDF-null) 108 Figure 4-1: Haplotype structure of the IL-6 gene 151 Figure 4-2: Phylogenetic relationship of IL-6 haplotypes 152 Figure 4-3: 28-day mortality rates by IL-6 clade 153 Figure 4-4: Kaplan-Meier mortality analysis by IL-6 clade 154 Figure 4-5: Multiple-system organ dysfunction in critically ill patients 155-156 Figure 4-6: Kaplan-Meier mortality analysis by IL-6 clade in the validation cohort of critically ill patients 157 Figure 4-7: A/C/G, G/G/G and G/C/C haplotype clades are associated with increased occurrence of a vasodilatory syndrome following cardiac surgery 158 xi Figure 4-8: A/C/G, G/G/G and G/C/C haplotype clades are associated with increased occurrence of a vasodilatory syndrome following cardiac surgery in a subgroup of patients on vasopressors 159 Figure 4-9: IL-6 haplotype clades are not associated with hours spent in the CSICU after cardiac surgery 160 xii LIST OF SYMBOLS, NOMENCLATURE AND ABBREVIATIONS ALI - Acute lung injury ANOVA - Analysis of Variance APACHE II - Acute Physiology and Chronic Health Evaluation score II APACHE III - Acute Physiology and Chronic Health Evaluation score III ARDS - Acute respiratory distress syndrome CABG - Coronary artery bypass grafting C D 1 4 - Cluster of differentiation 14 cDNA - Coding DNA CPB - Cardiopulmonary bypass DNA - Deoxyribonucleic acid ELISA - Enzyme-linked immunosorbant assay GCSF - Granulocyte colony stimulating factor htSNP - Haplotype tag single nucleotide polymorphism ICU - Intensive care unit IDF - IRAK4-deficient fibroblast IL-lra - Interleukin-1 receptor antagonist IL-6 - Interleukin-6 IL-8 - Interleukin-8 IRAKI - Interleukin-1 receptor associated kinase I R A K 4 - Interleukin-1 receptor associated kinase LPS - Lipopolysaccharide MALDItof - Matrix Assisted Laser Desorption /Ionization- Time Of Flight MBL - Mannose-binding lectin MCP-1 - Monocyte chemotactic protein-1 xiii MyD88 - Myeloid differentiation factor 88 NAF - Normal adult fibroblast N F K B - Nuclear factor kappa B PAMP - Pathogen-associated molecular pattern PCR-RFLP - Polymerase chain reaction - restriction fragment length polymorphism PGN - Peptidoglycan RNA - Ribonucleic acid SDM-PCR - Site-directed mutagenesis polymerase chain reaction SD - Standard deviation SIRS - Systemic inflammatory response syndrome SNP - Single nucleotide polymorphism SSP-PCR - Sequence-specific primers polymerase chain reaction TDT - Transmission-disequilibrium test TIR - Toll-like/Interleukin-1 receptor TLR2 - Toll-like receptor 2 TLR3 - Toll-like receptor 3 GLOSSARY Acute lung injury (ALI): A respiratory syndrome that includes need for mechanical ventilation, diffuse bilateral infiltrates on chest x-ray, a Pa02/Fi02 <300 (a measure of oxygenation), no evidence of volume overload or congestive heart failure, and occurring in the presence of an underlying cause. Acute respiratory distress syndrome (ARDS): More severe ALI with the Pa02/Fj02 <200. Allele: One of the different forms of a gene that can appear at a single locus. Codon: A section of DNA (three nucleotide pairs in length) that codes for a single amino acid. Exon: Any non-intron section of the coding sequence of a gene. Together the exons constitute the RNA and are translated into protein. Haplotype: Patterns of several SNPs that are in linkage disequilibrium with one another within a gene or segment of DNA, and are thus inherited as a unit. Haplotype clade: Evolutionarily related groups of haplotypes. Hardy-Weinberg equilibrium: The stable frequency distribution of genotypes A A , Aa, and aa, in the proportions p , 2pq, and q , respectively (where p and q are the frequencies of the alleles A and a), that is a consequence of random mating in the absence of mutation, migration, natural selection, or random drift. Intron: A segment of largely unknown function within a gene. This segment is initially transcribed, but the transcript is not found in the functional mRNA. Linkage disequilibrium: Physical coupling of alleles on the same chromosome that limits recombination and prevents their independent assortment when transmitted to the next generation. Mutation: A variation in the DNA sequence in a gene that occurs rarely (<1%). xv Population stratification: A confounding factor in genetic association studies. The confounding occurs when individuals are selected from two genetically different populations in different proportions in cases and controls, or exposed and unexposed individuals. Thus, the individuals are not matched for their genetic background. This may cause spurious associations, or it may mask true associations like any other unknown confounder. Sepsis: SIRS as the result of a known or suspected infection. Septic shock: Sepsis plus hypotension leading to decreased tissue perfusion and tissue hypoxia. Single nucleotide polymorphism: A variation in the DNA sequence in a gene that occurs relatively commonly (>1%). Systemic inflammatory response syndrome (SIRS): The systemic response to infection or trauma manifested by 2 or more of the following: (1) temperature less than 36°C, or greater than 38°C, (2) tachycardia (heart rate >90 beats/min), (3) tachypnea (respiratory rate >20 breaths/min), and (4) perturbations in white blood cell count. Transition: A type of nucleotide-pair substitution involving the replacement of a purine with a purine, or a pyrimidine with a pyrimidine - for example, GC to AT. Transversion: A type of nucleotide-pair substitution involving the replacement of a purine with a pyrimidine, or vice versa - for example, GC to TA. xvi ACKNOWLEDGEMENTS This work is supported by the Canadian Institutes of Health Research and the Heart and Stroke Foundation of British Columbia and Yukon. I received salary support from a Michael Smith Foundation for Health Research and the Canadian Institutes of Health Research. I would like to thank Dr. Jim Russell and Dr. Keith Walley for their support and mentorship, both in the lab and clinically. I would like to thank all the members of the Walley/Russell/Dorscheid lab for all their help and suggestions and for providing an exciting and stimulating work environment. I would like to thank Dr. Mark Wilkinson, Dr. Anthony Chow and Dr. Keith Walley for guiding me as members of my thesis supervisory committee. Most importantly, I would like to thank my family for all their love and support. xvii CO-AUTHORSHIP STATEMENT Chapter 1: Dr. Russell and I co-wrote section 1.3 - Measures of Genetic Variation - as part of a review entitled "Issues with Polymorphism Analysis in Sepsis" that has been accepted for publication in Clinical Infectious Diseases. Dr. Russell outlined the original review in an abstract for the Sepsis Symposium in Cambridge, England in 2004. I performed a literature review and wrote the original review. Dr. Russell edited and added to the review. Chapter 2: Chapter 2 is a copy of the published manuscript: Polymorphisms in CD14, mannose-binding lectin, and Toll-like receptor-2 are associated with increased prevalence of infection in critically ill adults; Crit Care Med. 2005 Mar;33(3):638-44, with additional data concerning the association of CD14 with mortality in a validation cohort. I selected CD14, M B L and TLR2 as candidate genes for an association study based on previous literature and the focus on innate immunity of Dr. Russell's grants. I performed all the genetic analysis and selected haplotype tag SNPs and SNPs previously described in the literature to be genotyped. In the derivation cohort, I genotyped polymorphisms in CD14, MBL, and TLR2 by PCR-RFLP and PCR-SSP. I did the statistical analysis to test for associations between SNPs and haplotypes of CD14, M B L and TLR2 and clinical outcome in the derivation and validation cohorts of critically ill patients. I wrote the original manuscript and made all the figures. Dr. Walley and Dr. Russell assisted with the statistical analysis and edited the manuscript. Chapter 3: Chapter 3 is a manuscript in preparation: A haplotype clade of Interleukin-1 Receptor Associated Kinase 4 (IRAK4) containing a non-synonymous polymorphism (ALA428THR) xviii is associated with altered innate immune inflammatory responses in vitro and bacterial infection in critically ill patients. I selected IRAK4 as a candidate gene for an association study based on previous literature and the focus on innate immunity of Dr. Russell's SCCOR grant. I performed all the genetic analysis and selected haplotype tag SNPs to be genotyped. I did the statistical analysis to test for associations between SNPs and haplotype clades of IRAK4 and clinical outcome in the cohort of critically ill patients. I designed the experiments to test for the association of the C/T/A haplotype clade with in vitro B-lymphocyte and fibroblast response to toll-like receptor ligands. Andy (Ho-Pan) Sham optimized and carried out the stimulation of the B-lymphocyte cell lines with CpG. Andy (Ho-Pan) Sham performed the original transfection of the IRAK4-deficient fibroblasts with the wild-type IRAK4 expression plasmid and subsequent stimulation. I carried out the site-directed mutagenesis experiments to exchange the allele at position 29429 of the IRAK4 gene. I performed the transfection of IRAK4-deficient fibroblasts with both the A and G allele expression plasmids and stimulated the cells with lipopolysaccharide and peptidoglycan. I measured IL-6 in the supernatant of the stimulated cells by ELISA. I did the statistical analysis for all the in vitro experiments. I wrote the original manuscript and made all the figures. Dr. Walley and Dr. Russell assisted with the statistical analysis and edited the manuscript. Chapter 4: Chapter 4 is a copy of the published manuscript: The association of interleukin 6 haplotype clades with mortality in critically ill adults; Arch Intern Med. 2005 Jan 10;165(l):75-82, with additional data concerning the validation cohort and cardiopulmonary bypass patient cohort. I selected Interleukin-6 as a candidate gene for an association study based on previous literature and the focus on inflammation of Dr. Russell and Dr. Walley's grants. I performed xix all the genetic analysis and selected haplotype tag SNPs to be genotyped. I did the statistical analysis to test for associations between haplotypes of IL-6 and clinical outcome in the derivation and validation cohorts of critically ill patients and in the cohort of cardiopulmonary bypass patients. I measured the post-operative serum cytokine levels of the cardiac surgery patients. I wrote the original manuscript and made all the figures. Dr. Walley, Dr. Manocha, and Dr. Russell assisted with the statistical analysis and edited the manuscript. xx C H A P T E R 1: I N T R O D U C T I O N 1.1 The Systemic Inflammatory Response Syndrome and Sepsis The Systemic Inflammatory Response Syndrome (SIRS) is believed to affect one-third of all in-hospital patients, more than 50% of all intensive care patients, and greater than 80% of surgical patients SIRS is defined as the presence of at least 2 of the following 4 criteria: 1) fever (>38°C) or hypothermia (<36 °C), 2) tachycardia (>90 beats/minute), 3) tachypnea (>20 breaths/minute), PaCC>2 < 32 mm Hg, or need for mechanical ventilation, and 4) leukocytosis (total leukocyte count >12,000 mm ) or leukopenia (<4,000 mm" ) 2 . SIRS due to a bacterial, viral, fungal, or protozoal infection is termed sepsis. Sepsis has a mortality rate of about 30%, and when sepsis leads to multiple organ dysfunction and shock (septic shock), mortality increases to 50%-60% 2 . Each year more people die of sepsis than of acute myocardial infarction 3 . The systemic inflammatory response syndrome and sepsis are biologically complex disorders that remain incompletely understood. SIRS can ultimately be thought of as an imbalance in pro-inflammatory and anti-inflammatory processes, and also in pro-coagulant and anti-coagulant processes (Figure 1-1). Because many important pathways of the innate immunity, inflammatory, and coagulation systems are involved and interact, it has been very difficult to develop effective treatments. Treatments targeting any one of these pathways are often double-edge swords. For example, anti-inflammatory steroid treatment often leads to delayed wound healing. To date, there is only one evidence-based treatment available for patients with severe sepsis, activated protein C (APC) 4 . Parts of this chapter have been published. Sutherland A M , Russell JA. Issues with Polymorphism Analysis 1 in Sepsis. Clin Infect Dis. 2005 Nov 15;41 Suppl 7:S396-402. 1.2 Genetic Variability of the Host Immune and Inflammatory Responses Although many advances have been made in the understanding of the pathophysiology of sepsis, only one definitive therapy exists, and is only available for the sickest patients. Many treatments have improved survival in certain patient populations, yet the overall mortality rate for septic patients remains high 3 ' 5 ' 6 . Sepsis is a complex disease that can result from the response to any class of invading microorganism, and there is considerable inter-individual variability in the degree of activation of the innate immunity and inflammatory responses to infection. Numerous studies have shown that outcome from infection is heritable 7"9. For example, the genetic contribution to death from infection exceeded the genetic contribution to cancer risk by 5 fold 1 0 . If patients predisposed to an inappropriate immune response to infection could be prospectively identified (for example, by testing their genotype) then individually tailored therapy could be targeted to patients at higher risk by genotype to decrease risk of death. In addition, detrimental side effects could be avoided by not treating other patients who are not at risk by genotype. For example, anti-tumor necrosis factor-alpha (TNFa) could be more effective in patients who have a genotype/haplotype associated with greater TNFa production. It is highly likely that genetic information will be used by the clinician to define clinical subtypes of disease, and to stratify patients according to their risk of poor outcome. Genotype could also be used to determine the optimal drug and its dose and duration for treatment of each individual, while minimizing adverse effects. This knowledge will be instrumental in both preventive medicine (by predicting prognosis and using preventative strategies) and in treatment decisions. Innate immunity and inflammatory genes are important candidate genes in examining variability of the host response to infection as they are the first line of defense against a pathogen, and are potent initiators of the inflammatory response. Innate immunity receptors 2 are a critical point of direct contact between the host and invading pathogens (Figure 1 -2). Innate immunity receptors are pattern recognition receptors (PRRs); they recognize specific conserved components of invading microorganisms, or pathogen-associated molecular patterns (PAMPs). Important soluble PRRs include mannose-binding lectin, which binds mannose groups on pathogens and activates the alternative or "lectin" complement pathway n ' 1 2 , and lipololysaccharide binding protein (LPB) which recognizes and binds lipopolysaccharide (LPS) from Gram-negative bacteria 1 3 ' 1 4 and presents it to soluble and membrane-bound Cluster of differentiation 14 (CD14) l 3 ' ' 4 . CD14 is found on the surface of moncytes/macrophages and hepatocytes in association with toll-like receptor 4 (TLR4) l 5 ' 1 6 . Membrane-bound CD 14 is cleaved from these cells and shed into the blood stream as soluble CD14 (sCD14) 1 7 where it helps to present LPS to TLR4 on other immune cells 1 5 " 1 8 . CD14 has no intracellular domain and so it must interact with TLR4 and the adaptor protein MD2 to activate intracellular signaling 1 5 1 8 1 9 . CD14 and TLR4/MD2 form the LPS receptor complex 1 5 ~ 1 8 . TLR4 is a member of the toll-like receptor family, one of the most important families of innate immunity receptors. Toll is a transmembrane receptor first identified in Drosophila . Stimulation of Toll in Drosophila leads to the production of the antifungal peptide Drosomycin 2 1 . To date, ten homologous toll-like receptors have been identified in humans (TLR1-TLR10) 2 2 2 7 (Table 1-1). Toll and human TLRs are characterized by a leucine-rich repeat (LRR) in their extracellular domain 2 8 and a toll-like/interleukin-1 receptor (TIR) domain in their intracellular domain 2 9 . This TIR domain is also found in the intracellular domain of the interleukin-1 receptor (IL-lr) and the IL-18 receptor . Activation of TLRs by the binding of their respective ligands to their extracellular domains causes an intracellular 3 signaling cascade which ultimately leads to activation of nuclear factor kappa B ( N F K B ) and other transcription factors and transcription of pro-inflammatory cytokines 3 0 - 3 3 (Figure 1 -3). Specifically, the activation of TLRs by their specific ligands causes intracellular Myeloid differentiation primary-response protein 88 (MyD88) to be recruited by its TIR domain to the TIR domain of the TLR at the cell membrane. MyD88 interacts with TLRs to recruit the serine/threonine IL-lr associated kinases 1 and 4 (IRAKI and IRAK4) to the receptor complex 3 4 ' 3 5 . IRAK4 is essential for MyD88-dependent signaling and is needed for all TLR signalling with the exception of TLR3 3 6 . IRAK4 induces phosphorylation of IRAKI. IRAKI then further auto-phosphorylates and associates with tumour-necrosis-factor receptor-associated factor 6 (TRAF6) 3 4 ' 3 5 ' 3 7 " 3 9 . IRAKI and TRAF6 dissociate from the receptor and interact with transforming-growth factor-P-activated kinase 1 (TAK1), TAK1-binding protein 1 (TAB1) and TAB2, inducing phosphorylation of TAB2 and TAK1 3 4 ' 3 5 ' 3 7 ~ 3 9 . IRAKI is degraded and the TRAF6-TAK1-TAB1-TAB2 complex translocates to the cytosol 3 4' 3 5' 3 7" 3 9. In the cytosol the complex associates with ubiquitin ligases leading to the ubiquitylation of TRAF6 and activation of TAK1 3 4 ' 3 5 ' 3 7 " 3 9 . TAK1 phosphorylates both mitogen-activated protein (MAP) kinases and the inhibitor of nuclear factor-KB (iKB)-kinase (IKK) complex made of IKK-a, IKK-p and IKK-y 3 4 , 3 5 ' 3 7 " 3 9 . I K B is phosphorylated by the IKK complex and is ubiquinated and degraded, allowing N F - K B to translocate to the nucleus and induce expression of key innate immunity and inflammatory genes, for example, Interleukin-6 (IL-6) 3 4 , 3 5 ' 3 7 " 3 9 . Mitogen-activated protein kinase pathways are also activated and induce activation of the stress kinases ERK1/2, JNK, and p38. The activation of these stress kinases leads to the activation of the transcription factors SRE, API and CREB that have minor roles in controlling transcription of various inflammatory molecules 4 0 ~ 4 2 . 4 Polymorphisms within innate immunity and inflammatory genes may lead to an inappropriate inflammatory response to an infection 4 3 - 4 5 . Polymorphisms in the promoter regions of genes may cause altered levels of expression of cytokines, chemokines, coagulant proteins, and other proteins central to the immune response, while polymorphisms within exons may lead to synthesis of dysfunctional proteins. Polymorphisms within the bactericidal/ permeability increasing protein (BPI), lipopolysaccharide binding protein (LBP), CD 14, TLR2, TLR4, and mannose-binding lectin (MBL) have been tested for association with outcomes of sepsis and infection. Many of the findings have not been clear, or have conflicted with previous studies 4 6 . Variability among individuals' genetic makeup plays a large role in inter-individual variability in susceptibility and response to disease. To truly associate genotype with phenotype or clinical outcome, we must now show biologically plausible mechanisms through which genetic variations cause phenotypic variation, and we must validate associations in separate and independent groups of individuals or patients. Understanding the role of genetic variation in disease will be central to medicine in the coming years, but only if we can understand how genetic variation is causing variability of clinical phenotype. 1.3 Measures of Genetic Variation With the completion of the sequencing of the human genome a new project is underway to document the 0.1%, or 3 million bases, of the human genome that are variable between individuals (http://www.ncbi.nlm.nih.gov/projects/SNP/) 4 7 ' 4 8 . The majority of variation in the human genome is in the form of single nucleotide polymorphisms (SNPs), or single base changes in the DNA sequence. Other forms of polymorphism such as insertion and deletion are of functional importance as well. This variable 0.1% of the genome must 5 account for the inter-individual genetic variation we see in quantitative traits and risk of disease 4 9 ' 5 0 . While many genes have been characterized that are responsible for rare, Mendelian diseases, we are still struggling to understand the contribution of genetic variation to complex disease such as sepsis. In addition to the difficulty of assessing the individual contributions of potentially hundreds of genetic variants to a complex disease, we must also take into account the variability of environmental factors. Critically ill patients acquire infections from a number of sources, such as community-acquired pneumonia, line infections or surgical site infections, and these infections may affect different sites of the body (ie. lung, blood, gastrointestinal, urinary tract). Additionally these infections can be caused by any number of different organisms, many of which are resistant to antibiotics, including multidrug-resistant Staphylococcus aureus. A patient's clinical phenotype and outcome may depend heavily on the type and severity of infection they have acquired, how soon the infection is detected and whether it is treated appropriately, whether with antibiotics, medical treatment or surgery. Understanding the contribution of environmental variability to complex disease and how it interacts with genetic variability is essential to understanding of the role of genetic variability in determining outcome from complex diseases such as sepsis. SNP Analysis in Genetic Association Studies of Sepsis A fundamental principle of genetic association studies is the measurement of association of a unit of the human genome with a phenotype. The first generation of gene association studies focused on SNPs as the unit of genetic variation that is measured in each individual and then associated (or not) with a clinical phenotype, such as death. The effort to find genetic variants that are responsible for susceptibility to severe infection and sepsis has 6 been almost entirely gene association studies in moderate size case control or cohort studies. However, association studies based on the functional plausibility of single SNPs may overlook many polymorphisms essential for proper expression of genes for reasons to be developed below. Investigators have evaluated single SNPs usually in regulatory or coding regions that alter expression of a gene or produce an altered protein structure that may be dysfunctional 5 0 . While it is logical to assume that polymorphisms associated with altered disease susceptibility occur in regulatory or coding sequences, this appears to be overly simplistic and underestimates the unknown potential for other SNPs of as yet unknown function to alter outcome. Because of the large number of SNPs and because of variable design of clinical studies, there are critical statistical issues that must be optimized to "prove" that a SNP is associated with a clinical outcome. Another problem with the use of putative functional SNPs (meaning non-synonymous coding SNPs or promoter region SNPs) as the unit of measurement in gene association studies is that we still do not completely understand the functional significance of the vast majority of SNPs. In large genes there are often exonic SNPs and SNPs in known regulatory regions of the gene such as the promoter. However, we still do not have a complete understanding of the regulatory structure of genes. It is now known that SNPs in intronic regions may be of vital importance to the correct splicing of genes. In addition, other SNPs may be important in the three-dimensional regulation of gene transcription and chromatin structure. Furthermore, synonymous exonic SNPs may affect the efficiency of mRNA translation. Without a complete understanding of the regulation of transcription and translation of a gene, an association study strategy based on the selecting probable functional SNPs (non-synonymous and promoter SNPs) as candidates may overlook many polymorphisms essential for proper expression of genes and proteins 5 1 . 7 Another problem with the use of known SNPs as the unit of genetic measurement in gene association studies is that a single SNP approach measures the association of the known SNP and all other SNPs that are in linkage disequilibrium (LD) with the known measured SNP with the clinical phenotype. A single SNP-based genetic association study is optimized when the SNP evaluated is the SNP responsible for a change in phenotype or is in high LD with the causal SNP 5 2 ' 5 3 . Alternatively, a SNP may be mistakenly characterized as a disease susceptibility allele when it is in fact in linkage disequilibrium with, and thus simply marking, the true causal allele 5 3 . Our limited knowledge of transcriptional regulation and the structure of linkage disequilibrium, in addition to incomplete detection of all polymorphic sites within the human genome may be in part responsible for the lack of reproducibility of many genetic association studies in sepsis. The Advantages and Disadvantages of Functional Variants The use of functional variants (potentially deleterious SNPs) as the unit of genetic information in association studies has the advantage of biologic plausibility of the hypothesis, may aid in efficient selection of hypotheses, and in the interpretation of the results of an association study. However, the use of functional variants is limited by our incomplete understanding of gene function, of biologic pathways, of post-translational modification and of gene - gene interaction, and by how all of these processes are altered by genetic variation. 8 Haplotype-based Approach to Genetic Association Studies in Sepsis A comparison of SNP- and haplotype-based genetic association studies is shown in Table 1-1. The evolution from SNP-based association studies to haplotype-based association studies has been facilitated by two factors. First, SNP markers of adequate density have been identified for most genes. Indeed, the Hap Map project aims to identify haplotype blocks across the entire human genome. The second factor leading to use of haplotypes in association studies has been the development and verification of robust linkage disequilibrium (LD) analyses to infer haplotypes from unphased genotype data (data in which alleles at polymorphic sites have not been assigned to a chromosome) of individuals. Our group has taken the haplotype-based approach to narrow down the search for SNPs that are associated with sepsis. That is, we group polymorphisms into haplotypes, which are then tested for association to outcome 5 4 . Haplotypes are sets of SNPs that are in LD with one another within a gene or segment of DNA, and are thus inherited as a unit 5 4 ' 5 5 . A disease causing mutation occurs on the background of an ancestral haplotype and is in complete LD with all other alleles at other polymorphic sites in the haplotype 5 6 . Thus, haplotypes serve as markers of all detected and undetected SNPs within the haplotype 5 5 . Haplotype tag SNPs (htSNPs) that are unique to a specific haplotype can be selected using a number of different algorithms to tag common haplotypes capturing the most genetic diversity based on linkage disequilibrium among SNPs 5 7 ~ 6 0 . Such an approach allows us to test all detected and undetected genetic variation within a gene for association to outcome in sepsis in a more efficient and cost-effective manner than genotyping every detected SNP in a candidate gene region. In using haplotypes as the unit of genetic variation in genetic association studies, no knowledge is required of a SNPs functional effects on transcriptional 9 regulation or protein function 5 3 . If a haplotype is associated with a phenotype, then we can work backwards to identify the causal SNP within a haplotype. A significant advantage of a haplotype compared to SNP-based genetic association study is the increased statistical power to associate genotype with phenotype of haplotype-based studies 5 5 . Importantly, the power of a haplotype-based association test is less subject than a single SNP to the effects of evolutionary forces such as drift, mutation and recombination as the haplotype is passed down through the generations 5 5 . These forces tend to increase the variability of LD between any single SNP and the disease allele. By examining multiple markers simultaneously in a haplotype the overall LD is less variable and exists in simpler patterns 5 5 . The major weakness of a haplotype-based approach is the lack of high-throughout methods to molecularly determine haplotype structure from sequence data 6 1 . High throughput genotyping methods can tell us which two alleles the individual carries, but cannot provide us with haplotype information, or which combination of alleles are present on each chromosome. Haplotypes can only be determined experimentally (ie. by separating paternal and maternal chromosomes before genotyping) through considerable cost or by directly genotyping family members to prove the haplotype composition of each patient (or subject) in a study 6 2 . By directly genotyping pedigrees we can determine the parental haplotypes and use tests such as the transmission-disequilibrium test to measure whether one allele or haplotype is transmitted from a heterozygous parent to an affected offspring more often than would be expected under Mendelian transmission 6 3 . If one allele or haplotype is 63 in excess it is probable that this allele or haplotype is in LD with the disease allele . 10 Because it is not usually possible to obtain DNA from an adequate sample of family members of patients for association studies in sepsis, we must rely on statistical methods to infer probable haplotypes using unphased genotypic data in unrelated individuals (http://pga.mbt.washington.edu, http://www.innateimmunity.net). There has been enormous progress in optimizing this indirect approach to determination of haplotypes. The software program PHASE predicts probable haplotypes and estimates the uncertainty associated with each phase call 6 4 . PHASE has been shown to be highly accurate compared to direct determination of haplotypes in family-based studies, and thus reconstruction of haplotypes experimentally or by genotyping additional family members may not be necessary 6 4 . Haplotypes are tagged by haplotype tag SNPs that are selected by a number of methods. All of the methods share a common basis; there is strong linkage disequilibrium and limited haplotype diversity in small genomic regions of a gene and its adjacent upstream and downstream segments. One efficiency of the use of haplotype tag SNPs is that once the haplotype tag SNPs are selected for the known haplotypes, there is essentially little or no yield to increasing the number of haplotype tag SNPs. That is, a test for the association of an allele with an altered clinical phenotype does not increase the number of degrees of freedom (which would alter statistical power) when the haplotype diversity has been completely reflected by the haplotype tag SNPs (when every SNP in the haplotype is either directly tested or is completely represented by a tag SNP). To date, the use of indirect association by use of inferred haplotypes and haplotype tag SNPs is the most cost efficient approach to studying association of an entire candidate gene with altered clinical phenotype. The efficiency is particularly helpful when studying a large number of candidate genes such as in a known inflammatory pathway. One minor 11 limitation of this haplotype-based approach in gene association studies is that rare (less than 5 percent frequency) variants may not be detected, thus running the risk of false negative studies if the rare variant is highly penetrant. Cladistic Analysis in Genetic Association Studies in Sepsis While most SNPs have only two alleles to compare, multiple haplotypes exist for all studied genes. Many haplotypes are rare (<5%) and testing multiple rare haplotypes for association to phenotype reduces power to find an association. To correct for this loss of power when studying rare haplotypes, haplotypes can be grouped into clades based on their evolutionary history 6 0 . Haplotype clades are evolutionarily-related groups of haplotypes which have shared ancestry in the chromosomal region near the disease allele 5 < 5. As the original founder haplotype on which the disease mutation occurs is passed down through generations, LD is broken down (ie. decreases) and rare new haplotypes are generated 5 6 . Recombination occurs less frequently over small genetic distances and so the founder haplotype is maintained close to (in high LD with) the disease allele. Haplotypes carrying the same disease mutation share a more recent common ancestor (MRCA) than haplotypes carrying different alleles and so will be grouped together in phylogenetic analysis 5 6 Thus, grouping haplotypes into clades increases power to associate genotype with clinical phenotype and susceptibility to disease 6 0 . MEGA2 is a robust, reproducible, widely-accepted molecular evolutionary genetics analysis software package we have used to make pair wise comparisons between the sequences of haplotypes generated by PHASE to generate phylogenetic trees outlining the relationships of the haplotypes 6 5 . M E G A 2 uses traditional phylogeny reconstruction. Other software is available that takes into account the presence of recombination within haplotypes, 12 the smaller numbers of variable loci present within a population than among populations, and the fact that the ancestral haplotypes will be the most frequent sequences sampled in a population-based study 6 6 ~ 6 8 . This emerging software may prove more useful in determining the cladistic structure of the human genome. Just as haplotype tag SNPs can be used to tag haplotypes, so SNPs that are unique to a particular haplotype clade can be selected and genotyped to identify haplotypes belonging to each haplotype clade 5 6 ' 6 0 . This use of haplotype tag SNPs to tag haplotype clades further increases the efficiency and cost-effectiveness of genetic association studies. The cladistic approach also increases statistical power. If a particular haplotype clade is found to be associated with susceptibility to sepsis or increased mortality, for example, additional haplotype tag SNPs can be selected and measured to distinguish the individual haplotypes from within a clade. Thus, it is possible to associate one unique haplotype with clinical phenotype. In specific conditions, this is an important further step because grouping haplotypes into clades may dilute the strength of the genotype-phenotype association if not all the haplotypes of a clade carry the disease allele. Population Substructure and False Positive Genetic Association One of the most important problems to be aware of in genetic association studies is the risk of false-positive associations due to population substructure, also known as population admixture and population stratification 6 9 " 7 2 . Ethnic heterogeneity within a study population can lead to spurious associations between genotype and phenotype 7 0 ' 7 1 . Random genetic drift and new mutations render substantial differences in allele frequency between populations that have been isolated for many generations 6 9 . Thus, the cladistic relationships of haplotypes and LD patterns among populations may differ significantly 6 9 . Case-control 13 studies are particularly susceptible to the problem of population substructure if cases and controls are sampled from heterogeneous populations 6 9 . For example, consider an allele that is strongly associated with increased risk of infection in Japanese patients but is not be associated with increased risk of infection in Caucasians (and indeed, may not occur at all in Caucasians). If patients of these two ethnicities were pooled for analysis in the same cohort, an allele that occurs at higher frequency within one ethnic group may be falsely associated with risk of disease if this group has other reasons (genetic or environmental) to increase risk of developing the disease. For example, if smoking (an environmental factor) was more common in one ethnic group than another and if smoking increased the risk of sepsis, then a pure genetic association study could find a higher risk of SNPs that differ in allele frequency between ethnicities and increased risk of sepsis. However, this would be a false positive genetic association of the SNPs with disease because the association was actually between smoking and disease. At this time we include only Caucasians in our cohort studies of critically ill adults to decrease the risk of false positive associations due to population stratification. As we increase the sample size of other ethnicities within our cohort, we will have to generate population-specific haplotypes and test for the association of these haplotypes to outcome of sepsis in ethnically homogenous cohorts. 1.4 Hypothesis My central hypothesis is: Allelic variants of key innate immunity and inflammatory genes alter protein levels and function, and are predictive of risk of severe infection and of outcome (mortality and organ dysfunction) in critically ill adults who have SIRS. 14 1.5 Objectives and Experimental Approach 1. To associate haplotypes and haplotype clades of key innate immunity and inflammatory genes with risk of severe infection and outcome in a derivation (hypothesis generating) cohort of critically ill patients with SIRS. 2. To genotype haplotypes within innate immunity and inflammatory genes found to be associated with adverse outcome in the critically ill derivation cohort in a separate validation (hypothesis testing) cohort of critically ill patients in order to validate associations between genotype and phenotype. 3. To genotype haplotypes within innate immunity and inflammatory genes found to be associated with adverse outcome in the critically ill derivation cohort in a separate cohort of cardiac surgery patients and to measure the pre- and post-operative serum concentrations of cytokines in these patients in order to validate associations between genotype, intermediate phenotype and clinical outcome (ie. to establish biological plausibility). 4. To measure and compare activation of the immune response in cells of different IRAK4 genotype after stimulation with bacterial products to elucidate the mechanism behind the association of the novel non-synonymous SNP IRAK4 G29429A (Ala428Thr) with increased risk of Gram-positive infection. The first step in a genetic association study is to carefully select candidate genes for association to clinical measures. There are thousands of genes that play a role in the immune response against severe infection and allelic variants in any number of these genes may influence a patient's outcome from critical illness. A small number of candidate genes had to 15 be selected. In order to examine the effect of genetic variation in a set of genes representative of different aspects of host clearance of microorganisms three innate immunity pattern recognitions receptors, one intermediate signaling molecule and a downstream effector molecule were selected for investigation. The innate immunity genes Toll-like receptor 2 (TLR2), mannose-binding lectin (MBL) and cluster of differentiation 14 (CD14) were chosen as candidate genes because they play vital roles in the initial recognition of invading pathogens. TLR2 recognizes peptidoglycan and lipotechoic acid from Gram positive bacteria 7 3 ' 1 4 . M B L is an acute phase serum protein that recognizes mannose groups a wide array of microorganisms and activates the lectin complement pathway in response n ' 7 5 . CD14 is found on the membranes of monocytes/macrophages, neutrophils and hepatocytes and also as a soluble serum protein 7 6 . With TLR4 and MD2, CD14 forms the LPS recognition complex 7 4 . CD14 does not have an intracellular domain and so requires interaction with TLR4 to induce signaling events in response to LPS binding 7 4 . Variants of these genes may alter a patient's ability to recognize and clear pathogens. Interleukin-1 receptor associated kinase 4 (IRAK4) was chosen as a candidate gene as it is the central signaling molecule through which toll-like receptors and the inflammatory cytokine interleukin-1 signal 3 6 . Variants of 1RAK4 may alter an individual's ability to activate a response to the recognition of pathogens by TLRs. The pluripotent cytokine Interleukin-6 (IL-6) was selected as it is a key downstream effector molecule in immunity and healing. IL-6 is involved in the regulation of the acute inflammatory response and in modulation of specific immune responses including B- and T-lymphocyte cell differentiation 7 7 . Peripheral blood concentration of IL-6 has been correlated to outcome in sepsis and septic shock 7 8 ' 7 9 ' . SNPs in these genes have been well-characterized at high density by the SeattleSNPs Program for Genomic Applications. The SeattleSNPs program re-sequenced hundreds of candidate gene regions in the DNA of 42 healthy Caucasian and African American 16 individuals from the Coriell Cell Repository. Tens of thousands of SNPs have been catalogued and the data made publicly available (NHLBI Programs for Genomics Applications: http://pga.mbt.washington.edu, http://www.innateimmunity.net). However, there are too many SNPs in each gene to assess seperately and we do not know their potential functions. In order to avoid making prior assumptions about the functional significance of SNPs, and to optimize the power to detect associations between genetic variants and clinical outcome (See section 1.3), haplotype clades were used as the unit of genetic variation to test for association to clinical outcomes from critical illness. Haplotypes of TLR2, M B L , CD14, IRAK4 and IL-6 were inferred using PHASE software and the cladistic structure was determined using MEGA2 software. A minimum set of htSNPs that define all major haplotype clades of the selected genes (meaning one htSNP to define each haplotype clade) was then selected for further genotyping. Patients' genotypes at the selected htSNPs were determined by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP), MALDItof, or real-time PCR assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System. For quality assurance and quality control samples of known genotype from the Coriell Cell Repository were genotyped with patient samples and 10% of the patient samples were re-genotyped. Samples that were genotyped by more than one method were in complete concordance. Once candidate genes and htSNPs of the candidate genes were selected a prospective derivation cohort of 250 critically ill Caucasian patients was recruited from the tertiary mixed medical-surgical Intensive Care Unit (ICU) of St. Paul's Hospital in which to test the hypothesis that allelic variants of innate immunity and inflammatory genes are associated with risk of severe infection and clinical outcome from critical illness. Patients were included in the derivation cohort if they had had at least 2 of 4 SIRS criteria. Because the 17 selected innate immunity candidate genes are important in the early recognition of invading microorganisms and the initiation of the immune response, the prevalence, type and source of infections at admission to ICU were recorded and tested for association to innate immunity gene polymorphisms. All patients had microbiological cultures taken at admission to ICU. Patients were categorized as having an infection at admission to ICU if any of their cultures at admission were positive, or if they were suspected of having an infection and put on antibiotics by the attending physician. Infections were categorized as Gram-positive, Gram-negative, mixed, fungal or other. Sources of infection were categorized as respiratory (sputum), gastrointestinal, genitourinary (urine), endovascular (blood, lines, valves), or from skin, soft tissues or wounds. The association of specific innate immunity gene htSNPs to risk, type and source of infection was then determined by Chi-square analysis. The presence, type and source of infections after admission to ICU were not recorded. Additionally, clinical data was collected over 28 days after ICU admission to measure primary and secondary outcome variables. The primary outcome variable was 28-day survival, and the secondary outcome variables were days alive and free of organ dysfunction and organ support, prevalence of infection, and days alive and free of SIRS (Table 1-2). Patients were given a score of one day alive and free (DAF) for every day that they were both alive and free of organ dysfunction (normal or mild dysfunction), organ support, or SIRS. DAF was scored as 0 if the patient had organ dysfunction (moderate or worse) or was not alive. Every day during the 28-day observation after ICU admission was scored in this way. Thus, the lowest score possible for each variable is zero and the highest score possible is 28. A low score indicates more organ dysfunction, while a high score indicates less organ dysfunction. This scoring system attempts to take into account the confounding effect of early deaths on organ dysfunction scoring. 18 In order to test for the association of genetic variants of TLR2, M B L , CD14, IRAK4 and IL-6 with clinical outcome, prevalence of positive microbiological cultures at admission to ICU and rates of dichotomous outcomes (28-day mortality, sepsis and shock at onset of SIRS) were compared among haplotype clades using a Chi-squared test. Differences in continuous outcome variables were compared among haplotype clades by ANOVA. Data that was not normally distributed was logarithmically (logio) transformed. To adjust for other confounders on mortality (age, sex, and medical vs. surgical diagnosis), 28-day mortality was further compared among haplotype clades using a Cox regression analysis together with Kaplan-Meier analysis of censored survival data. To measure the risk of poor outcome associated with carrying a specific haplotype clade, haplotype clade relative-risk was calculated. Genotype distributions were tested for Hardy-Weinberg equilibrium to check for population stratification. Genetic association studies are beset by a number of unique problems. One important problem faced by all genetic association studies is lack of reproducibility. This lack of reproducibility may be caused by false-positive associations resulting from multiple testing, by small or heterogeneous sample groups or by poorly defined phenotypes. In order to validate our results and minimize these problems, haplotype clades found to be associated with clinical outcome in the derivation cohort of critically ill adults were tested for association to clinical outcome in a larger separate validation cohort of critically ill adults. The larger validation cohort had increased power to detect associations between genotype and phenotype, and reduced the chance of type I error due to multiple testing because only associations that were previously found to be statistically significant in the derivation cohort were tested in the validation cohort. 19 Another problem faced by genetic association studies is that the association of a genetic variant with a clinical outcome measure does not provide a biological mechanism for the association. In particular, the use of haplotype clades as the unit of genetic variation in a genetic association study does not indicate that a specific functional SNPS is causing a change in phenotype. Therefore, in addition to validating genetic association studies in separate cohorts to show statistical plausibility, it is important to assess the biological plausibility of any association. One way of doing this is by associating genotype with an intermediate phenotype, for example, patients' serum cytokine concentrations. However, meaningful comparisons of serum cytokine concentrations among critically ill patients are difficult as patients may experience a systemic inflammatory response syndrome or sepsis as a result of many different inflammatory stimuli, infectious pathogens or surgical complications. Additionally, it may not be known precisely when the inflammatory response was initiated. So as to obtain more meaningful comparisons of serum cytokine concentrations among genotypes, a separate cohort of 1 0 0 0 patients undergoing elective coronary artery bypass graft (CABG) surgery with cardiopulmonary bypass (CPB) was recruited. Haplotype clades of innate immunity and inflammatory cytokine genes that were found to be associated with altered clinical outcome in patients with sepsis should also be clinically relevant in patients who experienced another cause of SIRS; cardiovascular surgery with CPB. CPB and associated cerebral injury trigger multiple pathways, including activation of complement 8 0 , coagulation and inflammatory 8 1 cascades, resulting in a systemic inflammatory response. The complement, coagulation, and inflammatory cascades are key interacting components of the innate immune response 8 2 . Every cardiac surgery patient experienced the same inflammatory stimulus (CABG surgery with CPB), and the exact time of the inflammatory stimulus was known and recorded. Therefore, it was possible to gather clinical data and blood samples at precise time-points before and after surgery. 2 0 Haplotype tag SNPs of haplotype clades found to be associated with clinical outcome in critically ill patients were genotyped in the cohort of cardiac surgery patients and tested for association to clinical and intermediate phenotypes in order to lend biological plausibility to genetic associations detected in the cohorts of critically ill patients. The primary outcome variable in the cardiovascular surgery cohort was the occurrence of a post-surgical vasodilatory syndrome resulting from the inflammatory response. The frequency of the occurrence of this vasodilatory syndrome was compared among haplotype clades using a Chi-squared test. As an intermediate phenotype serum concentrations of inflammatory mediators were measured in a subset of the cardiopulmonary bypass patients. Serum levels of inflammatory cytokines were determined in 200 CPB patients before surgery and 0, 4, 12 and 24 hours post-surgery using a Luminex Immunoassy/Liquichip system for bead-based protein assays, and were compared at each time point among haplotype clades by ANOVA. The association of clinically important haplotypes with altered serum concentrations of inflammatory mediators may provide insight as to the mechanism by which genetic variation alters clinical outcome in inflammatory diseases. Clinically important genotypes may act by altering inflammatory mediator levels. Beyond just demonstrating biological plausibility for a genetic association, it is vital to demonstrate a mechanism for the association before the association may be considered real. In an attempt to elucidate a mechanism for the association between the IRAK4 haplotype clade marked by the htSNP G29429A (Ala428Thr) and increased prevalence of Gram-positive infection at admission to ICU in critically ill patients, B-lymphocyte cell lines and fibroblasts of known 1RAK4 genotypes were stimulated with TLR ligands and their immune response was measured and compared by IRAK4 haplotype clade. B-lymphocyte 21 cell lines (n=43 unique human samples) from the Coriell Cell Repository were used for these experiments because they were readily available and their genotype was known. B-lymphocytes, however, do not express TLR2 or TLR4, and so CpG, a synthetic bacterial DNA sequence, was used to stimulate the immune response of B-lymphocytes through TLR9. The B-lymphocytes were stimulated with lOuM or 50uM CpG for 24 hours and IL-6 secretion in the supernatant was measured by ELISA as a measure of immune response signaling through IRAK4. IL-6 is an easily measured indicator of the immune response as it is produced neither too early in the inflammatory response to practically measure, nor too late. The concentrations of IL-6 measured in the supernatant of the CpG-stimulated B-lymphocyte cell lines were not normally distributed and so were logarithmically transformed and compared between IRAK4 G29429A alleles by Student's t-test. The B-lymphocyte cell lines are from 43 unique individuals and so vary at many different SNP sites, and these variations may interact with the G29429A polymorphism. In order to eliminate the effects of potentially thousands of different polymorphisms interacting with the IRAK4 G29429A polymorphism, constructs containing the two different IRAK4 G29429A variants were transfected into genetically identical IRAK4-deficient dermal fibroblast cell lines. The IRAK4-deficient fibroblasts had been isolated from a local pediatric patient with a genetic IRAK4 deficiency ~. While fibroblasts are not the primary cells of the innate immunity system, they express TLR2, TLR4 and CD 14 and normal fibroblasts produce a cytokine response to TLR ligands . Fibroblasts serve important functions in the immune response such as walling off infections and tissue repair. The immune response of the IRAK4-deficient fibroblasts transfected with IRAK4 constructs carrying the 29429G or A allele was quantified following stimulation with lipopolysaccharide (LPS) or peptidoglycan (PGN) by measuring supernatant IL-6 concentration by ELISA. Supernatant IL-6 22 concentration between cell lines transfected with either the 29429A or G allele was compared by Student's t-test. Using this experimental approach we hope to detect several genetic variants of key innate immunity and inflammatory genes that may contribute to alterations in immune response in vitro and in vivo. Knowing which genetic variants contribute to patients' risk of adverse outcome, and through what molecular mechanisms they act, may one day allow individualized application of anti-inflammatory therapy to improve clinical outcome of sepsis and other inflammatory diseases. 23 1.6 Tables and Figures Table 1-1. Comparison of SNP-based and haplotype-based genetic association studies SNP-based Haplotype-based • Candidate gene selection • Candidate gene selection • SNPs of known or presumed functional significance • Assumes SNP evaluated is most likely the causal SNP • Misses associations of SNPs not in linkage disequilibrium with selected measured SNP(s) • Limited power • Need for intermediate phenotype studies for biologic plausibility • Haplotype tag SNPs to mark haplotypes and haplotype clades • No prior assumptions about functional significance of SNPs • Power optimized by use of haplotypes and clades • SNP markers of adequate density available • Need for intermediate phenotype studies for biologic plausibility • Haplotypes are usually inferred • If haplotype is associated with outcome, need for further studies to identify causal SNP and functional significance 24 Table 1-2. Clinical phenotypes used as primary and secondary outcome variables in the derivation and validation cohorts of critically ill Caucasians Outcome Variable A. Risk of Infection at Admission to ICU - Prevalence of positive microbiological cultures and infection Source of positive microbiological cultures B. Primary Outcome Variables Survival C. Secondary Outcome Variables - Organ dysfunction - Duration of SIRS Measurement - % of patients with positive microbiological cultures or with clinical evidence of infection (ie. on antibiotics) at admission to ICU; % Gram-positive, % Gram-negative - Positive culture from blood or lines, sputum, urine, abscess fluid, cerebrospinal fluid or wound drainage fluid - 28-day survival - Days alive and free of organ dysfunction (Cardiovascular, respiratory, renal, hepatic, central nervous system) - Days alive and free of organ system support (vasopressors, ventilator, dialysis) - Days alive and free or 2/4, 3/4 or 4/4 SIRS criteria 2 25 Table 1-3. Human Toll-like receptors (TLRs) and their ligands TLR Ligands TLR1/2 Triacyl lipopeptides (Pam3CSK4) (Bacteria, Mycobacteria)84 TLR2 Lipoprotein/lipopeptides 1 8 ' 8 5 Peptidoglycan (Gram-positive bacteria)73 73 Lipoteichoic acid (Gram-positive bacteria) Zymosan (Fungi) 8 6 TLR3 Poly (I-C) dsRNA (Virus) 8 7 TLR4 LPS 8 8 TLR5 Flagellin (Flagellated bacteria) TLR6/2 Diacyl lipopeptides (Mycoplasma) 9 0 TLR7 Synthetic compounds ' TLR8 Synthetic compounds 9 2 TLR9 Unmethylated CpG DNA (Bacteria, virus, yeast) ' 26 Figure 1-1. Figure 1-1 has been removed because of copyright restrictions. Original Source: Holmes CL, Russell JA, Walley KR. Chest. 2003 Sep;124(3):l 103-15. Figure 1-1. Innate immunity, inflammation and coagulation in systemic inflammatory response syndrome. Gram positive and Gram negative bacteria have cellular components such as lipotechoic acid, LPS and zymosan that bind to pattern recognition proteins of the innate immunity pathway including CD14 and TLR4. This leads to complex intracellular signaling ultimately causing transcription factors such as N FKB to move from the cytoplasm to the nucleus and then to increase synthesis of cytokines such as TNF-a, IL-1, IL-6, and IL-10. The pro-inflammatory cytokines (TNF-a, IL-1 and IL-6) directly and indirectly injure endothelium by increasing binding of neutrophils to the endothelium. TNF-a, IL-1 and IL-6 also stimulate .coagulation by increasing tissue factor (TF) which activates the extrinsic pathway of coagulation. TF ultimately generates a thrombin, a thrombin, factor Xa, and fibrin increase TNF-a, IL-1, and IL-8 synthesis and release from peripheral blood mononuclear cells and endothelial cells, a thrombin converts fibrinogen (not shown) to fibrin and generates fibrin clots. Fibrin clots are broken down by fibrinolysis by plasminogen activator (not shown) actions on plasminogen. Plasminogen activator is inhibited by plasminogen activator inhibitor-1 (PAI-1). Activated protein C has potentially beneficial actions in the vascular bed of the lung and other tissues because it inhibits factors Va and Xa which limits a.-thrombin generation, inhibits PAI-1, and limits binding of neutrophils to endothelium. 27 Figure 1-2. Act ivat ion of Innate Immunity Receptors by Pathogens. Activation of innate immune receptors leads to activation of the inflammatory response. Lipopolysaccharide (LPS) from Gram negative bacteria bind Lipopolysaccharide-binding protein (LPB) in the blood 1 3 , M . LBP transfers LPS to soluble and membrane bound CD 14, enhancing sensitivity to LPS 100-1000 fold 1 3 ' 1 4 . CD14 binds LPS and interacts with Toll-like receptor 4 (TLR4)/MD2 which activate intracellular signalling pathways (See Figure 2). TLR2 recognizes and binds Lipotechoic acid (LTA) from Gram positive bacteria and activated intracellular signalling pathways 8 5 ' 9 4 , 9 5 (See Figure 2). Mannose-binding lectin (MBL) binds mannose groups on a wide variety of pathogens and activates the "alternative" or lectin complement pathway 1 1 2 . CpG, a bacterial DNA motif, is recognized by intracellular TLR9 96, 97 28 Figure 1-3. Activation of Innate Immunity Signaling Pathways through Toll-like Receptors 2 and 4 (TLR2 and TLR4). Lipopolysaccharide (LPS), peptidoglycan (PPG) and other bacterial cell wall components bind mannose binding lectin (MBL), CD 14, and lipopolysaccharide-binding protein (LBP) in serum, and toll-like receptors on immune cells and activate intracellular signalling pathways 9 8 ' 9 9 . LPS activation of TLR4 and PPG activation of TLR2 are shown as a prototype for innate immunity signalling. LPS binds LBP in the serum 1 3 , 1 4 . LBP transfers LPS to soluble and membrane bound CD14 , 3 i ' 1 4 CD14 has no intracellular signalling domain, and so interacts with TLR4/MD2 to activate intracellular signalling15"18. MD2 is an adaptor protein that is required for TLR4 signalling 1 9 . PPG binds TLR2. MyD88 interacts with the TLR4/MD2 or TLR2 receptor complex by its TIR domain and recruits the serine/threonine IL1 receptor associated kinases 1 and 4 (IRAKI and 4) to the receptor complex 3 4 ' 3 5 . IRAKI and 4 are phosphorylated and associate with TNF receptor associated factor 6 (TRAF6) activating signalling cascades ultimately leading to 34 35 37 39 phosphorylation of IKKp and activation of N F - K B Cytokine, chemokine, cell-adhesion molecule, nitric oxide synthase and procoagulant protein expression is controlled by the I K K - N F - K B signalling pathway. Mitogen-activated protein kinase pathways are also activated and induce activation of the stress kinases ERK1/2, JNK, and p38 that have minor roles in controlling transcription of various inflammatory molecules 4 ( M 2 . 29 1.7 References 1. Brun-Buisson C. The epidemiology of the systemic inflammatory response. Intensive Care Med 2000;26 Suppl 1 :S64-74. 2. Bone RC. The sepsis syndrome. Definition and general approach to management. Clin Chest Med 1996; 17:175-81. 3. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001 ;29:1303-10. 4. Bernard GR, Vincent JL, Laterre PF et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001 ;344:699-709. 5. Martin GS, Mannino D M , Eaton S, Moss M . The epidemiology of sepsis in the United States from 1979 through 2000. NEnglJMed2003;348:1546-54. 6. Angus DC, Wax RS. Epidemiology of sepsis: an update. Crit Care Med 2001;29:S109-16. 7. Burgner D, Levin M. Genetic susceptibility to infectious diseases. Pediatr Infect DisJ2003;22:1-6. 8. Bellamy R, Hill AV. Genetic susceptibility to mycobacteria and other infectious pathogens in humans. Curr Opin Immunol 1998;10:483-7. 9. Choi EH, Zimmerman PA, Foster CB et al. Genetic polymorphisms in molecules of innate immunity and susceptibility to infection with Wuchereria bancrofti in South India. Genes Immun 2001;2:248-53. 10. Sorensen TI, Nielsen GG, Andersen PK, Teasdale TW. Genetic and environmental influences on premature death in adult adoptees. N Engl J Med 1988;318:727-32. 11. Tabona P, Mellor A, Summerfield JA. Mannose binding protein is involved in first-line host defence: evidence from transgenic mice. Immunology 1995;85:153-9. 12. Tenner AJ, Robinson SL, Ezekowitz RA. Mannose binding protein (MBP) enhances mononuclear phagocyte function via a receptor that contains the 126,000 M(r) component of the Clq receptor. Immunity 1995;3:485-93. 13. Hailman E, Lichenstein HS, Wurfel M M et al. Lipopolysaccharide (LPS)-binding protein accelerates the binding of LPS to CD14. J Exp Med 1994; 179:269.-77. 14. Schumann RR, Zweigner J. A novel acute-phase marker: lipopolysaccharide binding protein (LBP). Clin Chem Lab Med 1999;37:271-4. 15. Chow JC, Young DW, Golenbock DT, Christ WJ, Gusovsky F. Toll-like receptor-4 mediates lipopolysaccharide-induced signal transduction. J Biol Chem 1999;274:10689-92. 16. Wright SD, Ramos RA, Tobias PS, Ulevitch RJ, Mathison JC. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 1990;249:1431-3. 17. Labeta MO, Durieux JJ, Fernandez N, Herrmann R, Ferrara P. Release from a human monocyte-like cell line of two different soluble forms of the lipopolysaccharide receptor, CD14. Eur J Immunol 1993;23:2144-51. 18. Brightbill HD, Libraty DH, Krutzik SR et al. Host defense mechanisms triggered by microbial lipoproteins through toll-like receptors. Science 1999;285:732-6. 19. Shimazu R, Akashi S, Ogata H et al. MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like receptor 4. J Exp Med 1999;189:1777-82. 30 20. Hashimoto C, Hudson K L , Anderson KV. The Toll gene of Drosophila, required for dorsal-ventral embryonic polarity, appears to encode a transmembrane protein. Cell 1988;52:269-79. 21. Lemaitre B, Nicolas E, Michaut L, Reichhart JM, Hoffmann JA. The dorsoventral regulatory gene cassette spatzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 1996;86:973-83. 22. Medzhitov R, Preston-Hurlburt P, Janeway CA, Jr. A human homologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 1997;388:394-7. 23. Rock FL, Hardiman G, Timans JC, Kastelein RA, Bazan JF. A family of human receptors structurally related to Drosophila Toll. Proc Natl Acad Sci USA 1998;95:588-93. 24. Takeuchi O, Kawai T, Sanjo H et al. TLR6: A novel member of an expanding toll-like receptor family. Gene 1999;231:59-65. 25. Chuang TH, Ulevitch RJ. Cloning and characterization of a sub-family of human toll-like receptors: hTLR7, hTLR8 and hTLR9. Eur Cytokine Netw 2000;11:372-8. 26. Chuang T, Ulevitch RJ. Identification of hTLRlO: a novel human Toll-like receptor preferentially expressed in immune cells. Biochim Biophys Acta 2001 ;1518:157-61. 27. Du X, Poltorak A, Wei Y, Beutler B. Three novel mammalian toll-like receptors: gene structure, expression, and evolution. Eur Cytokine Netw 2000;11:362-71. 28. Bell JK, Mullen GE, Leifer CA, Mazzoni A, Davies DR, Segal DM. Leucine-rich repeats and pathogen recognition in Toll-like receptors. Trends Immunol 2003;24:528-33. 29. Slack JL, Schooley K, Bonnert TP et al. Identification of two major sites in the type I interleukin-1 receptor cytoplasmic region responsible for coupling to pro-inflammatory signaling pathways. J Biol Chem 2000;275:4670-8. 30. Re F, Strominger JL. Toll-like receptor 2 (TLR2) and TLR4 differentially activate human dendritic cells. J Biol Chem 2001 ;276:37692-9. 31. Wang Q, Dziarski R, Kirschning CJ, Muzio M , Gupta D. Micrococci and peptidoglycan activate TLR2—>MyD88->IRAK—>TRAF—>NIK—>IKK~>NF-kappaB signal transduction pathway that induces transcription of interleukin-8. Infect Immun 2001 ;69:2270-6. 32. Liu Y, Wang Y, Yamakuchi M et al. Upregulation of toll-like receptor 2 gene expression in macrophage response to peptidoglycan and high concentration of lipopolysaccharide is involved in NF-kappab activation. Infect Immun 2001;69:2788-96. 33. Arbibe L, Mira JP, Teusch N et al. Toll-like receptor 2-mediated NF-kappa B activation requires a Racl-dependent pathway. Nat Immunol 2000;1:533-40. 34. Li X, Qin J. Modulation of Toll-interleukin 1 receptor mediated signaling. J Mol Med 2005. 35. Palsson-McDermott E M , O'Neill LA. Signal transduction by the lipopolysaccharide receptor, Toll-like receptor-4. Immunology 2004;113:153-62. 36. Suzuki N, Suzuki S, Yeh WC. IRAK-4 as the central TIR signaling mediator in innate immunity. Trends Immunol 2002;23:503-6. 37. Dower SK, Qwarnstrom EE. Signalling networks, inflammation and innate immunity. Biochem Soc Trans 2003;31:1462-71. 38. Muzio M , Polntarutti N , Bosisio D, Prahladan MK, Mantovani A. Toll like receptor family (TLT) and signalling pathway. Eur Cytokine Netw 2000; 11:489-90. 39. Serbina NV, Kuziel W, Flavell R, Akira S, Rollins B, Pamer EG. Sequential . MyD88-independent and -dependent activation of innate immune responses to intracellular bacterial infection. Immunity 2003;19:891-901. 31 40. Beutler B. Inferences, questions and possibilities in Toll-like receptor signalling. Nature 2004;430:257-63. 41. Matsuguchi T, Masuda A, Sugimoto K, Nagai Y, Yoshikai Y. JNK-interacting protein 3 associates with Toll-like receptor 4 and is involved in LPS-mediated INK activation. Embo J2003;22:4455-64. 42. Cao Z, Henzel WJ, Gao X. IRAK: a kinase associated with the interleukin-1 receptor. Science 1996;271:1128-31. 43. Child NJ, Yang IA, Pulletz MC et al. Polymorphisms in Toll-like receptor 4 and the systemic inflammatory response syndrome. Biochem Soc Trans 2003;31:652-3. 44. Garred P, Madsen HO, Marquart H et al. Two edged role of mannose binding lectin in rheumatoid arthritis: a cross sectional study. J Rheumatol 2000;27:26-34. 45. Soborg C, Madsen HO, Andersen AB, Lillebaek T, Kok-Jensen A, Garred P. Mannose-binding lectin polymorphisms in clinical tuberculosis. J Infect Dis 2003;188:777-82. 46. Lin MT, Albertson TE. Genomic polymorphisms in sepsis. Crit Care Med 2004;32:569-79. 47. Sachidanandam R, Weissman D, Schmidt SC et al. A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature 2001;409:928-33. 48. Collins FS, Guyer MS, Charkravarti A. Variations on a theme: cataloging human DNA sequence variation. Science 1997;278:1580-1. 49. Zondervan KT, Cardon LR. The complex interplay among factors that influence allelic association. Nat Rev Genet 2004;5:89-100. 50. Risch N, Merikangas K. The future of genetic studies of complex human diseases. Science 1996;273:1516-7. 51. Chakravarti A. It's raining SNPs, hallelujah? Nat Genet 1998;19:216-7. 52. Long AD, Grote MN, Langley CH. Genetic analysis of complex diseases. Science 1997;275:1328; author reply 1329-30. 53. Long AD, Langley CH. The power of association studies to detect the contribution of candidate genetic loci to variation in complex traits. Genome Res 1999;9:720-31. 54. Zhang K, Calabrese P, Nordborg M , Sun F. Haplotype block structure and its applications to association studies: power and study designs. Am J Hum Genet 2002;71:1386-94. 55. Akey J, Jin L, Xiong M. Haplotypes vs single marker linkage disequilibrium tests: what do we gaml Eur J Hum Genet 2001;9:291-300. 56. Durrant C, Zondervan KT, Cardon LR, Hunt S, Deloukas P, Morris AP. Linkage disequilibrium mapping via cladistic analysis of single-nucleotide polymorphism haplotypes. Am J Hum Genet 2004;75:35-43. 57. Cousin E , Genin E , Mace S et al. Association studies in candidate genes: strategies to select SNPs to be tested. Hum Hered 2003;56:151 -9. 58. Gabriel SB, Schaffher SF, Nguyen H et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225-9. 59. Johnson GC, Esposito L, Barratt BJ et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001 ;29:233-7. 60. Templeton AR, Weiss K M , Nickerson DA, Boerwinkle E , Sing CF. Cladistic structure within the human lipoprotein lipase gene and its implications for phenotypic association studies. Genetics 2000;156:1259-75. 32 61. Kwok PY, Xiao M. Single-molecule analysis for molecular haplotyping. Hum Mutat 2004;23:442-6. 62. Sobel E, Lange K. Descent graphs in pedigree analysis: applications to haplotyping, location scores, and marker-sharing statistics. Am J Hum Genet 1996;58:1323-37. 63. Laird NM, Lange C. Family-based designs in the age of large-scale gene-association studies. Nat Rev Genet 2006;7:385-94. 64. Stephens M , Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001;68:978-89. 65. Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bio informatics 2001;17:1244-5. 66. Crandall KA, Templeton AR. Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics 1993;134:959-69. 67. Watterson GA, Guess HA. Is the most frequent allele the oldest? Theor Popul Biol 1977;11:141-60. 68. Clement M , Posada D, Crandall KA. TCS: a computer program to estimate gene genealogies. Mol Ecol 2000;9:1657-9. 69. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet 2001;2:91 -9. 70. Cardon LR, Palmer LJ. Population stratification and spurious allelic association. Lancet 2003;361:598-604. 71. Marchini J, Cardon LR, Phillips MS, Donnelly P. The effects of human population structure on large genetic association studies. Nat Genet 2004;36:512-7. 72. Freedman M L , Reich D, Penney K L et al. Assessing the impact of population stratification on genetic association studies. Nat Genet 2004;36:388-93. 73. Schwandner R, Dziarski R, Wesche H, Rothe M , Kirschning CJ. Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll-like receptor 2. JBiol Chem 1999;274:17406-9. 74. Aderem A, Ulevitch RJ. Toll-like receptors in the induction of the innate immune response. Nature 2000;406:782-7. 75. Gordon A C , Waheed U, Hansen T K et al. Mannose-binding lectin polymorphisms in severe sepsis: relationship to levels, incidence, and outcome. Shock 2006;25:88-93. 76. Griffin JD, Ritz J, Nadler L M , Schlossman SF. Expression of myeloid differentiation antigens on normal and malignant myeloid cells. J Clin Invest 1981;68:932-41. 77. Borden EC, Chin P. Interleukin-6: a cytokine with potential diagnostic and therapeutic roles. J Lab Clin Med 1994;123:824-9. 78. Martin C, Boisson C, Haccoun M, Thomachot L, Mege JL. Patterns of cytokine evolution (tumor necrosis factor-alpha and interleukin-6) after septic shock, hemorrhagic shock, and severe trauma. Crit Care Med 1997;25.T 813-9. 79. Hack CE, De Groot ER, Felt-Bersma RJ et al. Increased plasma levels of interleukin-6 in sepsis. Blood 1989;74:1704-10. 80. Chenoweth DE, Cooper SW, Hugh TE, Stewart RW, Blackstone EH, Kirklin JW. Complement activation during cardiopulmonary bypass: evidence for generation of C3a and C5a anaphylatoxins. N EnglJ Med 1981;304:497-503. 81. Wan S, LeClerc JL, Vincent JL. Inflammatory response to cardiopulmonary bypass: mechanisms involved and possible therapeutic strategies. Chest 1997;112:676-92. 33 82. Uthaisangsook S, Day NK, Bahna SL, Good RA, Haraguchi S. Innate immunity and its role against infections. Ann Allergy Asthma Immunol 2002;88:253-64; quiz 265-6,318. 83. Davidson DJ, Currie AJ, Bowdish D M et al. IRAK-4 mutation (Q293X): rapid detection and characterization of defective post-transcriptional TLR/IL-1R responses in human myeloid and non-myeloid cells. J Immunol 2006;177:8202-11. 84. Takeuchi O, Sato S, Horiuchi T et al. Cutting edge: role of Toll-like receptor 1 in mediating immune response to microbial lipoproteins. J Immunol 2002;169:10-4. 85. Aliprantis AO, Yang RB, Mark MR et al. Cell activation and apoptosis by bacterial lipoproteins through toll-like receptor-2. Science 1999;285:736-9. 86. Underhill D M , Ozinsky A, Smith KD, Aderem A. Toll-like receptor-2 mediates mycobacteria-induced proinflammatory signaling in macrophages. Proc Natl Acad Sci USA 1999;96:14459-63. 87. Alexopoulou L, Holt A C , Medzhitov R, Flavell RA. Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3. Nature 2001;413:732-8. 88. Poltorak A, He X, Smirnova I et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998;282:2085-8. 89. Mizel SB, West AP, Hantgan RR. Identification of a sequence in human toll-like receptor 5 required for the binding of Gram-negative flagellin. J Biol Chem 2003;278:23624-9. 90. Takeuchi O, Kawai T, Muhlradt PF et al. Discrimination of bacterial lipoproteins by Toll-like receptor 6. Int Immunol 2001 ;13:933-40. 91. Hemmi H, Kaisho T, Takeuchi O et al. Small anti-viral compounds activate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat Immunol 2002;3:196-200. 92. Lee J, Chuang TH, Redecke V et al. Molecular basis for the immunostimulatory activity of guanine nucleoside analogs: activation of Toll-like receptor 7. Proc Natl Acad Sci USA 2003; 100:6646-51. 93. Akira S, Hemmi H. Recognition of pathogen-associated molecular patterns by TLR family. Immunol Lett 2003;85:85-95. 94. Lien E, Sellati TJ, Yoshimura A et al. Toll-like receptor 2 functions as a pattern recognition receptor for diverse bacterial products. J Biol Chem 1999;274:33419-25. 95. Travassos LH, Girardin SE, Philpott DJ et al. Toll-like receptor 2-dependent bacterial sensing does not occur via peptidoglycan recognition. EMBO Rep 2004;5:1000-6. 96. Baiyee EE, Flohe S, Lendemans S et al. Expression and function of Toll-like receptor 9 in severely injured patients prone to sepsis. Clin Exp Immunol 2006;145:456-62. 97. Dasari P, Nicholson IC, Hodge G, Dandie GW, Zola H. Expression of toll-like receptors on B lymphocytes. Cell Immunol 2005;236:140-5. 98. Akira S, Takeda K. Toll-like, receptor signalling. Nat Rev Immunol 2004;4:499-511. 99. Yoshimura A, Lien E, Ingalls RR, Tuomanen E, Dziarski R, Golenbock D. Cutting edge: recognition of Gram-positive bacterial cell wall components by the innate immune system occurs via Toll-like receptor 2. J Immunol 1999;163:1-5. 34 CHAPTER 2: POLYMORPHISMS IN INNATE IMMUNITY RECEPTORS ARE ASSOCIATED WITH INCREASED PREVALENCE OF INFECTION IN CRITICALLY ILL ADULTS 2.1 Introduction The pre-programmed innate immune response contributes to rapid clearance of microorganisms to prevent or contain infection. Activation of the innate immune response to infection varies significantly between individuals 1 - 3 which may have important clinical implications. Genotype has been shown to contribute substantially to outcome of infectious disease 4"6. Thus variation in key innate immunity genes may explain variation in individuals' responses to infection. If so, it would be useful to evaluate genotypes of critically ill patients to understand individual susceptibility and response to infection, which could lead ultimately to specific patient-tailored therapy based on genotype. Cluster of differentiation 14 (CD14) is an innate immunity receptor for lipopolysaccharide (LPS), peptidoglycan and lipoteichoic acid. CD14 is found in association with Toll-like receptors on the surface of monocytes, macrophages, neutrophils, and hepatocytes 7' 8 , and as a soluble form in serum 9 . The serum level of soluble CD14 is increased during septic shock and is associated with greater mortality , 0 ' n . A C-to-T transition at position -159 in the promoter of the CD14 gene has been associated with increased density of membrane-bound CD 14 on monocytes and increased serum levels of soluble CD14 1 2 ' 1 3 . This polymorphism of CD14 may alter host recognition and clearance of pathogens. A version of this chapter has been published. Sutherland A M , Walley KR, Russell JA. Polymorphisms in 35 CD14, mannose-binding lectin, and Toll-like receptor-2 are associated with increased prevalence of infection in critically ill adults. Crit Care Med. 2005 Mar;33(3):638-44. Mannose-binding lectin (MBL) binds repeating arrays of sugar groups on microbial surfaces with its lectin domains 1 4 . M B L activates the "alternative", or lectin, complement pathway in an antibody-independent manner. M B L also is an opsonin that enhances phagocytosis of a variety of microorganisms 1 5 . M B L binds yeasts, viruses and Gram-negative bacteria, and also Gram-positive bacteria, but with low affinity 1 6 " 1 8 . Three structural mutations have been found in exon 1 of the M B L gene at codons 52, 54, and 57 and the minor alleles are referred to as alleles D, B, and C respectively 1 9 ~ 2 1 . The A allele indicates the wild type at each of these positions. G-to-C transversions have also been identified at position -550 (genotype H/L) and -221 (genotype Y/X) in the M B L promoter 90 99 OA region . M B L polymorphisms occur as 6 different haplotypes " . These M B L haplotypes are associated with different serum concentrations of M B L (Table 2-1) 2 2 ~ 2 4 . Individuals having two copies of low M B L level haplotypes have increased incidence of infection and sepsis 2 5 ' 2 6 . Toll-like receptor 2 (TLR2) was previously believed to be a receptor for peptidoglycan from the cells walls of Gram-positive bacteria 2 7 " 2 9 . It is now recognized that commercial preparations of peptidoglycan used in studies of TLR2 activation contained other cell wall components 3 0 , and that TLR2 was activated by these impurities 3 1 . TLR2 is in fact a pattern recognition receptor for lipoteichoic acid from the cell membrane of Gram-positive 31 32 bacteria ' . The receptor is a point of direct contact between the host neutrophils, macrophages/monocytes, and T cells with Gram-positive bacteria 3 3 . Binding of lipoteichoic acid to TLR2 causes an intracellular signaling cascade that ultimately leads to activation of the nuclear factor kappa B ( N F K B ) pathway and increased transcription of pro-inflammatory cytokines 3 2 ~ 3 4 . 36 To-date there have been no studies of the association of SNPs of CD14, M B L and TLR2 with clinical outcomes in a single cohort of critically ill patients. We hypothesized that polymorphisms in the genes of these innate immunity receptor genes are associated with first, increased prevalence of positive bacterial cultures and sepsis, and second, with the type of bacteria (Gram-positive, Gram-negative, other), and finally with prevalence of septic shock and 28-day survival of critically ill patients with systemic inflammatory response syndrome (SIRS). We found that specific alleles of CD14, MBL, and TLR2 were each associated with increased risk of infection and sepsis, and with specific classes of bacteria, but not with altered survival in a single derivation cohort of critically ill patients. In a separate, larger validation cohort we found that C-159T CD14 was not associated with increased risk of infection and sepsis, and with specific classes of bacteria, but was associated with altered survival. 2.2 Methods This study was approved by the Research Ethics Board of Providence Health Care and the University of British Columbia. Study Populations All patients admitted to the intensive care unit (ICU) of St. Paul's Hospital were screened for inclusion. The ICU is a mixed medical-surgical ICU in a tertiary care, university affiliated teaching hospital of the University of British Columbia. Patients were included in the study if they met at least two out of four SIRS criteria: 1) fever (>38°C) or hypothermia (<36 °C), 2) tachycardia (>90 beats/minute), 3) tachypnea (>20 breaths/minute), PaCCh <32 mm Hg, or need for mechanical ventilation, and 4) leukocytosis (total leukocyte count >12,000 mm3) or leukopenia (<4,000 mm3) 3 5 . Patients were included in the cohort on 37 the calendar day on which the SIRS criteria were met. To decrease the confounding influence of population admixture secondary to ethnic diversity on associations between genotype and phenotype, only Caucasian patients were studied 3 6 . Of these consecutive critically ill patients admitted to St. Paul's Hospital ICU, 252 met the inclusion criteria for our study (having at least two out of four SIRS criteria and Caucasian). These patients were genotyped for polymorphisms in CD14, MBL, and TLR2 and were used as our derivation cohort for analysis. A large group of patients admitted to the ICU of St. Paul's Hospital after analysis of the initial derivation cohort were screened for inclusion into the study using the same criteria as above and were used as a validation cohort. Patient data were screened daily and Caucasian patients were included in the validation cohort (n=694) if they met at least two of four SIRS criteria. These patients were genotyped for the CD 14 polymorphism. Clinical Phenotype In all cohorts we assessed risk of infection as the prevalence of a positive bacterial culture at admission to ICU. The primary outcome variable was 28-day survival and secondary outcome variables were sepsis upon admission to the ICU, and septic shock upon admission to the ICU. Sepsis was defined as the presence of two or more SIRS criteria plus the presence of a known or suspected infection during the 24-hour period. Septic shock was defined by sepsis plus significant hypotension (systolic blood pressure < 90 mm Hg or the need for vasopressors). Baseline demographics recorded were age, gender, medical or surgical diagnosis on admission to the ICU (based on the Acute Physiology And Chronic 38 Health Evaluation (APACHE) III diagnostic codes) , and the admission A P A C H E II score 38 Each of the four SIRS criteria was recorded as present or absent upon admission to the ICU. Microbiological cultures were taken as part of routine medical care for any patients who were suspected of having an infection. As these are cohorts of critically ill patients with SIRS, most patients had cultures taken. Cultures that were judged to be positive due to contamination or colonization by the clinician were excluded. Positive cultures were categorized as Gram-positive, Gram-negative, mixed (polymicrobial), fungal, or other. Sources of positive cultures were categorized as respiratory (sputum), gastrointestinal, genitourinary (urine), endovascular (blood, lines, valves), or from skin, soft tissues or wounds. Selection and Genotyping of SNPs CD14: We genotyped the previously described C-159T CD14 promoter polymorphism in our derivation and validation patient cohorts 1 3 . CD14 C-159T was genotyped by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) in the derivation cohort. The promoter region of the CD14 gene was amplified by PCR using a set of primers designed according to the published sequence (Table 2-2). PCR was performed and analyzed using the restriction fragment length polymorphism (RFLP) technique. PCR was performed in a volume of 10 uL using 200ng of genomic DNA. The DNA was denatured at 94°C for 2 minutes, and PCR cycling was set at 94°C for 30 seconds, 59°C for 30 seconds, and 72°C for 40 seconds for 35 cycles with a final extension step at 72°C for 5 minutes. The resulting PCR product was 497bp and was analyzed using restriction fragment length polymorphism (RFLP) with the restriction endonuclease Ava II 39 (recognition site 5'GGWCC3', W=A/T) according to the manufacturers instructions (New England Biolabs, Beverly, MA). A C at base -159 alters the recognition site to 5'GGSCC3' and the DNA PCR product will not be cut by Ava II. Thus, a CD 14 promoter region sequence with -159C will result in a 497bp fragment, while a sequence with -159T will result in 353bp and 144bp fragments. PCR-RFLP fragments were electrophoresed on a 2% agarose gel at 120V for 90 minutes and visualized with ethidium bromide staining and ultraviolet illumination. CD14 C-159T was genotyped in the validation cohort by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Inc.) (Table 2-2).. MBL: We genotyped our patient cohort for the previously described X / Y polymorphism at position -221 in the promoter 2 2 and structural mutations in exon 1 at codons 52, 54, and 57 (alleles D, B, and C respectively) of M B L 1 9 ~ 2 1 . The A allele indicates the wild type structural allele. Haplotypes of these 4 polymorphisms were inferred for all 39 patients using PHASE software ' . DNA was amplified by site-directed mutagenesis polymerase chain reaction (SDM-PCR) for the exon 1 alleles B, C, and D, and sequence-specific primer PCR (SSP-PCR) for Y/X promoter alleles 2 2 , 2 3 (Table 2-2). All general and site-directed mutagenesis PCR procedures were initiated by a 2 minute denaturation step at 94°C and completed by a 2 minute extension step at 72 °C. The SSP-PCR procedures were also initiated by a 2 minute denaturation step at 94 °C, but were completed by a 5 minute extension step at 72 °C. The temperature cycles for the different types of PCR were as follows: for general and SDM-PCR, 35 cycles of 30 seconds at 94 °C, 30 seconds at 60 °C, 40 and 10 seconds at 72 °C; for SSP-PCR, 35 cycles of 20 seconds at 94 °C, 50 seconds at 65 °C, and 20 seconds at 72 °C. The B, C and D alleles were analyzed using the restriction fragment length polymorphism (RFLP) technique. The M B L B and C alleles were detected by overnight BanI and MboII (New England Biolabs, MA) restriction enzyme digestion of the 329 basepair general PCR product amplified by the M B L exon 1 PCR primers, followed by 2% agarose gel electrophoresis at 120V for 80 minutes. Ban! cleaves the A allele into two fragments, (245bp and 84bp), and leaves the B allele undigested. MboII specifically cleaves the C allele into two fragments (266bp and 63bp), leaving the A allele undigested. The D allele was detected by RFLP performed on the SDM-PCR product. Overnight Mw/(New England Biolabs, MA) digestion of the 121 bp product cleaved the D allele into 2 fragments (21bp and lOObp), and left the A allele undigested. After digestion, the SDM-PCR products were separated by 3% agarose gel electrophoresis at 120V for 90 minutes. Bands were detected by ethidium bromide staining and ultraviolet illumination. TLR2: We selected a SNP in the TLR2 promoter region using a haplotype-based approach to SNP selection. We used unphased genotypic data from 23 Caucasians from the Coriell Cell Repository (from www.innateimmunity.net/IIPGA/IIPGASNPs) 4 0 to infer haplotypes of TLR2 using PHASE software 3 9 . We then used M E G A 2 software 4 1 to infer a phylogenetic tree to identify major haplotype clades. Haplotypes were sorted into clades according to this phylogenetic tree and this haplotype structure was inspected to choose htSNPs 4 2 , 4 3 . We chose an htSNP that identified the 2 major haplotype clades of TLR2 in Caucasians (rs4696480) (Figure 2-1). This SNP was then genotyped in our patient cohort to define major clades of TLR2. The promoter region of the TLR2 gene was amplified by PCR 41 with sequence-specific primers (SSP-PCR) designed according to the published sequence (Table 2-2). PCR was performed separately for each allele in a volume of 10 uL using 200ng of genomic DNA. The DNA was denatured at 94°C for 2 minutes, and PCR cycling was set at 94°C for 30 sees, 65°C for 20 sees, and 72°C for 10 sees for 35 cycles with a final extension step at 72°C for 5 minutes. Genomic DNA was only amplified by a sequence-specific primer if the specific allele was contained in the sequence. The resulting PCR products were 217bp long. PCR products were electrophoresed on a 2% agarose gel. Samples that were homozygous for one allele or the other at position -16933 of the TLR2 gene produced only one band, while samples that were heterozygous for the -16933T/A SNP produced 2 bands on separate gels when detected by ethidium bromide staining and ultraviolet illumination. Blood Collection and Processing Discarded blood from routine clinical laboratory tests was collected for all critically ill patients in the derivation cohort. The buffy coat was extracted from whole blood and samples transferred into 1.5 mL cryotubes and stored at -80°C. DNA was extracted from the buffy coat using the Qiagen DNA Blood Mini Kit. The genotypic analysis was performed in a blinded fashion, without clinical information. Statistical Analysis We used a prospective cohort study design. Rates of dichotomous outcomes (presence or absence of positive bacterial culture, type of bacterial culture (Gram-positive, Gram-negative, other), sepsis and septic shock at admission, and 28-day survival) were compared between genotypes using a chi-squared test. Differences in continuous outcome variables (age, A P A C H E II score) between genotypes were tested using analysis of variance 42 (ANOVA). 28-day survival was further compared between genotypes while adjusting for other confounders (age, sex, and medical vs. surgical diagnosis) using a Cox regression analysis in addition to a Kaplan-Meier analysis. Genotype relative risk for survival was calculated for each genotype. Genotype distributions were tested for Hardy-Weinberg equilibrium to test for population stratification in the cohort 4 4. We report the mean and 95% confidence intervals. Statistical significance was set at p < 0.05. The data was analyzed using SPSS 11.5 for Windows (SPSS Inc, Chicago, IL, 2003). 2.3 Results Each gene was tested separately for association of genotype with clinical phenotype in the derivation cohort. There were no significant differences by genotype (for any of the genes evaluated) in age, gender (percent female), medical vs. surgical diagnosis for admission, or severity of illness at time of admission (as estimated by A P A C H E II score) (Table 2-4). Only CD14 C-159T was tested for association to clinical measures in the validation cohort. There were no differences by CD 14 C-159T genotype in age, gender, medical vs. surgical diagnosis for admission, or severity of illness at time of admission in the validation cohort (Table 2-5). CD14 247 patients in the derivation cohort were successfully genotyped for the C-159T CD14 polymorphism. C-159T CD14 genotypes occurred in the derivation cohort with frequencies similar to those previously reported in Caucasians 1 3 ' 4 5 > 4 6 (Table 2-4) and were in Hardy-Weinberg equilibrium indicating there was no population stratification. 43 In the derivation cohort the CD14 -159TT genotype was associated with significantly increased prevalence of positive bacterial cultures at admission to the ICU (p<0.02) (Figure 2-2). Both the CD14 -159TT and the -159CT genotypes were associated with significantly increased prevalence of Gram-negative cultures compared to the -159CC genotype (p<0.03) (Figure 2-3). The frequencies of all types and sources of positive microbiological cultures by CD14 genotype are displayed in Table 2-5. CD14 genotypes were not, however, associated with a significantly different prevalence of sepsis or septic shock upon admission to the ICU, nor with 28-day survival (Figure 2-4). The CD14 -159TT genotype was associated with 62% 28-day survival rate while the CD14 -159CT and -159CC genotypes were associated with a 68% survival rate, but this difference was not statistically significant (power was only 12%). The association of CD14 C-159T with prevalence of positive bacterial cultures at admission to ICU, and specifically with Gram-negative cultures was further tested in the larger validation cohort of critically ill Caucasians. 694 patients in the validation cohort were successfully genotyped for the C-159T CD14 polymorphism. C-159T CD14 genotypes occurred in the validation cohort with frequencies similar to those previously reported in Caucasians and to those reported in the derivation cohort 1 3 ' 4 5 ' 4 6 (Table 2-5) and are in Hardy-Weinberg equilibrium. In the validation cohort CD14 C-159T genotype was not associated with prevalence of positive cultures at admission to ICU nor with type of culture at admission to ICU. The frequencies of all types and sources of positive microbiological cultures by CD14 genotype are displayed in Table 2-5. However, the CD14 -159TT genotype was significantly associated with decreased 28-day survival in the validation cohort (Figure 2-5). The CD14 -44 159TT genotype was associated with a 58% 28-day survival rate while the CD14 -159CT and CC genotypes were associated with a 67% 28-day survival rate (p<0.03). In the larger cohort we had 55% power to detect a 9% difference in survival. In examining the Kaplan-Meier survival plot it appears that the survival rates of the two groups diverge after day 8 of ICU admission (Figure 2-5). It is important to note, however, that the validation cohort had significantly higher A P A C H E II scores than the derivation cohorts for all CD 14 genotypes (Table 2-7) MBL 222 patients were successfully genotyped for the chosen alleles within the M B L gene. The M B L X/Y promoter polymorphism and the exon 1 A, B, C and D structural alleles were each in Hardy-Weinberg equilibrium. Five common and two rare haplotypes of the X / Y polymorphism and the A, B, C and D structural polymorphisms were inferred in our patient cohort using PHASE software (Table 2-3). O signifies a null haplotype, or any haplotype carrying at least one exon 1 coding polymorphism. The frequencies of these haplotypes were 22 24 similar to frequencies previously reported in Caucasians " . It has been well established that individuals who carry at least one copy of the Y A A A haplotype or who have 2 copies of the X A A A haplotype have higher serum levels of M B L than do individuals who carry the X A A A haplotype in combination with an O haplotype, or 2 oo o i AH copies of an O haplotype (Table 2-1) ' . Therefore, we used this approach to group patients in our cohort into a "high M B L " haplotype group and a "low M B L " haplotype group according to the pair of M B L haplotypes that they carried (Table 2-4). 45 Patients in the low MBL haplotype group had significantly increased prevalence of positive bacterial cultures at admission to the ICU (p<0.02) (Figure 2-6). There was no association of MBL haplotype with a specific class of microorganism (Gram-positive vs. Gram-negative) (data not shown). Patients in the low MBL haplotype group did not have significantly increased rates of sepsis or septic shock at admission to the ICU (data not shown). The frequencies of all types and sources of positive microbiological cultures by MBL haplotype are displayed in Table 2-8. 28-day survival did not differ significantly between the low MBL haplotype and high MBL haplotype groups (Figure 2-7). TLR2 237 patients were successfully genotyped for the TLR2 T-16933A polymorphism. The genotype frequencies of this polymorphism were similar to frequencies reported for other available Caucasian data (from www.innateimmunity.net/IIPGA/IIPGASNPs), and the genotypes were in Hardy-Weinberg equilibrium (Table 2-4). We found that the TLR2 -16933AA genotype was associated with significantly increased prevalence of sepsis upon admission to the ICU (p<0.03) (Figure 2-8) and specifically with increased prevalence of Gram-positive infections (p<0.04) (Figure 2-9). TLR2 -16933AA was not associated with increased prevalence of positive bacterial cultures or septic shock upon admission to the ICU, nor with a significant difference in 28-day survival (Figure 2-10). The frequencies of all types and sources of positive microbiological cultures by TLR2 T-16933A genotype are displayed in Table 2-9. 46 2.4 Discussion We found that CD14 -159TT and low MBL haplotype pairs of MBL were associated with increased rates of positive bacterial cultures on admission to the ICU, and that TLR2 -16933AA was associated with increased prevalence of sepsis at ICU admission in a derivation cohort of 252 critically ill Caucasians with SIRS. Furthermore, CD14 -159CT and -159TT were associated with increased prevalence of Gram-negative infections while TLR2 -16933AA was associated with increased prevalence of Gram-positive infections. MBL haplotype pairs were not associated with prevalence of a specific type of infection. None of these polymorphisms of CD14, MBL or TLR2 were associated with significantly increased prevalence of septic shock at admission or with 28-day survival in the derivation cohort. In a larger validation cohort of 694 critically ill Caucasians we found that the CD14 C-159T SNP was not associated with increased prevalence of positive bacterial cultures at admission or prevalence of Gram-negative infections at admission to ICU. CD14 -159TT was associated, however, with significantly decreased 28-day survival in the validation cohort. We have shown that polymorphisms in three key innate immunity receptor genes are associated with increased prevalence of infection in a derivation cohort of critically adults, marked either by positive microbial cultures or sepsis at admission to the ICU. In contrast, polymorphisms in CD 14, MBL, and TLR2 were not associated with increased septic shock or altered survival in the derivation cohort, which initially suggested that polymorphisms in innate immunity receptors may increase susceptibility to infection, but do not increase risk of shock or death in critically ill patients who have SIRS. However, our subsequent finding that CD14 -159TT was associated with decreased survival in the validation cohort suggests that this polymorphism may increase risk of death in critically ill patients who have SIRS, particularly later in the course (after day 8) of their ICU stay. These results may be 47 confounded by the fact that the validation cohort had significantly higher A P A C H E II scores than the derivation cohort. To our knowledge, this is the first report of association of polymorphisms of CD14, M B L , and TLR2 with prevalence and type of infection and outcome in a common prospective cohort of critically ill patients. Several other groups have tested the association of polymorphisms in the CD14 and M B L genes separately with incidence of and outcome from sepsis in separate case-control studies. Two studies reported that the C-159T polymorphism was not associated with risk of severe sepsis in trauma patients or with increased risk of Gram-negative infection in critically ill patients 4 8 ' 4 9 , while a third reported that the -159T allele was more frequent in patients with septic shock than in healthy controls and that the -159TT genotype was a risk factor for death in septic shock patients 4 5 . Thus, the association of CD14 -159TT with increased risk of death (ie. decreased survival) in the validation cohort is consistent with Gibot and colleagues 4 5 . Three structural mutations in exon 1 of the M B L gene (Gly54Asp, Gly57Glu, and Cys52Arg) 2 0 ' 5 0 and 2 promoter variants (G-550C and G-221C) of the M B L gene are associated with altered levels of serum MBL. M B L polymorphisms occur as 6 different haplotypes, or sets of SNPs that are in linkage disequilibrium with one another and are inherited as a unit. These M B L haplotypes are associated with different serum concentrations of M B L (Table 2-1) 2 2 ' 2 3 ' 4 7 . There have been conflicting reports of the association between M B L haplotypes and outcome from sepsis. One early study found that low M B L haplotypes were associated with increased prevalence of sepsis and septic shock, and increased mortality rates in ICU patients 2 6 . A later study by Gordon et al. found, however, that while low M B L haplotypes were more common in patients with severe sepsis 48 and septic shock compared to healthy controls, there was no difference in haplotype frequency between survivors and non-survivors of sepsis and septic shock 5 1 . These latter findings are in agreement with our own, in that low M B L haplotypes are associated with risk of infection and sepsis, but not with mortality in critically ill patients. There have been no studies of the association of the TLR2 T-16933A polymorphism with infection or sepsis. Other polymorphisms in TLR2 have been associated with significantly increased risk of Gram-positive infections, and hypo-responsiveness to bacterial peptides 5 2 ~ 5 4 , but not with mortality from severe Staphylococcus aureus infection 5 5 . There are no previous studies of the association of TLR2 SNPs and risk of death in the critically ill. Our finding of association of innate immunity receptor genotype and increased prevalence of positive bacterial cultures and sepsis is consistent with recent animal model studies. Anti-CD 14 antibodies administered prior to bacterial challenge increased bacterial load in the lungs, but attenuated risk of septic shock, in rabbits 5 6 ' 5 7 . CD14 also alters lethality of murine models of Gram-negative and LPS-induced shock 5 8 ' 5 9 . Anti-CD14 58 antibodies attenuated acute lung injury in mice after intratracheal administration of LPS , while transgenic mice expressing human CD14 were found to be hypersensitive to LPS-induced shock 5 9 . TLR2-deficient mice are more susceptible to Gram-positive and Mycobacterial infection and have reduced clearance of bacteria 6 0 ' 6 1 . Taken together, these studies suggest that polymorphisms of innate immunity receptors could be associated with impaired clearance of bacteria. Notably, our finding that polymorphisms in CD14 and TLR2 are associated with increased prevalence of Gram-negative and Gram-positive bacteria, respectively, suggests 49 that polymorphisms of innate immunity receptors could be associated with impaired clearance of bacteria for which they are specific. Interestingly, we did not find that M B L haplotype pairs were associated with prevalence of a particular type of infection. While MBL's main role is in the recognition and clearing of Gram-negative bacteria, it also recognizes Gram-positive bacteria, yeasts and viruses and that may explain why M B L haplotype pairs were not associated with susceptibility to Gram-positive vs. Gram-negative bacteria. Our finding that CD14 -159TT was not associated with increased risk of Gram-negative infection but was associated with increased risk of death in the validation cohort is not entirely inconsistent with our results from the derivation cohort. Although not statistically significant, the Kaplan-Meier survival plots of the derivation cohort resemble the Kaplan-Meier plots of the validation cohort (Figures 2-4 and 2-5). In the derivation cohort, as in the validation cohort, the survival curve of patients with the CD 14 -159TT genotype diverges from that of patients with the CD14 -159CC or CT genotypes after about day 7 and is consistently lower over the remainder of the 28-day study period. The initial derivation cohort was underpowered to detect a difference in survival (power was 12%). This highlights the importance of performing association studies in very large cohorts of patients in order to detect small but clinically relevant differences in outcome. It is also important to note that the validation cohort had significantly higher A P A C H E II scores at admission to the study. The cohorts were gathered sequentially, and this difference in baseline illness severity indicates that in the future, a better study design would be to collect all patients and then randomly assign them to a derivation or validation cohort to avoid differences in baseline characteristics that may be the result of changes in practice or hospital policy. However, although there was no association of CD14 C-159T genotype and risk of Gram-negative 50 infection at admission to ICU in the validation cohort, it is interesting to note that the difference in survival rates between CD14 genotypes appears after day 8 of ICU admission, a time when risk of nosocomial infection increases. Numerous studies have shown that between 5 and 35% of ICU patients acquire nosocomial infections from mechanical ventilation, intra-abdominal infections following trauma or surgery, and bacteremia derived fry from intravascular devices . These nosocomial infections contributed to 88,000 deaths and cost $4.5 billion in the U.S. in 1995 6 2 . In two large studies of nosocomial infection rates in ICU patients, Pseudomonas aeruginosa was found to be the most prevalent cause of ventilator-associated pneumonia and lower respiratory tract infections . Although we do not have the data to test our hypothesis, it is tempting to speculate that patients with the CD14 -159TT genotype may be at increased risk for secondary, nosocomial Gram-negative infections and that these nosocomial infections contribute to their decreased survival rates late in the ICU stay (after day 7). There are several strengths of our gene association study that minimized common limitations of genetic association studies. First, our prospective study of a large cohort of critically ill patients (n=252) reduced the chance of Type I error (finding a spurious association) compared to studies with smaller sample sizes. Second, to avoid spurious associations we included only Caucasians in our cohort of critically ill adults, thus limiting the chance of population stratification due to ethnic heterogeneity . Third, we studied three separate but related innate immunity genes in the same patients, which allows direct comparison of the results of each individual gene association with outcome. CD14, MBL and TLR2 are upstream candidate genes in the innate immunity pathway and thus genetic variants of these genes are likely to have important consequences in recognition and clearance of 51 invading pathogens. Finally, we tested CD14 C-159T for association to clinical outcomes in two separate cohorts in an attempt to validate our initial findings. There are limitations of our gene association study that are relevant. We did not measure either ribonucleic acid (RNA) expression or protein levels of CD 14, MBL, or TLR2 so we do not know the functional consequences of the polymorphisms of these genes in the systemic inflammatory response syndrome. Fortunately, several groups have examined the effects of the CD14 C-159T polymorphism and M B L haplotypes on relevant protein levels. CD14 -159T has been associated with increased density of membrane-bound CD14 on monocytes and increased serum levels of soluble CD14 1 2 ' , 3 ' 6 4 . Pairs of low M B L level haplotypes have been associated with decreased serum levels of M B L 2 2 - 2 4 and with increased incidence of infection, as in chronic necrotizing pulmonary aspergillosis 6 5 . We found that polymorphisms of three important innate immunity genes (CD 14, MBL, TLR2) were associated with increased prevalence of infection at admission to the ICU [prevalence of positive bacterial cultures (CD14, MBL)] or sepsis (TLR2), but not with septic shock, organ dysfunction, or survival in a derivation cohort of critically ill adults. Importantly, we found that polymorphisms in CD 14 and TLR2 are associated with increased prevalence of Gram-negative and Gram-positive bacteria respectively. In a larger validation cohort of critically ill patients we found that the promoter polymorphism of CD14 was associated with decreased survival. It is possible that specific polymorphisms of innate immunity receptors alter host recognition and clearance of bacteria leading to increased prevalence of infections and perhaps increased risk of nosocomial infections causing decreased late survival. 52 2.5 Tables and Figures Table 2-1. MBL haplotypes associated with high, moderate, or low levels of serum M B L Serum M B L G-221C Codon 52 Codon 54 Codon 57 Haplotype (Y/X) A/D A/B A/C Y A High Y A A A X A Moderate X A A A O Low Y A A C O Low Y A B A 0 Low Y D A A 0 Low Y D B A 53 Table 2-2. Primer and probe sequences used for genotyping CD14 C-159T, MBL A,B,C,D,X&Y alleles and TLR2 T-16933A in the validation cohort using the ABI Prism 7900HT Sequence Detection System SNP Genotyping . Method Primer Sequence CD14 C-159T PCR-RFLP Forward Reverse G T G C C A A C A G A T G A G G T T C A C G C C T C T G A C A G T T T A T G T A A T C CD14 C-159T RT-PCR (TaqMan, ABI) Primer L Primer R Probe VIC Probe F A M C T A G A T G C C C T G C A G A A T C C T T C C C T T C C T T T C C T G G A A A T A T T G C A C C T G T T A C G G T C C C C C T C T G T T A C G G C C C C C C T M B L exon 1 General PCR Forward Reverse G T A G G A C A G A G G G C A T G C T C C A G G C A G T T T C C T C T G G A A G G M B L allele D SDM-PCR Forward Reverse C A A C G G C T T C C C A G G C A A A G A C G C G A T C C C C A G G C A G T T T C C T C T G G A A G G M B L allele X SSP-PCR Forward Reverse C A T T T G T T C T C A C T G C C A C C A C A T T C C T T G T G A C A C T G C G M B L allele Y SSP-PCR Forward Reverse C A T T T G T T C T C A C T G C C A C G A C A T T C C T T G T G A C A C T G C G TLR2 T-16933A SSP-PCR Forward T Forward A Reverse A T T G A A G G G C T G C A T C T G G T A T T G A A G G G C T G C A T C T G G A G C T G A G A G G T G G A A C C T T T T * VIC and F A M are fluorophores used to label the allele-specific hybridization probes. 54 Table 2-3. MBL Haplotype Frequencies in the Derivation Cohort of Critically 111 Adults with SIRS Haplotype G-221C (Y/X) Codon 52 A/D Codon 54 A/B Codon 57 A/C Frequency YA Y A A A 53.4% XA X A A A 19.4% 0 X A A C 0.2% 0 Y A A C 2.1% O Y A B A 15.7% O Y D A A 8.9% O Y D B A 0.2% 55 Table 2-4. CD14, MBL, and TLR2 genotype/haplotype frequencies and derivation cohort patient baseline characteristics by genotype/haplotype Baseline Characteristics Frequency Age (years) (meaniSD) % Female % Medical Diagnosis APACHE II (mean±SD) CD14 n=247 CC 33% 58±17 31 78 21±9 CT 44% 61±15 39 66 21±8 TT 23% 58±18 31 69 19±8 P NS NS NS NS MBL n=222 High 80% 60±16 37 67 21±8 Low 20% 56±18 32 73 20±9 P NS NS NS NS TLR2 n=237 A A 25% 60±15 37 58 23±9 T A 50% 59±17 33 76 21±8 TT 25% 58±17 38 74 20±10 P NS NS ' NS NS * Individuals who carry at least one copy of the M B L Y A A A haplotype or have two copies of the M B L X A A A haplotype have "high" serum concentrations of M B L compared to individuals who carry the X A A A haplotype in combination with an O haplotype, or two copies of an O haplotype 2 2 ' 2 3 ' 4 1 . 56 Table 2-5. CD14 genotype frequencies and validation cohort patient baseline characteristics by genotype Baseline Characteristics Frequency Age (years) % Female % Medical A P A C H E II (mean±SD) Diagnosis (mean±SD) CD14 n=694 CC 26% 58±16 41 74 24±9 CT 48% 56±26 37 78 23±9 TT 26% 57±17 36 80 24±9 p NS NS NS NS 57 Table 2-6. Frequency of types and sources of positive microbiological cultures by CD14 C-159T genotype in the derivation and validation cohorts of critically ill patients Derivation Cohort CD14-159CC CD14-159CT CD14-159TT Type of Culture Gram-positive 14.8% 18.5% 29.3% Gram-negative 13.6% 28.7% 29.3% Mixed 6.2% 1.9% 3.4% Fungal 1.2% 0.9% 0% Other 2.5% 1.9% 3.4% Source of Culture Respiratory 13.6% 25.9% 32.8% Gastrointestinal 6.2% 6.5% 5.2% Skin, soft tissue 6.2% 0.9% 5.2% Genitourinary 0% 2.8% 8.6% Endovascular 13.6% 15.7% 13.8% Validation Cohort CD14-159CC CD14-159CT CD14-159TT Type of Culture Gram-positive 23.6% 23.7% 27.0% Gram-negative 9.0% 9.8% 11.8% Mixed 6.7% 5.9% 2.2% Fungal 1.7% 0.9% 1.7% Other 1.7% 0.3% 0% Source of Culture Respiratory 19.7% 14.5% 17.4% Gastrointestinal 2.8% 0.6% 0.6% Skin, soft tissue 3.4% 1.2% 3.9% Genitourinary 1.7% 1.5% 1.7% Endovascular 14.0% 21.9% 19.1% 58 Table 2-7. Comparison of derivation and validation cohort baseline characteristics by CD14 C-159T genotype CD14 C-159T Genotype Age Derivation Validation P % Female Derivation Validation P APACHEII Derivation Validation P % Medical Derivation Validation P CC 58±17 58±16 0.8 31% 41% 0.2 21±9 24±9 0.04 78% 74% 0.5 CT 61±15 56±26 0.3 39% 37% 0.8 21±8 23±9 0.05 66% .78% 0.02 TT 58±18 57±17 0.6 31% 36% 0.5 19±8 24±9 4x10"4 69% 80% 0.1 Mean 59±16 57±22 0.1 34% 38% 0.4 21±8 24±9 2x10"3 70% 77% 0.05 59 Table 2-8. Frequency of types and sources of positive microbiological cultures by MBL Haplotype pair in the derivation cohort of critically ill patients Low M B L Haplotype Pairs High M B L Haplotype Pairs Type of Culture Gram-positive 25.0% 16.9% Gram-negative 31.8% 24.2% Mixed 4.5% 3.4% Fungal 2.3% 0.6% Other 4.5% 1.7% Source of Culture Respiratory 29.5% 22.5% Gastrointestinal 13.6% 4.5% Skin, soft tissue 4.5% 3.9% Genitourinary 2.3% 3.4% Endovascular 18.2% 12.9% * Individuals who carry at least one copy of the M B L Y A A A haplotype or have two copies of the M B L X A A A haplotype have "high" serum concentrations of M B L compared to individuals who carry the X A A A haplotype in combination with an O haplotype, or two copies of an O haplotype 2 2 ' 2 3 ' 4 1 . 60 Table 2-9. Frequency of types and sources of positive microbiological cultures by TLR2 T-16933A genotype in the derivation cohort of critically ill patients TLR2 -16933AA TLR2 -16933TA TLR2 -16933TT Type of Culture Gram-positive 30.0% 16.8% 15.5% Gram-negative 25.0% 24.4% 19.0% Mixed 0% 2.5% 10.3% Fungal 0% 1.7% 0% Other 1.7% 2.5% 3.4% Source of Culture Respiratory 25.0% 20.2% 25.9% Gastrointestinal 1.7% 7.6% 8.6% Skin, soft tissue 8.3% 0.8% 5.2% Genitourinary 5.0% 4.2% 0% Endovascular 18.3% 15.1% 8.6% 61 -16933 -16693 -16692 -15731 -15607 596 638 1349 1622 1891 2257 Clade Frequency A 1 0.48 A A A A T 2 0.52 T T T J htSNP Figure 2-1. Haplotype structure of the Tol l- l ike receptor 2 (TLR2) gene. Haplotypes of the TLR2 gene were inferred from unphased genotype data from 23 Caucasians using PHASE software. Columns are polymorphic sites of TLR2. Rows are haplotypes of TLR2 ordered by phylogenetic relationship. Yellow boxes are minor alleles and blue boxes are major alleles. Cladistic relationships of TLR2 haplotypes were determined using MEGA2 phylogenetic software. There are 2 major haplotype clades of TLR2, marked clades 1 and 2. T-16933A was chosen as a haplotype clade tag single nucleotide polymorphism (htSNP) as it distinguished between the 2 major haplotype clades. 62 Figure 2-2. Prevalence of positive bacterial cultures at ICU admission by Cluster of differentiation 14 (CD14) C-159T genotype. Bacterial cultures were taken as part of routine medical care for any patients who were suspected of having an infection. Buffy coat was extracted from discarded whole blood. DMA was isolated by Qiagen DNA Blood Mini Kit, the CD14 promoter amplified by polymerase chain reaction (PCR), and subjected to restriction fragment length polymorphism (RFLP). Association of CD14. C -159T genotype with prevalence of positive bacterial cultures was tested by chi-square analysis and relative risk was calculated. 63 100 -, CO Ui o c E CO O *p<0.03 RR=2.13 CC CT CD14 C-159T Genotype Figure 2-3. Prevalence of Gram-negative bacterial cultures at ICU admission by CD14 C-159T genotype. Cultures were categorized as Gram-positive, Gram-negative, fungal, or other. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit, the CD14 promoter amplified by polymerase chain reaction (PCR), and subjected to restriction fragment length polymorphism (RFLP). Association of CD14 C -159T genotype with type of bacterial culture was tested by chi-square analysis and relative risk was calculated. 64 Figure 2-4. Kaplan-Meier survival analysis by CD14 genotype in the derivation cohort. Patients' survival was scored for 28 days or until hospital discharge. Kaplan-Meier analysis showed that patients homozygous for CD 14 -159T did not have different survival rates over the 28-day observation period (p-0A2) compared to patients homozygous or heterozygous for CD14-159C. 65 Figure 2-5. Kaplan-Meier survival analysis by CD14 genotype in the validation cohort. Patients' survival was scored for 28 days or until hospital discharge. Kaplan-Meier analysis showed that patients homozygous for CD 14 -159T (TT genotype) had decreased survival rates over the 28-day observation period (p=0.03) compared to patients homozygous or heterozygous for CD14 -159C (CC + CT genotypes). 66 Figure 2-6. Prevalence of positive bacterial cultures at ICU admission by Mannose-binding lectin (MBL) haplotype pair. Bacterial cultures were taken as part of routine medical care for any patients who were suspected of having an infection. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. DNA was amplified by site-directed mutagenesis polymerase chain reaction (SDM-PCR) for the exon 1 alleles B, C, and D, and subjected to restriction fragment length polymorphism (RFLP). The Y/X promoter alleles were amplified by sequence-specific primer PCR (SSP-PCR). Individuals who carry at least one copy of the M B L Y A haplotype or have two copies of the M B L X A haplotype have "high" serum concentrations of M B L compared to individuals who carry the X A haplotype in combination with an O haplotype, or two copies of an O haplotype ' ' . Association of low or high M B L haplotype pairs with prevalence of positive bacterial cultures was tested by chi-square analysis and relative risk was calculated. 67 28-Day Survival by MBL Haplotype Pair 100 o n o l ) £ 60 « • • • • e P a i r s a: a • • •' 40 Haplotyp a: igh M B L airs ow M B L airs r-• . • • • • ' , • D 20 n. • •. Figure 2-7. Kaplan-Meier survival analysis by MBL haplotype pair. Patients' survival was scored for 28 days or until hospital discharge. Kaplan-Meier analysis showed that patients with low M B L level haplotype pairs did not have different survival rates over the 28 day observation period (p=0.34) compared to patients with high M B L level haplotype pairs. 68 T A + T T n = 1 7 7 A A n=60 TLR2 T-16933A Genotype Figure 2-8. Prevalence of sepsis at ICU admission by TLR2 T-16933A genotype. A patient was considered septic at admission to ICU if two or more SIRS criteria were present plus a known or suspected infection. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. The TLR2 promoter was amplified by PCR with sequence-specific primers (SSP-PCR) so that genomic DNA was only amplified by a sequence-specific primer if the specific allele was contained in the sequence. Association of TLR2 T-16933A genotype with prevalence of sepsis at admission to ICU was tested by chi-square analysis and relative risk was calculated. 69 100 -, 80 ~ 60 40 *p<0.04 RR=1.83 TA + TT AA TLR2 Genotype Figure 2-9. Prevalence of Gram-positive bacterial cultures at ICU admission by TLR2 T-16933A genotype. Cultures were categorized as Gram-positive, Gram-negative, fungal, or other. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. The TLR2 promoter was amplified by PCR with sequence-specific primers (SSP-PCR) so that genomic DNA was only amplified by a sequence-specific primer if the specific allele was contained in the sequence. Association of TLR2 T-16933A genotype with type of bacterial culture was tested by chi-square analysis and relative risk was calculated. 70 Figure 2-10. Kaplan-Meier survival analysis by TLR2 T-16933A genotype. Patients' survival was scored for 28 days or until hospital discharge. Kaplan-Meier analysis showed that patients homozygous for TLR2 -16933A and patients homozygous or heterozygous for TLR2 -16933T did not have different survival rates over the 28 day observation period (p=0.77). 71 2.6 References 1. Burgner D, Levin M. Genetic susceptibility to infectious diseases. Pediatr Infect DisJ2003;22:1-6. 2. Bellamy R, Hill AV. Genetic susceptibility to mycobacteria and other infectious pathogens in humans. Curr Opin Immunol 1998;10:483-7. 3. Choi EH, Zimmerman PA, Foster CB et al. Genetic polymorphisms in molecules of innate immunity and susceptibility to infection with Wuchereria bancrofti in South India. Genes Immun 2001;2:248-53. 4. Majetschak M, Obertacke U, Schade FU et al. Tumor necrosis factor gene polymorphisms, leukocyte function, and sepsis susceptibility in blunt trauma patients. Clin Diagn Lab Immunol 2002;9:1205-11. 5. Mira JP, Cariou A, Grail F et al. Association of TNF2, a TNF-alpha promoter polymorphism, with septic shock susceptibility and mortality: a multicenter study. Jama 1999;282:561-8. 6. Read RC, Pullin J, Gregory S et al. A functional polymorphism of toll-like receptor 4 is not associated with likelihood or severity of meningococcal disease. J Infect Dis 2001;184:640-2. 7. Chow JC, Young DW, Golenbock DT, Christ WJ, Gusovsky F. Toll-like receptor-4 mediates lipopolysaccharide-induced signal transduction. J Biol Chem 1999;274:10689-92. 8. Wright SD, Ramos RA, Tobias PS, Ulevitch RJ, Mathison JC. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 1990;249:1431-3. 9. Labeta MO, Durieux JJ, Fernandez N, Herrmann R, Ferrara P. Release from a human monocyte-like cell line of two different soluble forms of the lipopolysaccharide receptor, CD14. Eur J Immunol 1993;23:2144-51. 10. Landmann R, Zimmerli W, Sansano S et al. Increased circulating soluble CD 14 is associated with high mortality in gram-negative septic shock. J Infect Dis 1995;171:639-44. 11. Burgmann H, Winkler S, Locker GJ et al. Increased serum concentration of soluble CD 14 is a prognostic marker in gram-positive sepsis. Clin Immunol Immunopathol 1996;80:307-10. 12. Baldini M , Lohman IC, Halonen M , Erickson RP, Holt PG, Martinez FD. A Polymorphism* in the 5' flanking region of the CD14 gene is associated with circulating soluble CD 14 levels and with total serum immunoglobulin E. Am J Respir Cell Mol Biol 1999;20:976-83. 13. Hubacek JA, Rothe G, Pit'ha J et al. C(-260)->T polymorphism in the promoter of the CD14 monocyte receptor gene as a risk factor for myocardial infarction. Circulation 1999;99:3218-20. 14. Weis WI, Drickamer K. Trimeric structure of a C-type mannose-binding protein. Structure 1994;2:1227-40. 15. Kuhlman M, Joiner K, Ezekowitz RA. The human mannose-binding protein functions as an opsonin. J Exp Med 1989;169:1733-45. 16. Schelenz S, Malhotra R, Sim RB, Holmskov U, Bancroft GJ. Binding of host collectins to the pathogenic yeast Cryptococcus neoformans: human surfactant protein D acts as an agglutinin for acapsular yeast cells. Infect Immun 1995;63:3360-6. 72 17. Tabona P, Mellor A, Summerfield JA. Manriose binding protein is involved in first-line host defence: evidence from transgenic mice. Immunology 1995;85:153-9. 18. van Emmerik LC, Kuijper EJ, Fijen CA, Dankert J, Thiel S. Binding of mannan-binding protein to various bacterial pathogens of meningitis. Clin Exp Immunol 1994;97:411-6. 19. Lipscombe RJ, Sumiya M , Hill A V et al. High frequencies in African and non-African populations of independent mutations in the mannose binding protein gene. Hum Mol Genet 1992;1:709-15. 20. Madsen HO, Garred P, Kurtzhals JA et al. A new frequent allele is the missing link in the structural polymorphism of the human mannan-binding protein. Immunogenetics 1994;40:37-44. 21. Sumiya M , Super M , Tabona P et al. Molecular basis of opsonic defect in immunodeficient children. Lancet 1991;337:1569-70. 22. Madsen HO, Garred P, Thiel S et al. Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein. J Immunol 1995;155:3013-20. 23. Steffensen R, Thiel S, Vanning K, Jersild C, Jensenius JC. Detection of structural gene mutations and promoter polymorphisms in the mannan-binding lectin (MBL) gene by polymerase chain reaction with sequence-specific primers. J Immunol Methods 2000;241:33-42. 24. Crosdale DJ, Oilier WE, Thomson W et al. Mannose binding lectin (MBL) genotype distributions with relation to serum levels in UK Caucasoids. Eur J Immunogenet 2000;27:111-7. 25. Fidler KJ, Wilson P, Davies JC, Turner MW, Peters MJ, Klein NJ. Increased incidence and severity of the systemic inflammatory response syndrome in patients deficient in mannose-binding lectin. Intensive Care Med 2004;30:1438-45. 26. Garred P, J JS, Quist L, Taaning E, Madsen HO. Association of mannose-binding lectin polymorphisms with sepsis and fatal outcome, in patients with systemic inflammatory response syndrome. J Infect Dis 2003;188:1394-403. 27. Aliprantis AO, Yang RB, Mark MR et al. Cell activation and apoptosis by bacterial lipoproteins through toll-like receptor-2. Science 1999;285:736-9. 28. Brightbill HD, Libraty DH, Krutzik SR et al. Host defense mechanisms triggered by microbial lipoproteins through toll-like receptors. Science 1999;285:732-6. 29. Lien E, Sellati TJ, Yoshimura A et al. Toll-like receptor 2 functions as a pattern recognition receptor for diverse bacterial products. J Biol Chem 1999;274:33419-25. 30. Takeda K, Akira S. Toll receptors and pathogen resistance. Cell Microbiol 2003;5:143-53. 31. Travassos LH, Girardin SE, Philpott DJ et al. Toll-like receptor 2-dependent bacterial sensing does not occur via peptidoglycan recognition. EMBO Rep 2004;5:1000-6. 32. Lotz S, Aga E, Wilde I et al. Highly purified lipoteichoic acid activates neutrophil granulocytes and delays their spontaneous apoptosis via CD14 and TLR2. / Leukoc Biol 2004;75:467-77. 33. Ellingsen E, Morath S, Flo T et al. Induction of cytokine production in human T cells and monocytes by highly purified lipoteichoic acid: involvement of Toll-like receptors and CD14. Med Sci MomY 2002;8:BR149-56. 34. Morath S, Stadelmaier A, Geyer A, Schmidt RR, Hartung T. Synthetic lipoteichoic acid from Staphylococcus aureus is a potent stimulus of cytokine release. J Exp Med 2002;195:1635-40. 73 35. Bone RC. The sepsis syndrome. Definition and general approach to management. Clin Chest Med 1996;17:175-81. 36. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet 2001;2:91-9. 37. Knaus WA, Wagner DP, Draper EA et al. The A P A C H E III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100:1619-36. 38. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. A P A C H E II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 39. Stephens M , Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001 ;68:978-89. 40. Rats RU, et al. TLR2. 2003 vol: Innate Immunity PGA, NHLBI Program in Genomic Applications; 2002. 41. Kumar S, Tamura K, Jakobsen IB, Nei M . MEGA2: molecular evolutionary genetics analysis software. Bio informatics 2001;17:1244-5. 42. Johnson GC, Esposito L, Barratt BJ et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001 ;29:233-7. 43. Gabriel SB, Schaffner SF, Nguyen H et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225-9. 44. Guo SW, Thompson EA. Performing the exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics 1992;48:361-72. 45. Gibot S, Cariou A, Drouet L, Rossignol M, Ripoll L. Association between a genomic polymorphism within the CD14 locus and septic shock susceptibility and mortality rate. Crit Care Med2002;30:969-73. 46. Lichy C, Meiser H, Grond-Ginsbach C, Buggle F, Dorfer C, Grau A. Lipopolysaccharide receptor CD14 polymorphism and risk of stroke in a South-German population. J Neurol 2002;249:821-3. 47. Crosdale DJ, Poulton K V , Oilier WE, Thomson W, Denning DW. Mannose-binding lectin gene polymorphisms as a susceptibility factor for chronic necrotizing pulmonary aspergillosis. J Infect Dis 2001 ;184:653-6. 48. Heesen M , Bloemeke B, Schade U, Obertacke U, Majetschak M. The -260 C ->T promoter polymorphism of the lipopolysaccharide receptor CD 14 and severe sepsis in trauma patients. Intensive Care Med 2002;28:1161 -3. 49. Agnese D M , Calvano JE, Hahm SJ et al. Human toll-like receptor 4 mutations but not CD14 polymorphisms are associated with an increased risk of gram-negative infections. J Infect Dis 2002;186:1522-5. 50. Wallis R, Cheng JY. Molecular defects in variant forms of mannose-binding protein associated with immunodeficiency. J Immunol 1999;163:4953-9. 51. Gordon A C , Waheed U, Hansen TK et al. Mannose-binding lectin polymorphisms in severe sepsis: relationship to levels, incidence, and outcome. Shock 2006;25:88-93. 52. Bochud PY, Hawn TR, Aderem A. Cutting edge: a Toll-like receptor 2 polymorphism that is associated with lepromatous leprosy is unable to mediate mycobacterial signaling. J Immunol 2003; 170:3451 -4. 53. Kang TJ, Lee SB, Chae GT. A polymorphism in the toll-like receptor 2 is associated with IL-12 production from monocyte in lepromatous leprosy. Cytokine 2002;20:56-62. 74 54. Lorenz E, Mira JP, Cornish KL, Arbour NC, Schwartz DA. A novel polymorphism in the toll-like receptor 2 gene and its potential association with staphylococcal infection. Infect Immun 2000;68:6398-401. 55. Moore CE, Segal S, Berendt AR, Hill A V , Day NP. Lack of association between Toll-like receptor 2 polymorphisms and susceptibility to severe disease caused by Staphylococcus aureus. Clin Diagn Lab Immunol 2004; 11:1194-7. 56. Frevert CW, Matute-Bello G, Skerrett SJ et al. Effect of CD14 blockade in rabbits with Escherichia coli pneumonia and sepsis. J Immunol 2000;164:5439-45. 57. Opal SM, Palardy JE, Parejo N, Jasman RL. Effect of anti-CD14 monoclonal antibody on clearance of Escherichia coli bacteremia and endotoxemia. Crit Care Med 2003;31:929-32. 58. Tasaka S, Ishizaka A, Yamada W et al. Effect of CD 14 blockade on endotoxin-induced acute lung injury in mice. Am J Respir Cell Mol Biol 2003;29:252-8. 59. Ferrero E, Jiao D, Tsuberi BZ et al. Transgenic mice expressing human CD14 are hypersensitive to lipopolysaccharide. Proc Natl Acad Sci USA 1993;90:2380-4. 60. Echchannaoui H, Frei K, Schnell C, Leib SL, Zimmerli W, Landmann R. Toll-like receptor 2-deficient mice are highly susceptible to Streptococcus pneumoniae meningitis because of reduced bacterial clearing and enhanced inflammation. J Infect Dis 2002;186:798-806. 61. Drennan MB, Nicolle D, Quesniaux VJ et al. Toll-like receptor 2-deficient mice succumb to Mycobacterium tuberculosis infection. Am J Pathol 2004;164:49-57. 62. Eggimann P, Pittet D. Infection control in the ICU. Chest 2001 ;120:2059-93. 63. Gaynes R, Culver DH, Banerjee S, Edwards JR, Henderson TS. Meaningful interhospital comparisons of infection rates in intensive care units. Am J Infect Control 1993;21:43-4. 64. Karhukorpi J, Yan Y, Niemela S et al. Effect of CD14 promoter polymorphism and H. pylori infection and its clinical outcomes on circulating CD 14. Clin Exp Immunol 2002;128:326-32. 65. Yoon HJ, Shin JH, Yang SH et al. Association of the CD14 gene -159C polymorphism with progression of IgA nephropathy. J Med Genet 2003;40:104-8. 75 CHAPTER 3: A HAPLOTYPE CLADE OF INTERLEUKIN-1 RECEPTOR ASSOCIATED KINASE 4 (IRAK4) CONTAINING A NON-SYNONYMOUS POLYMORPHISM (ALA428THR) IS ASSOCIATED WITH ALTERED INNATE IMMUNE INFLAMMATORY RESPONSES IN VITRO AND BACTERIAL INFECTION IN CRITICALLY ILL PATIENTS 3.1 Introduction The innate immune system is our first line of defense against invading microorganisms. Pattern recognition receptors of the innate immune system, most importantly Toll-like receptors (TLRs), recognize conserved domains of pathogens, or pathogen associated molecular patterns (PAMPs) ' ' 2 . Upon interaction with a PAMP, TLRs interact with a network of intracellular signalling molecules through their Toll-like/Interleukin-1 receptor (TIR) domain leading to the activation of the inflammatory response. The inflammatory response is critical for isolation and clearance of pathogens; however, excessive inflammation may result in the clinical phenotype of sepsis syndrome and multi-system organ failure. There is considerable inter-individual variability in the degree of activation of the innate immunity and inflammatory responses to infection 3"5. Genotype has been shown to contribute significantly to outcome in diseases involving innate immunity 6 - 8 . For example, the genetic contribution to the outcome from infectious diseases exceeded the genetic contribution to cancer risk by 5 fold 9 Genetic variation within innate immunity genes may play a role in an individual's susceptibility to and outcome from severe infection. A version of this chapter has been published. Sutherland A M , Walley KR, ManochaS, Russell JA. 76 The association of interleukin 6 haplotype clades with mortality in critically ill adults. Arch Intern Med. 2005 Jan 10;165(l):75-82. signalling and is essential for all TLR signalling with the exception of TLR3 1 0 . IRAK4 is serine-threonine kinase that interacts with MyD88 and then phosphorylates downstream molecules, ultimately leading to the activation and nuclear translocation of nuclear-factor-KB and the subsequent transcription of inflammatory mediators such as interluekin-6 (IL-6) 1 0 . Paediatric patients with rare mutations of IRAK4 experience recurrent Gram-positive infections and leukocytes from these children are hypo-responsive to TLR ligands in vitro 1 7 . To date there have been no studies of the role of common genetic variants of IRAK4 in susceptibility to infection. We hypothesized that common genetic variants of IRAK4 may be associated with susceptibility to and outcome from infection in critically ill adults. In order to test our hypotheses we recruited a prospective cohort of critically ill patients with systemic inflammatory response syndrome (SIRS). We inferred haplotypes of IRAK4 and grouped them into clades, or evolutionarily related sets of haplotypes 1 8 . We then selected haplotype tag SNPs (htSNPs) that uniquely defined all haplotype clades of IRAK4 to be genotyped and associated to phenotype in the cohort of critically ill patients19. We found that a haplotype clade tagged by the A allele of the htSNP G29429A (Ala428Thr) was associated with increased prevalence of Gram-positive infection at admission to ICU. In order to elucidate the mechanism behind this association we tested the response of B-lymphocyte and fibroblast cell lines of different IRAK4 G29429A genotypes to TLR ligands. The lymphocyte cell lines were readily available from the Coriell Cell Repository, and were the same cell lines that had been previously genotyped by the SeattleSNPs Program for Genomics Applications, so we knew their IRAK4 genotypes. The fibroblasts had been isolated and cultured from an IRAK4-deficient paediatric patient and were a gift from one of our collaborators at the BC Children and Women's Hospital. We found first that the 29429A allele was associated with decreased B-lymphocyte response to CpG (as measured by IL-6 77 secretion), a TLR9 ligand, and secondly that IRAK4-deficient fibroblasts transfected with an IRAK4 expression plasmid containing the 29429A allele produced less IL-6 in response to lipopolysaccharide (LPS) and peptidoglycan (PGN). Taken together, these data suggest that the IRAK4 haplotype clade marked by 29429A (428Thr) alters susceptibility to Gram-positive bacteria, by decreasing cellular response to TLR ligands. 3.2 Methods This study was approved by the Providence Health Care / University of British Columbia Research Ethics Board. Patient Cohort 1029 consecutive patients admitted to the tertiary mixed medical-surgical Intensive Care Unit (ICU) of St. Paul's Hospital were screened for inclusion into this study. We only report the results for the Caucasian patients (n=820) in order to decrease the potential confounding influence of population admixture secondary to ethnic diversity on associations 9 0 between genotype and phenotype . Patients were screened daily and patients were included in the study cohort (n=775) if they met at least two of four SIRS criteria and were successfully genotyped at all three htSNPs of IRAK4. The SIRS criteria were 1) fever (>38 C) or hypothermia (<36 C), 2) tachycardia (>100 beats/min in the absence of beta blockers), 3) tachypnea (>20 breaths/min) or need for mechanical ventilation, and 4) leukocytosis (total leukocyte count >12,000/uL) or leukopenia (total leukocyte count <4,000/uL) 2 1 . Patients were included in this cohort on the calendar day that 2/4 SIRS criteria were met. 78 Clinical Phenotypes We assessed the risk of the presence of a positive bacterial culture at admission to ICU by IRAK4 haplotype clade. Cultures were taken from respiratory (sputum), gastrointestinal, genitourinary (urine), or endovascular (blood, lines, valves) sources, or from skin, soft tissues or wounds. Cultures were categorized as Gram-positive, Gram-negative, mixed, fungal or other. Our primary outcome variable was 28-day survival. Secondary outcome variables were sepsis upon admission to the ICU and septic shock upon admission to the ICU. Sepsis was defined as the presence of two or more SIRS criteria plus the presence of a known or suspected infection during the 24-hour period. Septic shock was defined by sepsis plus significant hypotension (systolic blood pressure <90 mm Hg or the need for vasopressors). Patients were followed for 28 days for survival. Baseline demographics recorded were age, gender, medical or surgical diagnosis on admission to the ICU (based on the A P A C H E III diagnostic codes), and the admission APACHE II score 2 2 . Each of the four SIRS criteria were recorded as present or absent upon admission to the ICU. Bacterial cultures were taken as part of routine medical care for all patients at admission to the ICU. Cultures that were judged to be positive due to contamination or colonization by the attending physician were excluded. Positive bacterial cultures were categorized as Gram-positive, Gram-negative, fungal, or other. The source of a positive culture was categorized as respiratory (sputum), gastrointestinal (peritoneal fluid, abscess drainage, biliary tract), skin (soft tissues or wounds), genitourinary (urine), endovascular (blood) or other. 79 Blood Collection and Processing for Genomic DNA Discarded blood from routine clinical laboratory tests was collected. Samples were centrifuged at 400g for 15 minutes at room temperature. The buffy coat was collected and transferred into 2.0 mL cryotubes and stored at -80°C. DNA was extracted from the buffy coat using the Qiagen DNA Blood Mini Kit (Qiagen, Mississauga, ON). Haplotypes and Selection of htSNPs We used unphased genotypic data from 23 Caucasians from the Coriell Cell Repository (from pga.mbt.washington.edu) to infer haplotypes of the 1RAK4 gene using PHASE software (Figure 1) 1 8 , 2 3 . We then used M E G A 2 software to infer a phylogenetic tree to identify major haplotype clades 2 4 . Haplotypes were sorted into clades according to this phylogenetic tree and this haplotype structure was inspected to choose a minimum set of "haplotype tag" single nucleotide polymorphisms (htSNPs) (Figure 3-1) 2 5 ' 2 6 . We chose 3 htSNPs that identified 4 major haplotype clades of IRAK4 in Caucasians. The first SNP was an intronic C-to-G substitution at nucleotide 23338 relative to the start of sequencing (NCBI ID: rs4251513). The second SNP was an intronic T-to-C substitution at nucleotide 24472 relative to the start (rs4251520). The third SNP was a G-to-A substitution at nucleotide 29429 (rs4251545) (NCBI IRAK4 accession number AF155U8) that results in an alanine at amino acid position 428 in exon 11 being replaced with a threonine. These three SNPs were then genotyped in our 775 patient cohort and PHASE was used to identify haplotypes of each patient. Genotyping The genotypic analysis was performed in a blinded fashion, without clinical information. Patients' genotypes were determined by real-time polymerase chain reaction 80 (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Inc., Foster City, CA) 2 7 (Table 3-1). Five nanograms of patients' genomic DNA was used per genotyping reaction in a 384-well plate. We genotyped DNA with known genotype from 23 lymphocyte cell lines from the Coriell Cell Repository using ABI Prism 7900HT Sequence Detection System and found complete concordance between our genotyping and the SeattleSNPs genotyping at all three positions of the IRAK4 gene 2 3 . Mechanistic Studies We measured and compared activation of the immune response in cells of different IRAK4 genotype after stimulation with bacterial products to elucidate the mechanism behind the association of the non-synonymous SNP IRAK4 G29429A (Ala428Thr) with Gram-positive infection. Forty-three B-lymphocyte cell lines (n=43 individual normal subjects) from the individuals sequenced by the SeattleSNPs Programs for Genomic Applications were purchased from the Coriell Cell Repository (See Appendix 3-1). Because B-lymphocytes most highly express TLR9 in humans we decided to stimulate them with CpG, which is recognized by TLR9 which is present in B-lymphocytes 2 8 , 2 9 . CpG oligonucleotide Type B is a specific sequence with high GC content found uniquely in bacterial DNA. CpG signals through intracellular TLR9 and activates the MyD88-IRAKl-IRAK4-NFKB signalling pathway. To ensure that any B-lymphocyte immune response was specific to CpG, and not simply a result of internalizing foreign DNA, NCpG oligonucleotide was used as a negative control. NCpG oligonucleotide control is the same size as CpG oligonucleotide Type B and contains the same amount of each nucleotide, but in a scrambled sequence (Table 3-2). 81 Initially, we found that the CpG and NCpG sequences caused a non-specific immune response; the B-lymphocytes produced high amounts of IL-6 when stimulated with NCpG oligonucleotide control. We then discovered earlier work that suggested that the phosphorothioate-modification of CpG and NCpG oligonucleotides (to increase their stability) may cause a non-specific immune response. To avoid a non-specific immune response we obtained phosphodiester-bonded CpG and NCpG sequences to which the B -lymphocytes had no non-specific immune response. The B-lymphocytes were cultured in suspension in RPMI with 10% fetal bovine serum ( F B S ) and 50ug/mL gentamicin. For CpG stimulation, cells were washed and gentamicin-free media was added. One-millilitre aliquots of the B-lymphocytes containing half a million cells were placed in a 12-well plate. The B-lymphocytes were then stimulated in triplicate with 1 OuM or 50uM synthetic CpG oligonucleotide Type B or NCpG oligonucleotide control for 24 hours. Following CpG or NCpG stimulation, IL-6 secretion in the B-lymphocyte supernatant was determined as a measure of the immune response using ELISA (R&D Systems, Minneapolis, MN). The B-lymphocyte cell lines from the Coriell Cell Repositories are from different individuals and so differ at many loci throughout the genome. In order to determine that the difference in immune response observed was solely attributable to the IRAK4 G29429A (Ala428Thr) SNP we obtained a wild-type IRAK4 expression plasmid from Amgen Inc. This was used to transfect IRAK4-deficient fibroblasts (IDFs) (a gift from our collaborators, Dr. David Speert and Dr. Stuart Turvey). These IRAK4-deficient fibroblasts do not activate N F - K B or produce cytokines in response to TLR ligands 1 7 . In order to demonstrate that the immune response of IDFs to toll-like receptor ligands could be rescued with the IRAK4 82 plasmid, IDFs that had been transfected with either the wild-type plasmid or the empty vector pGL3 were then stimulated with LPS or PGN. Normal adult fibroblasts (NAFs) (CRL-2522 from A T C C , Manassas, VA) were also stimulated with LPS or PGN. The supernatant concentration of IL-6 after stimulation was then measured by ELISA and compared between the transfected IDFs and NAFs. The IDFs were grown in T75 flasks in D M E M with 10% FBS and 50ug/mL gentamicin. The day before transfection, the IDFs were seeded to 6-well plates in media containing no antibiotics. The day of transfection the cells were washed and 2mL of serum-free media was added to each well. There were approximately 1 million cells per well. Four ug of the appropriate plasmid and lpg of Lipofectamine 2000 (Invitrogen, Burlington, ON) in 500uL of Opti-MEM (Invitrogen, Burlington, ON) were added to each well. Four hours post-transfection the media was removed and fresh media with serum and 50pg/mL of gentamicin was added. The cells were incubated overnight at 37°C. At that time NAFs were seeded to 6-well plates to use as positive controls and incubated at 37°C overnight (18 hours). Following transfection, the cells were washed with sterile PBS and ImL of fresh serum-free media was added. O.lug/mL of LPS (Invivogen) or lpg/mL of PGN (Invivogen) or serum-free media in triplicate was then added to the appropriate wells containing NAFs or transfected IDFs and incubated the plates for 6 hours at 37°C. After 6 hours the supernatant from the NAFs and transfected IDFs was collected and the cell lysates were harvested. IL-6 concentration in the supernatant was measured using ELISA (R&D Systems, Minneapolis, MN). To test whether the Ala428Thr polymorphism was associated with altered immune response the A allele was exchanged for the G allele at nucleotide 29429 in the IRAK4 expression plasmid using the Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA) so that all other SNP sites were the same between the two plasmids. Thus the 83 IDFs transfected with either the wild-type or mutated IRAK4 expression plasmids differed at only nucleotide 249249 of the IRAK4 gene and were otherwise genetically identical. Sequence-specific primers were designed containing the mutation based on the IRAK4 wild-type cDNA sequence according to guidelines provided by Stratagene (Table 3-3). Five ng of the IRAK4 expression plasmid was used as template DNA for the reaction. At higher concentrations of template DNA the site-directed mutagenesis did not work and we were left with the original plasmid containing the G at nucleotide 29429. The Quick-Change site-directed mutagenesis kit makes use of the enzyme Dpn I to digest parental methylated and hemimethylated DNA while leaving the non-methylated mutant strand intact. At high concentrations of template DNA the enzyme may be overwhelmed resulting in an incomplete digestion of the parental strands. Following transformation of the mutated plasmid into XL1 -Blue supercompetent cells for nick repair the cells were plated onto agar plates containing lOOug/mL Ampicillin. Individual colonies were then picked from the plates and grown at 37°C overnight in 4mL of LB broth with lOOug/mL Ampicillin. The mutated IRAK4 plasmids were purified using the QIAprep Miniprep kit (Qiagen Inc., Mississauga, ON) with a microcentrifuge according to the manufacturer's instructions. The DNA was eluted in water for sequencing at the University of British Columbia Nucleic Acid Protein Service Unit using a custom-designed primer (Table 3-3). The sequencing results were compared to the IRAK4 wild-type cDNA sequence using the Basic Local Alignment Search Tool (BLAST) bl2seq to align two sequences. Once the IRAK4 expression plasmid was successfully mutated to carry an A allele at nucleotide 294249 the IDFs were transfected with the IRAK4 29429G plasmid, the IRAK4 29429A plasmid or the pGL3 empty plasmid as a control. The IDFs were grown in T75 flasks in D M E M with 10% FBS and 50pg/mL gentamicin. The day before transfection, the 84 IDFs were seeded to 12-well plates in media containing no antibiotics. The day of transfection the cells were washed and lmL of serum-free media was added to each well. There were approximately half a million cells per well. Two u,g of the appropriate plasmid and 0.5 ug of Lipofectamine 2000 (Invitrogen, Burlington, ON) in 200uL of Opti-MEM (Invitrogen, Burlington, ON) were added to each well. Four hours post-transfection the media was removed and fresh media with serum and 50ug/mL of gentamicin was added. The cells were incubated overnight at 37°C. Following transfection the cells were washed with sterile PBS and lmL of fresh serum-free media was added. O.lug/mL of LPS or lug/mL of PGN or serum-free media in triplicate was then added to the appropriate wells and the plates were incubated for 6 hours at 37°C. After 6 hours the supernatant was collected. IL-6 concentration in the supernatant was measured using ELISA (R&D Systems, Minneapolis, MN). Statistical Analysis We used a cohort study design. Long and Langley 3 0 performed an extensive simulation study to determine the power of association-based studies when a dense, but not exhaustive, set of SNPs is available over a candidate gene region. A hidden polymorphism was assumed to contribute to the variance of a quantitative trait and the number of nearby SNPs was varied in an estimation of the number of subjects necessary to detect the causative locus. The power of SNP and haplotype association studies was also compared to the quantitative trait TDT (TDT-Q5). Under the assumptions of these models, a sample size of 500 or more is required to detect association with 80% power when 10% or less of the variance in the trait is explained by the locus . Our study of 775 Caucasians had >80% power to detect an absolute difference of 10% in prevalence of positive bacterial cultures at 85 admission to ICU, assuming a 40% baseline prevalence of positive bacterial cultures at admission with significance set at p<0.05. Rates of dichotomous outcomes (positive cultures at admission to ICU, sepsis and shock at onset of SIRS, 28-day survival) were compared between clades using Fisher's exact test. Differences in continuous outcome variables (age, A P A C H E II score) between haplotype clades were tested using ANOVA. 28-day survival was further compared between the two groups of patients while adjusting for other confounders (age, sex, and medical vs. surgical diagnosis) using a Cox regression analysis as well as a Kaplan-Meier analysis of censored survival data. Haplotype clade relative risk was calculated. Genotype distributions were tested for Hardy-Weinberg equilibrium 3 1 . Supernatant IL-6 concentrations following CpG stimulation of B-lymphocytes were compared among IRAK4 haplotype clades by ANOVA. Supernatant IL-6 concentrations after stimulation with LPS or PGN were compared among IDFs transfected with the wild-type IRAK4 plasmid or the PGL3 empty plasmid and normal adult fibroblasts using ANOVA. Supernatant IL-6 concentrations after stimulation with LPS or PGN were compared among IDFs transfected with the IRAK4 29429G plasmid, the IRAK4 29429A plasmid and the pGL7 empty plasmid by ANOVA. We report the mean and 95% confidence intervals. Statistical significance was set at p < 0.05. The data was analyzed using SPSS 14.0 for Windows (SPSS lnc, Chicago, IL, 2003). 86 3.3 Results Genotype-Clinical Phenotype Associations We were able to infer haplotypes from complete sequencing of 1RAK4 for 23 Caucasians in the Coriell Cell Repository 2 3 using PHASE software, and identified four major haplotype clades using MEGA2 software (Figure 3.1) 1 8 ' 2 4 . These 4 clades could be resolved by genotyping three htSNPs; C23338G, T24472C, and G29429A 26. 115 Caucasian patients with SIRS were successfully genotyped for all three htSNPs. All three SNPs were in Hardy-Weinberg equilibrium. The haplotype clade defined by 23338C/24472T/29429G (C/T/G) occurred with a frequency of 31%, the C/T/A haplotype clade occurred with a frequency of 9%, the C/C/G haplotype clade occurred with a frequency of 10%, and the G/T/G haplotype clade occurred with a frequency of 50%. These frequencies were similar to frequencies deduced from other available Caucasian data 2 3. For the 775 successfully genotyped individuals of the cohort of Caucasian patients who had at least 2 of 4 SIRS criteria, there were no differences in age, gender (percent female), or severity of illness at the time of admission (as estimated by the A P A C H E II score) by IRAK4 haplotype clade (Table 3-4). The C/T/G haplotype clade was associated with a slightly increased prevalence of medical diagnoses for admission to the ICU. On initial analysis the IRAK4 C/T/A haplotype clade was associated with increased prevalence of positive bacterial cultures at admission to the ICU compared to the other three haplotype clades (Figure 3-2A). For subsequent analysis we compared the C/T/A haplotype clade to all other clades grouped (Figure 3-2B). There was no difference in the prevalence of 87 medical diagnoses for admission to ICU between the C/T/A clade versus all the other clades (p>0.8). The C/T/A haplotype clade was associated with significantly greater prevalence of positive bacterial cultures at admission to ICU (p<0.03) (Figure 3-2), and specifically with increased prevalence of Gram-positive cultures (p<0.01) (Figure 3-3). Patients with at least one copy of the C/T/A haplotype clade had a relative risk of 1.2 for a Gram-positive culture at admission to ICU. The prevalence of positive bacterial cultures from an endovascular source (ie. positive blood cultures) was also significantly increased in patients who carried at least one copy of the C/T/A haplotype clade (p=0.0GT) (Figure 3-4). It is clinically significant to note that patients who carried at least one copy of the C/T/A clade had relative risk of 1.8 for a positive Gram-positive bacterial blood culture (p=0.006) (Figure 3-5). IRAK4 haplotypes were not associated with a significantly different prevalence of sepsis or septic shock upon admission to the ICU, or with 28-day survival (Table 3-5). IRAK4 haplotype clades were not associated with 28-day survival after adjusting for age, gender, APACHE II score, and surgical diagnosis for admission to ICU in a Cox proportional hazards regression analysis. Our study had >80% power (a=0.05) to detect an absolute difference of survival of 10% assuming that the cohort had an average survival rate of 60%. Mechanism With the purpose of explaining the association of the IRAK4 C/T/A haplotype clade with increased prevalence of Gram-positive infections we compared B-lymphocyte immune response to CpG by IRAK4 haplotype clade. The measured IL-6 concentrations were logarithmically (logio) transformed prior to statistical analysis as the data was not normally 88 distributed. The IRAK4 C/T/A haplotype clade was associated with significantly decreased IL-6 secretion compared to all other clades following stimulation of B-lymphocytes with lOuM CpG (raw values ± 95% C.L: 40.3pg/mL ± 32.3pg/mL vs. 85.8pg/mL ± 29.4pg/mL; log transformed values ± 95% C.I.: 1.13 ± 0.37 vs. 1.55 ± 0.18, p<0.04) or 50uM CpG (raw values ± 95% C.I.: 38.3pg/mL ± 103.3pg/mL vs. 81.6pg/mL ± 201.3pg/mL; log transformed values ± 9 5 % C.I.: 1.15 ±0 .41 vs. 1.55 ± 0.16, p<0.04) (Figure 3-6). In order to isolate the difference in immune response caused by the IRAK4 G29429A (Ala428Thr) SNP we transfected IDFs with IRAK4 expression plasmids containing either the 29429G or A allele and measured their immune response (as measured by supernatant IL-6 concentration) to TLR ligands. As a proof of principle we transfected IDFs with the wild-type IRAK4 expression plasmid containing the G allele (n=3) and compared the cell line's immune response to that of IDFs transfected with an empty vector (n=3) and to normal adult fibroblasts (NAFs) (n=3). While the NAFs had significantly higher IL-6 concentration in response to LPS and PGN (p<0.05) (Figure 3-7) compared to all IDFs, the IDFs transfected with the IRAK4 expression plasmid (containing the G allele of G294249A) produced significantly greater amounts of IL-6 in response to LPS and PGN stimulation than IDFs transfected with the empty vector (p<0.05) (Figure 3-7). Therefore, using the IRAK4 expression plasmid, we were able to successfully rescue the immune response of IDFs to TLR ligands (based on assessment of LPS- and PGN-stimulated IL-6 concentration). Once we had established that we could rescue the phenotype of the IDFs using the wild-type IRAK4 expression plasmid we successfully used site-directed mutagenesis to replace the IRAK4 29429G allele in the plasmid with an A at position 29429. We confirmed that we had exchanged the G for an A using the bl2seq tool of NCBI BLAST (Figure 3-8). 89 We successfully transfected IDFs with expression plasmids containing either the 29429G (IDF-G, n=3) or A allele (IDF-A, n=3) or with the empty expression vector (IDF-null, n=3). IDFs transfected with IRAK expression plasmids containing either the 29429G or A allele produced significantly higher levels of IL-6 in response to stimulation with LPS or PGN than IDFs transfected with the empty vector (p<0.05) (Figure 3-9). Although not significant, there was a trend to the IDF-A cell lines producing less IL-6 in response to stimulation with LPS (p=0.07) (Figure 3-9). There was no significant difference between the cell lines in the amount of IL-6 secreted in response to PGN (Figure 3-9). 3.4 Discussion We found that the IRAK4 haplotype clade marked by 23338C/24472T/29429A (C/T/A) was associated with increased prevalence of positive bacterial cultures on admission to the ICU in a cohort of 775 critically ill Caucasians with SIRS. Specifically the C/T/A clade was associated with increased prevalence of Gram-positive cultures and increased prevalence of positive blood cultures. No IRAK4 haplotype clade was associated with prevalence of sepsis or septic shock at admission, or with 28-day survival. Furthermore, we have demonstrated that the IRAK4 C/T/A clade is associated with decreased B-lymphocyte immune response to CpG (as measured by IL-6), and that fibroblasts transfected with the 29429A allele that marks the C/T/A clade show a trend to decreased IL-6 secretion in response to stimulation with LPS compared to IRAK4 29429G fibroblasts. To our knowledge, this is the first report of an association of genetic polymorphisms of IRAK4 with prevalence and type of bacterial cultures in a cohort of critically ill patients and supported by mechanistic evidence of decreased cellular immune response to TLR ligands. 90 Haplotypes within the C/T/A clade are uniquely distinguished from haplotypes in the other four clades of IRAK4 by 29429A. The substitution of an A for a G at position 29429 results in an alanine, a non-polar, alipathic amino acid, being replaced by a threonine, a polar, uncharged amino acid, at amino acid position 428 of the IRAK4 protein 2 3 . This amino acid substitution may disrupt IRAK4 signalling and result in a less effective immune response to invading pathogens so that critically ill patients carrying the C/T/A clade have increased risk of positive bacterial cultures. Our finding that the C/T/A haplotype clade of IRAK4 is associated with increased prevalence of positive bacterial cultures at admission to ICU is consistent with our previous studies 3 2 , and with recent animal studies 3 3 - 3 5 which suggest that polymorphisms of innate immunity genes could be associated with impaired clearance of 8bacteria. A number of studies have shown that rare germ-line mutations in IRAK4 cause recurring bacterial infections in children and deficiencies in cytokine production in response to a range of microbial-derived TLR agonists and to recombinant IL-lbeta or IL-18 1 2 ~ 1 6 ' 3 6 . Our finding that the IRAK4 C/T/A clade is associated with decreased B-lymphocyte immune response to CpG and with a trend to decreased immune response to LPS in IDFs transfected with the IRAK4 expression plasmid carrying the 29429A allele suggests that the replacement of Ala428 with a threonine decreases the cellular response to TLR ligands, perhaps impairing the host's ability to clear an infection. There are several strengths of our gene association study that minimized limitations of genetic association studies. Firstly, our large cohort of critically ill patients (n=775) reduced the risk of Type I error (finding a spurious association) compared to other studies of smaller sample size (n=20-100) 3 7 ~ 4 0 . Our large sample size also ensured that we had adequate power to determine that there was truly no association of IRAK4 haplotype clades with prevalence of sepsis and septic shock, or 28-day survival in our cohort of critically ill 91 patients. Second, we adjusted for confounders in our survival analysis (age, gender, A P A C H E II score, surgical diagnosis) and still found no association of IRAK4 haplotypes with 28-day survival. Third, to avoid spurious associations we included only Caucasians in our cohort of critically ill adults, thus limiting the risk of positive associations due to ethnic heterogeneity . Finally, by testing the immune response of cell lines of different IRAK4 haplotypes to TLR ligands, we have provided plausible mechanistic data to support our clinical association. An important strength of the design of our association study is our use of haplotypes and haplotype clades as the unit of genetic variation. Haplotype analysis is more effective in determining association of genotype with phenotype than is individual SNP analysis 4 1 , 4 2 . Haplotypes serve as markers of unidentified polymorphisms that may be the cause of phenotypic variation 4 1 . The phylogenetic history of haplotypes within a population may be determined and used to group haplotypes into clades using cladistic analysis 1 9 ' 30> 4 3 - 4 5 . Cladistic analysis has two unique strengths. First, grouping haplotypes into clades decreases the degrees of freedom, thereby increasing the statistical power to associate genotype with phenotype 4 2 . Second, grouping haplotypes into clades facilitates identification of causal SNPs. Another practical strength of our approach is that a small number of "haplotype tag" SNPs (htSNPs) can be used to distinguish haplotype clades, eliminating the need to genotype all SNPs within a gene 2 6' 4 6. There are several limitations of our gene association study. We have not examined how the 29429G/A (A428T) polymorphism affects the function or expression of the IRAK4 protein so we do not know the functional consequences of the C/T/A haplotype clade in the systemic inflammatory response syndrome. We also do not know how the amino acid 92 change at position 428 may affect the interaction of IRAK4 with other proteins in the MyD88-IRAKl -IRAK4 signalling pathway. Additionally, we were only able to show a trend to decreased IL-6 secretion in response to LPS stimulation of IRAK4-deficient fibroblasts transfected with the IRAK4 29429A allele (compared to the IRAK4 29429G allele). As this trend is in line with our finding of an association of the C/T/A clade with decreased IL-6 secretion in B-lymphocytes, we believe that it is an important result, although limited by the experimental design. The Ala428Thr amino acid change may only mildly disrupt IRAK4 function, and thus cause only minor changes in IL-6 secretion in response to stimulus with TLR ligands. Although we treated all cells identically at the same time, because we introduced IRAK4 into the IDFs using transient transfection it is possible that varying numbers of cells in each well of the transfection and stimulation experiments were viable and expressing IRAK4. This experimental variability may affect our ability to accurately measure the genetic variability in the immune response due to the IRAK4 G29429A polymorphism. In the future it may be appropriate to use stable transfection to introduce the IRAK4 gene into IDFs. Furthermore, although involved in wound repair, fibroblasts are not the ideal cells in which to be measuring variability of the innate immune response. The IRAK4-deficient fibroblasts were a generous gift to us from our collaborators and served to limit genetic heterogeneity at other loci. However, we may be able to observe greater genetic variability of the immune response to TLR ligands in monocytes or macrophages, cells that are much more sensitive to TLR ligands. In summary we have demonstrated a novel association between a haplotype clade of IRAK4 marked by G29429A (Ala428Thr) and increased risk of Gram-positive infection in a cohort of critically ill adults. We have shown that a possible explanation for this association is decreased cellular response to TLR ligands through the IRAK4 signalling pathway, 93 potentially inhibiting the host's ability to clear an infection. Future studies in a clinical cohort and in other cellular models will be needed to confirm these findings. 3.5 Tables and Figures Table 3-1. Primer and probe sequences used for genotyping IRAK4 htSNPs in the ABI Prism 7900HT Sequence Detection System htSNP C23338G Primer L T G T G C C T A T A G G A A G G A T C C A G A T T Primer R C C C T A C T G A A C A C A T C A T C T C A T T T C T Probe VIC C T C T T C C A T A G T A T C C T C - V I C Probe F A M C T C T T C C A T A C T A T C C T C - F A M T24472C Primer L C A T T T G C C T G G A G T G C C T T T C Primer R C A G A G G G T G A A A A G T G T G C T T A G T A Probe VIC T T T C C C T C A T A T A A C T A T T A A C Probe F A M C C T C A T A T A A C C A T T A A C G29429A Primer L G A A G A T T A T A T T G A T A A A A A G A T G A A T G A T G C T G A T T C C Primer R G C A G A C A T T G A C T A G C A A C A G A G T Probe VIC A C T T C A G T T G A A G C T A T G T Probe F A M A C T T C A G T T G A A A C T A T G T * VIC and F A M are fluorophores used to label the allele-specific hybridization probes. 95 Table 3-2. Synthetic bacterial DNA sequences (CpG) used for stimulation of B-lymphocytes Oligonucleotide Sequence CpG Oligonucleotide Type B 5' T C G T C G T T T T G T C G T T T T G T C G T T 3' (24mer) NCpG (Oligonucleotide Control) 5' T G C T G C T T T T G T G C T T T T G T G C T T 3' (24mer) 96 Table 3-3. Primers used for site-directed mutagenesis ( S D M ) and sequencing of the IRAK4 expression plasmid Primer Sequence SDM-sense 5 ' C C A C T T C A G T T G A A A C T A T G T A C T C T G T T G C T A G T C 3 ' SDM-anti-sense 5' G A C T A G C A A C A G A G T A C A T A G T T T C A A C T G A A G T G G 3' Sequencing primer 5' C G G G C T T C T G A G A A G T T T G C 3' 97 Table 3-4. IRAK4 haplotype frequencies and patient baseline characteristics by haplotype Baseline Characteristics Clade Frequency Age (yrs) % Female % Medical Dx A P A C H E II (mean±SD) (mean±SD) C/T/G 31% 57±16 37 80 ' 23±9 C/T/A 9% 59±16 36 74 23±9 C/C/G 10% 58±18 32 . 76 22±9 G/T/G 50% 58±17 37 73 23±9 P 0.4 0.7 0.05 0.7 98 Table 3-5. Prevalence of sepsis and septic shock at admission to ICU, and 28-day survival by IRAK4 haplotype clade Clade Sepsis at Septic Shock at 28-day Survival Days Alive Admission (%) Admission (%) (%) (mean±SD) C/T/G 81 55 66 21 C/T/A 81 53 63 20 C/C/G 74 52 ' 6 3 21 G/T/G 81 54 65 21 p 0.2 0.9 0.8 1.0 99 Figure 3-1. Haplotype structure of the Interleukin-1 receptor associated kinase 4 (IRAK4) gene. Haplotypes of the IRAK4 gene were inferred from unphased genotype data from 23 Caucasians using PHASE software. Columns are polymorphic sites of IRAK4. Rows are haplotypes of IRAK4 ordered by phylogenetic relationship. Yellow boxes are minor alleles and blue boxes are major alleles. Cladistic relationships of IRAK4 haplotypes were determined using MEGA2 phylogenetic software. There are 4 major haplotype clades of IRAK4, marked clades 1 through 4. C23338G, T24472C, and G29429A were chosen as a haplotype clade tag single nucleotide polymorphisms (htSNP) to distinguish between the 4 major haplotype clades. 100 A B C/T/A C/T/G C/C/G IRAK4 Haplotype Clade G/T/G C O "5 60 I < (0 3 o *p<0.03 RR=1.2 C/T/A clade No C/T/A clade C/T/A Haplotype Clade Figure 3-2. Prevalence of positive bacterial cultures at ICU admission by IRAK4 haplotype. Bacterial cultures were taken as part of routine medical care for any patients who were suspected of having an infection. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. Patients were genotyped at the IRAK4 htSNPs C23338G, T24472C, and G29429A by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System. Association of IRAK4 haplotype clades with prevalence of positive bacterial cultures was tested by chi-square analysis and the C/T/A clade appeared to be associated with increased prevalence of positive cultures (A). Subsequently, the prevalence of positive cultures at admission was compared between carriers of the IRAK4 C/T/A clade and non-carriers by chi-square analysis and haplotype clade relative risk was calculated (B). 101 C / T / A clade No C / T / A clade C/T/A Haplotype Clade Figure 3-3. Prevalence of Gram-positive bacterial cultures at ICU admission by IRAK4 haplotype clade. Cultures were categorized as Gram-positive, Gram-negative, fungal, or other. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. Patients were genotyped at the IRAK4 htSNPs C23338G, T24472C, and G29429A by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System. Association of the IRAK4 C/T/A haplotype clade with type of bacterial culture was tested by chi-square analysis and relative risk was calculated. 102 C/T/A clade No C/T/A clade C/T/A Haplotype Clade Figure 3-4. Prevalence of positive endovascular bacterial cultures (ie positive blood cultures) at ICU admission by IRAK4 haplotype clade. Culture sources were categorized as respiratory (sputum), gastrointestinal (peritoneal fluid, abscess drainage, biliary tract), skin (soft tissues or wounds), genitourinary (urine), endovascular (blood) or other. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. Patients were genotyped at the IRAK4 htSNPs C23338G, T24472C, and G29429A by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System. Association of the IRAK4 C/T/A haplotype clade with the source of positive bacterial culture was tested by chi-square analysis and haplotype relative risk was calculated. 103 Figure 3-5. Prevalence of Gram-positive endovascular bacterial cultures (ie positive blood cultures) at ICU admission by IRAK4 haplotype clade. Cultures were categorized as Gram-positive, Gram-negative, fungal, or other. Culture sources were categorized as respiratory (sputum), gastrointestinal (peritoneal fluid, abscess drainage, biliary tract), skin (soft tissues or wounds), genitourinary (urine), endovascular (blood) or other. Buffy coat was extracted from discarded whole blood. DNA was isolated by Qiagen DNA Blood Mini Kit. Patients were genotyped at the IRAK4 htSNPs C23338G, T24472C, and G29429A by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the AB1 Prism 7900HT Sequence Detection System. Association of the IRAK4 C/T/A haplotype clade with prevalence of positive Gram-positive blood cultures was tested by chi-square analysis and haplotype relative risk was calculated. 104 300 250 E 200 S 150 • C/T/A clade • No C/T/A clade Media NCpG 10uM Stimulant 50uM Figure 3-6. B-lymphocyte immune response to CpG (as measured by IL-6 supernatant concentration) by IRAK4 haplotype clade. B-lymphocytes of known IRAK4 genotype were stimulated with media, scrambled CpG (NCpG), lOuM or 50uM CpG, a TLR9 ligand, for 6 hours. Each stimulation was performed in triplicate. IL-6 supernatant concentration was measured using ELISA. Mean IL-6 concentrations were logarithmically (logio) transformed and compared between the C/T/A clade and all other IRAK4 clades using Student's t-test. The C/T/A clade was associated with significantly decreased IL-6 concentration (p<0.04). Graph displays mean IL-6 concentration ± 95% confidence interval. 105 • N A F • IDF-G H IDF-null L P S P G N Stimulant Figure 3-7. TLR-ligand stimulation of normal adult fibroblasts (NAFs) and IRAK4-deficient fibroblasts (IDFs). NAFs (n=3) and IDFs transfected with either the empty vector PGL3 (null) (n=3) or an IRAK4 expression plasmid carrying the 29429G allele (n=3) were stimulated for 6 hours with O.lpg/mL lipopolysaccharide (LPS) or lpg/mL peptidoglycan (PGN). Post-stimulation concentrations of interleukin-6 (IL-6) in the supernatant were measured by ELISA. Mean IL-6 concentration following either LPS or PGN stimulation was compared among NAFs, IDF-G and IDF-null cell lines by ANOVA. NAFs secreted significantly more IL-6 in response to LPS or PGN than IDF-G or IDF-null cell lines (p<0.05), while IDF-G cell lines had significantly greater IL-6 concentrations in response to LPS or PGN than IDF-null cell lines (p<0.05). Using the IRAK4 29429G expression plasmid we were able to successfully rescue the immune response of IDFs to TLR ligands. 106 Score 346 b i t s (180), Expect = le-92 I d e n t i t i e s = 182/183 (99%), Gaps = 0/183 (0%) Strand=Plus/Plus Query 1 G A A G A A A T T G A A G A T G A A G A A A A G A C A A T T G A A G A T T A T A T T G A T A A A A A G A T G A A T G A T 60 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I Sbjct 184 G A A G A A A T T G A A G A T G A A G A A A A G A C A A T T G A A G A T T A T A T T G A T A A A A A G A T G A A T G A T 243 Query 61 GC T G A T T C C AC T T C A G T T G AJJGi I I I I I I I I I I I I I I I I I I I I I Sbjct 2 44 GC T G A T T C C AC T T C A G T T G Ail Ai p T A T G T A C T C T G T T G C T A G T C A A T G T C T G C A T G A A A A G I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I T A T G T A C T C T G T T G C T A G T C A A T G T C T G C A T G A A A A G 120 303 Query 121 A A A A A T A A G A G A C C AGAC A T T A A G A A G G T T C A A C A G C T G C TGC A A G A G A T G A C AGC T T C T 180 I I II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I Sbjct 3 04 A A A A A T A A G A G A C C AGAC A T T A A G A A G G T T C A A C A G C T G C TGC A A G A G A T G A C AGC T T C T 3 63 Query 181 T A A 183 I I I Sbjct 3 64 T A A 3 66 Figure 3-8. BLAST result comparing sequences of wild-type IRAK4 expression plasmid containing the 29429G allele (query) with IRAK4 expression plasmid mutated to carry the 29429A allele (subject). The 29429G allele of the wild-type IRAK4 expression plasmid was mutated to carry the 29429A allele using site-directed mutagenesis. The two IRAK4 sequences are identical with the exception of the 29429 allele (box highlighted). 107 Figure 3-9. TLR-ligand stimulation of IRAK4-deficient fibroblasts (IDFs) transfected with IRAK4 expression plasmids containing either the 29429G or A alleles or with the empty vector. IDFs transfected with IRAK4 expression plasmids containing either the IRAK4 29429G (IDF-G) (n=3) or A (IDF-A) (n=3) alleles or the empty vector PGL3 (IDF-null) (n=3) were stimulated for 6 hours with O.lpg/mL lipopolysaccharide (LPS) or lpg/mL peptidoglycan (PGN). Post-stimulation concentrations of interleukin-6 (IL-6) in the supernatant were measured using ELISA. Mean IL-6 concentration was compared among IDF-G, IDF-A and IDF-null cell lines by ANOVA. IDF-G and IDF-A cell lines had significantly greater supernatant IL-6 concentrations compared to IDF-null cell lines (p<0.05). There was a trend to decreased IL-6 concentration in the IDF-A cell lines compared to the IDF-G cell lines in response to LPS stimulation (p=0.07). Graph displays mean IL-6 concentration ± 95% confidence interval. 108 3.6 References 1. Akira S, Takeda K, Kaisho T. Toll-like receptors: critical proteins linking innate and acquired immunity. Nat Immunol 2001 ;2:675-80. 2. Beutler B, Hoebe K, Du X, Ulevitch RJ. How we detect microbes and respond to them: the Toll-like receptors and their transducers. JLeukoc Biol 2003;74:479-85. 3. Burgner D, Levin M. Genetic susceptibility to infectious diseases. Pediatr Infect Dis J 2003;22:1-6. 4. Bellamy R, Hill A V . Genetic susceptibility to mycobacteria and other infectious pathogens in humans. Curr Opin Immunol 1998;10:483-7. 5. Choi EH, Zimmerman PA, Foster CB et al. Genetic polymorphisms in molecules of innate immunity and susceptibility to infection with Wuchereria bancrofti in South India. Genes Immun 2001;2:248-53. 6. Majetschak M , Obertacke U, Schade FU et al. Tumor necrosis factor gene polymorphisms, leukocyte function, and sepsis susceptibility in blunt trauma patients. Clin Diagn Lab Immunol 2002;9:1205-11. 7. Mira JP, Cariou A, Grail F et al. Association of TNF2, a TNF-alpha promoter polymorphism, with septic shock susceptibility and mortality: a multicenter study. Jama 1999;282:561-8. 8. Read RC, Pullin J, Gregory S et al. A functional polymorphism of toll-like receptor 4 is not associated with likelihood or severity of meningococcal disease. J Infect Dis 2001;184:640-2. 9. Sorensen TI, Nielsen GG, Andersen PK, Teasdale TW. Genetic and environmental influences on premature death in adult adoptees. N Engl J Med 1988;318:727-32. 10. Suzuki N, Suzuki S, Yeh WC. IRAK-4 as the central TIR signalling mediator in innate immunity. Trends Immunol 2002;23:503-6. 11. Yang K, Puel A, Zhang S et al. Human TLR-7-, -8-, and -9-mediated induction of IFN-alpha/beta and -lambda Is IRAK-4 dependent and redundant for protective immunity to viruses. Immunity 2005;23:465-78. 12. Currie AJ, Davidson DJ, Reid GS et al. Primary immunodeficiency to pneumococcal infection due to a defect in Toll-like receptor signalling. J Pediatr 2004;144:512-8. 13. Enders A, Pannicke U, Berner R et al. Two siblings with lethal pneumococcal meningitis in a family with a mutation in Interleukin-1 receptor-associated kinase 4. J Pediatr 2004; 145:698-700. 14. Chapel H, Puel A, von Bernuth H, Picard C, Casanova JL. Shigella sonnei meningitis due to interleukin-1 receptor-associated kinase-4 deficiency: first association with a primary immune deficiency. Clin Infect Dis 2005 ;40:1227-31. 15. Medvedev A E , Lentschat A, Kuhns DB et al. Distinct mutations in IRAK-4 confer hyporesponsiveness to lipopolysaccharide and interleukin-1 in a patient with recurrent bacterial infections. J Exp Med 2003;198:521-31. 16. Picard C, Puel A, Bonnet M et al. Pyogenic bacterial, infections in humans with IRAK-4 deficiency. Science 2003;299:2076-9. Epub 2003 Mar 13. 17. Davidson DJ, Currie AJ, Bowdish D M et al. IRAK-4 mutation (Q293X): rapid detection and characterization of defective post-transcriptional TLR/IL-1R responses in human myeloid and non-myeloid cells. J Immunol 2006;177:8202-11. 109 18. Stephens M , Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001 ;68:978-89. 19. Templeton AR, Weiss K M , Nickerson DA, Boerwinkle E, Sing CF. Cladistic structure within the human lipoprotein lipase gene and its implications for phenotypic association studies. Genetics 2000;156:1259-75. 20. Cardon LR, Bell JL Association study designs for complex diseases. Nat Rev Genet 2001;2:91-9. 21. Bone RC. The sepsis syndrome. Definition and general approach to management. Clin Chest Med 1996;17'Al'5-81. 22. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. A P A C H E II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 23. Rieder M CD ND. IRAK-4 Sequence Variation. 2004 vol: SeattleSNPs. NHLBI Program for Genomic Applications HL66682; 2002. 24. Kumar S, Tamura K, Jakobsen IB, Nei M . MEGA2: molecular evolutionary genetics analysis software. Bio informatics 2001 ;17:1244-5. 25. Gabriel SB, Schaffner SF, Nguyen H et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225-9. 26. Johnson GC, Esposito L, Barratt BJ et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001;29:233-7. 27. . Livak KJ. Allelic discrimination using fluorogenic probes and the 5' nuclease assay. Genet Anal 1999;14:143-9. 28. Baiyee EE, Flohe S, Lendemans S et al. Expression and function of Toll-like receptor 9 in severely injured patients prone to sepsis. Clin Exp Immunol 2006;145:456-62. 29. Dasari P, Nicholson IC, Hodge G, Dandie GW, Zola H. Expression of toll-like receptors on B lymphocytes. Cell Immunol 2005;236:140-5. 30. Long AD, Langley CH. The power of association studies to detect the contribution of candidate genetic loci to variation in complex traits. Genome Res 1999;9:720-31. 31. Guo SW, Thompson EA. Performing the exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics 1992;48:361-72. 32. Sutherland A M , Walley KR, Russell JA. Polymorphisms in CD14, mannose-binding lectin, and Toll-like receptor-2 are associated with increased prevalence of infection in critically ill adults. Crit Care Med2005;33:638-44. 33. Tasaka S, Ishizaka A, Yamada W et al. Effect of CD14 blockade on endotoxin-induced acute lung injury in mice. Am J Respir Cell Mol Biol 2003;29:252-8. 34. Opal SM, Palardy JE, Parejo N, Jasman RL. Effect of anti-CD14 monoclonal antibody on clearance of Escherichia coli bacteremia and endotoxemia. Crit Care Med 2003;31:929-32. 35. Echchannaoui H, Frei K, Schnell C, Leib SL, Zimmerli W, Landmann R. Toll-like receptor 2-deficient mice are highly susceptible to Streptococcus pneumoniae meningitis because of reduced bacterial clearing and enhanced inflammation. J Infect Dis 2002;186:798-806. 36. Day N, Tangsinmankong N, Ochs H et al. Interleukin receptor-associated kinase (IRAK-4) deficiency associated with bacterial infections and failure to sustain antibody responses. J Pediatr 2004;144:524-6. 37. Arnalich F, Lopez-Maderuelo D, Codoceo R et al. Interleukin-1 receptor antagonist gene polymorphism and mortality in patients with severe sepsis. Clin Exp Immunol 2002;127:331-6. 110 38. Gibot S, Cariou A, Drouet L, Rossignol M , Ripoll L. Association between a genomic polymorphism within the CD14 locus and septic shock susceptibility and mortality rate. Crit Care Med 2002;30:969-73. 39. Lowe PR, Galley HF, Abdel-Fattah A, Webster NR. Influence of interleukin-10 polymorphisms on interleukin-10 expression and survival in critically ill patients. Crit Care Med 2003;31:34-8. 40. Schaaf B M , Boehmke F, Esnaashari H et al. Pneumococcal septic shock is associated with the interleukin-10 - 1082 gene promoter polymorphism. Am J Respir Crit Care Med 2003; 168:476-80. 41. Akey J, Jin L, Xiong M . Haplotypes vs single marker linkage disequilibrium tests: what do we gain? Eur J Hum Gene? 2001;9:291-300. 42. Zhang K, Calabrese P, Nordborg M , Sun F. Haplotype block structure and its applications to association studies: power and study designs. Am J Hum Genet 2002;71:1386-94. 43. Templeton AR, Crandall K A , Sing CF. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 1992;132:619-33. 44. Templeton AR. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping or DNA sequencing. V. Analysis of case/control sampling designs: Alzheimer's disease and the apoprotein E locus. Genetics 1995;140:403-9. 45. Braaten O, Rodningen OK, Nordal I, Leren TP. The genetic algorithm applied to haplotype data at the L D L receptor locus. Comput Methods Programs Biomed 2000;61:1-9. 46. Weiss K M , Clark A G . Linkage disequilibrium and the mapping of complex human traits. Trends Genet 2002;18:19-24. I l l Appendix 3-1. Coriell Cell Repository IDs, SeattleSNP IDs and IRAK4 G29429A genotypes of B-lymphocyte cell lines used in mechanistic studies Coriell Cell Line ID Seattle SNPs ID G29429A Genotype GM17109 D009 A A GM17134 D034 A A GM07349 E009 A G GM10844 E014 A G GM10854 E015 A G GM10858 E011 A G GM17103 D003 A G GM17104 D004 A G GM17106 D006 A G GM17106 D006 A G GM17111 D011 A G GM17113 D013 A G GM17114 D014 A G GM17133 D033 A G GM17140 D040 A G GM06990 E022 G G GM07019 E023 GG GM07348 E020 GG GM10831 E017 GG GM10842 E007 GG GM10843 E018 GG GM10843 E018 GG GM10845 E003 GG GM10848 E012 GG GM10850 E019 GG GM10851 E008 G G GM10852 E021 G G GM10853 E004 GG GM10857 E010 GG GM10860 E005 GG GM12547 E002 G G GM12548 E013 GG GM12560 E001 GG GM17107 D007 GG GM17108 D008 G G GM17110 D010 GG GM17112 D012 GG GM17115 D015 GG GM17116 D016 GG GM17135 D035 GG GM17136 D036 GG GM17138 D038 GG GM17139 D039 GG CHAPTER 4: THE ASSOCIATION OF INTERLEUKIN 6 HAPLOTYPE CLADES WITH OUTCOME FROM THE SYSTEMIC INFLAMMATORY RESPONSE 4.1 Introduction A bacterial, viral, or other infection that results in a Systemic Inflammatory Response Syndrome (SIRS) is termed sepsis. Sepsis and septic shock have survival rates of approximately 70% and 35-45% respectively '. Activation of the inflammatory response to infection, including IL-6 production, varies significantly between individuals, in part due to genetic variation between individuals. Genotype has been shown to contribute substantially to outcome in sepsis 2 ' 3 ' 4 . The genetic contribution to death from sepsis exceeds the inherited genetic contribution to cancer risk of death by many fold and even exceeds the genetic contribution to cardiovascular disease risk 5 . Variation in key inflammatory genes such as IL-6 may explain variation in individuals' responses to infection. This variation has important clinical implications as therapeutic strategies targeting these pathways may be highly effective in patients having specific genotypes yet may have adverse effects in other patients with alternative genotypes. It is essential to now test genetic variants for association to outcome in complex inflammatory diseases, such as SIRS and sepsis, in order to incorporate genotype into the design of patient-tailored therapy. Genotype will be used to design therapeutic strategies to optimize outcome while minimizing adverse effects. It has been well established that a large proportion of patients experience SIRS following cardiac surgery with cardiopulmonary bypass (CPB). The systemic inflammatory response develops as a result of interaction of the blood with the cardiopulmonary bypass machine 6 , the release of endotoxins from the gut into the blood 1 , and ischemia-reperfusion injury 8 . Activation of inflammatory mediators may be responsible for the occurrence of a 113 vasodilatory syndrome following CPB 6 ' 9 - 1 ' . While it is known that the magnitude of the inflammatory response is controlled by many external factors including the duration of CPB, type of apparatus used, body temperature during CPB and pharmacologic agents used 1 2 - 1 6 ; patient morbidity following CPB is still difficult to predict and may be influenced by single nucleotide polymorphisms in important inflammatory genes in the same way that genetic variation in inflammatory genes may contribute to clinical outcome from SIRS and sepsis. lnterleukin-6 (IL-6) is a particularly important immunomodulatory cytokine in the pathogenesis and outcome of SIRS, sepsis, and septic shock. Increased plasma concentrations of IL-6 are found in patients with sepsis 1 7 ' 1 8 , and high IL-6 concentration is associated with increased mortality 1 9 . Furthermore, in a related critical illness, high plasma concentrations of IL-6 on day 1 of acute respiratory distress syndrome were also associated 9 0 with increased mortality . IL-6 is also a key mediator of the inflammatory response following cardiac surgery with CPB. It has been widely shown that IL-6 levels increase significantly following CPB 2 1 ~ 2 6 . High serum levels of IL-6 are associated with postoperative hyperdynamic circulatory dysregulation 2 6 , lactic acidosis 2 6 , increased bleeding 2 4 , prolonged respiratory support 2 4 ' 2 5 and with a greater degree of postoperative renal dysfunction 2 5 . In order to predict inter-individual responses to CPB a number of studies have begun to examine the influence of polymorphisms in the IL-6 gene with serum levels of IL-6 and clinical outcome following CPB. Therefore, we chose IL-6 as a candidate gene for this association study. There have been a number of studies examining how genetic variation within the IL-6 gene may affect outcome in sepsis and other inflammatory conditions. The C allele of a G-to-C single nucleotide polymorphism (SNP) at position -174 of the promoter region of the 114 IL-6 gene is associated with decreased levels of IL-6 1 1 . However, one study found there was no association between the -174G/C polymorphism and incidence of sepsis, although -174G 28 homozygosity was associated with improved survival rates in patients with sepsis . There have been conflicting findings in previous studies of IL-6 polymorphisms and IL-6 serum concentrations following cardiac surgery. Two early studies in the same group of patients showed an association between GG homozygosity at the -174G/C polymorphism and 2 S increased serum IL-6 concentrations ". GG homozygotes had longer stays in the ICU and hospital and increased rates of postoperative mortality, myocardial infarction, stroke, and 25 renal and pulmonary dysfunction . A third, smaller study found no association between the IL-6 -174G/C polymorphism and IL-6 serum concentrations 2 1 . Thus, the role of IL-6 genotype in SIRS, sepsis, septic shock and in the post-cardiac surgery inflammatory response is not clear. Recently, it has been recognized that single SNP analysis may be less informative than analysis based on haplotypes. Haplotypes are patterns of several SNPs that are in linkage disequilibrium with one another within a gene or segment of DNA, and are thus 9 0 ^ 0 inherited as a unit (Figure 4-1) ' . Single SNP approaches for genetic association studies only work well when the examined SNP is causal or is in significant linkage disequilibrium with the causal SNP. In contrast, haplotypes serve as markers for all measured and 9 0 unmeasured alleles within the haplotypes . A haplotype-based approach can narrow the search for causal SNPs and haplotypes can serve as predictors of disease severity 2 9 ' 3 0 . Multiple haplotypes exist for all studied genes, thus grouping haplotypes into haplotype clades (evolutionarily related groups of haplotypes) increases the statistical power to associate genotype with phenotype 3 1 . 115 Regardless of the unit of genetic variation used, it is important to validate genotype-phenotype associations in an independent cohort. Because multiple haplotype clades are tested for association to multiple outcome measures there is increased risk of finding false positive associations. To limit the risk of false positive associations, positive associations must be validated in a larger cohort to correct for multiple comparisons and a somewhat heterogeneous patient cohort, as well as to ensure proper power to rule out false negative associations. A major limitation of haplotype association studies is that association of haplotypes or haplotype clades with outcome does not identify the causal SNP or provide a mechanism through which genetic variation alters phenotype 2 9 , 3 2 . For example, the association of haplotypes of IL-6 with increased mortality does not provide a pathological mechanism to explain the increase in mortality. Even when a SNP associated with outcome is in a promoter region or exon and is found to alter gene expression or protein function in vitro, this does not prove its function in vivo. For these reasons it is important to measure intermediate phenotypes such as in vivo tissue RNA expression or serum protein levels . We therefore chose to study CPB patients to test for association between IL-6 haploytpe clades and intermediate phenotype in order to 1) obtain a baseline sample, 2) clearly define onset of SIRS as the start of surgery/CPB, 3) make repeated measures of serum concentrations of cytokines over 24 hours and 4) increase power to associate genotype with intermediate phenotype as cardiac surgery patients are less heterogeneous than critically ill patients with SIRS. Accordingly, our first aim was to determine whether haplotypes of IL-6 were associated with adverse outcomes in a derivation cohort of critically ill adults who had SIRS. 116 Our second aim was to determine whether the previously described IL-6 -174G/C polymorphism could account for any haplotype-clinical outcome association. To accomplish 33 this, the haplotype structure of the IL-6 gene was determined using publicly available data . Cladistic analysis was then used to group these haplotypes into related clades (Figure 4-1) and subsequently determined a minimum set of "haplotype tag" SNPs (htSNPs) (Figure 4-1) that defined all 4 major haplotype clades of the IL-6 gene 3 1* 3 4 , 3 5 . The association of these haplotype clades with 28-day mortality and organ dysfunction in a derivation cohort of critically ill adults who had SIRS was then tested. Subsequently, in order to validate associations of IL-6 haplotype clades with clinical outcome, the set of htSNPs was genotyped in a separate validation cohort of critically ill Caucasians and tested for the association of these haplotype clades with clinical outcome. Finally, to lend biological plausibility to genotype-phenotype associations, the IL-6 htSNPs were genotyped in a secondary cohort of coronary artery bypass graft (CABG) surgery with CPB patients and tested for the association of IL-6 haplotype clades with the occurrence of a vasodilatory syndrome after cardiac surgery and with pre- and post-operative serum levels of inflammatory mediators. 4.2 Methods This study was approved by the Providence Health Care / University of British Columbia Research Ethics Board. Derivation Cohort of Critically 111 Patients 337 consecutive patients admitted to the tertiary mixed medical-surgical Intensive Care Unit (ICU) of St. Paul's Hospital were screened for inclusion into the derivation cohort. We only report the results for the Caucasian patients (n=263) in order to decrease the potential confounding influence of population admixture secondary to ethnic diversity on 117 associations between genotype and phenotype 3 6 . Patient data were screened daily and patients were included in the derivation cohort (n=228) if they met at least two of four SIRS criteria and were successfully genotyped at all three htSNPs of IL-6. The SIRS criteria were 1) fever (>38 C) or hypothermia (<36 C), 2) tachycardia (>100 beats/min in the absence of beta blockers), 3) tachypnea (>20 breaths/min) or need for mechanical ventilation, and 4) leukocytosis (total leukocyte count > 12,000/uL) or leukopenia (total leukocyte count < 4,000/uL) \ Patients were included in this cohort on the calendar day that 2/4 SIRS criteria were met. Validation Cohort of Critically 111 Patients 775 consecutive patients admitted to the tertiary mixed medical-surgical Intensive Care Unit (ICU) of St. Paul's Hospital were screened for inclusion into the validation cohort. Patient data were screened daily and Caucasian patients were included in the validation cohort (n=441) if they met at least two of four SIRS criteria and were successfully genotyped at all three htSNPs of IL-6. Patients were included in this cohort on the calendar day that 2/4 SIRS criteria were met. Clinical Phenotype of Critically 111 Patients 28-day mortality was the primary outcome variable. Secondary outcome variables were days alive and free of SIRS, days alive and free of shock, and days alive and free of organ dysfunction (cardiovascular, respiratory, renal, hepatic, hematologic, and neurologic organ systems). Clinically significant organ dysfunction for each organ system was defined as present during a 24 hour period if there was evidence of at least moderate organ dysfunction using the Brussels criteria (Table 4-1) 3 7 . 118 Because data were not always available during each 24 hour period for each organ dysfunction variable, we used the "carry forward" assumption as defined previously 3 8 . Briefly, for any 24 hour period in which there was no measurement of a variable, we carried forward the measurement from the previous 24 hour period. If any variable was never measured, it was assumed to be normal. Clinical data were recorded for 28 days or until hospital discharge. To assess duration of organ dysfunction and SIRS and to correct organ dysfunction and SIRS scoring for deaths in the 28-day observation period, we calculated days alive and free of organ dysfunction. Briefly, during each 24 hour period (8 am to 8 am) for each variable, days alive and free was scored as 1 if the patient was alive and free of organ dysfunction (normal or mild dysfunction). Days alive and free was scored as 0 if the patient had organ dysfunction (moderate or worse) or was not alive. Every day over the 28-day observation after ICU admission was scored in this way. Thus, the lowest score possible for each variable was zero and the highest score possible was 28. A low score indicates more organ dysfunction, while a high score indicates less organ dysfunction. Baseline demographic data included age, gender, medical versus surgical diagnosis on admission (according to A P A C H E III diagnostic codes) , and admission APACHE II score. Cardiopulmonary Bypass Patient Cohort 911 consecutive patients undergoing elective coronary artery bypass graft surgery with cardiopulmonary bypass at St. Paul's Hospital were screened for inclusion into this study. Valve replacement patients and off-pump patients were excluded from the cohort. 119 We only report the results for the Caucasian patients (n=603) in. order to decrease the potential confounding influence of population admixture secondary to ethnic diversity on 36 associations between genotype and phenotype . Baseline demographics recorded were age, gender, the presence of diabetes mellitus, body mass index, smoking, use of angiotensin-converting enzyme inhibitors, use of beta-blockers, use of calcium channel blockers and use of immunosuppressives. Surgeon, pump-time, cross-clamp time, total surgery time, use of intra-operative milrinone, aprotinin and amicar were recorded. Clinical Phenotype of Cardiopulmonary Bypass Patients Kristof and Magder 4 0 found that low systemic vascular resistance index (SVRI) was a particularly useful clinical manifestation of systemic inflammation. They found that vasodilatory syndrome identified by low SVRI was associated with important clinical parameters including longer cross-clamp times and lower post-CPB platelet count 4 0 . Low systemic vascular resistance (SVR) defined as an indexed low systemic vascular resistance (SVRi) less than 1800 dyne*sec/cm5/m2 at 2 consecutive time points post-operatively 4 0 was the primary outcome variable (SVRi = [(MAP-CVP)*80]/Cl, where MAP is the mean arterial pressure, CVP is the central venous pressure and CI is the cardiac index). Because mortality after CPB is extremely low, we measured hours in cardiac surgery intensive care unit (CSICU) after surgery as a clinically relevant secondary outcome variable. Patients were followed until hospital discharge or death. 120 Intermediate Phenotype of Cardiopulmonary Bypass Patients Blood was drawn in 162 CPB patients as part of a separate randomized controlled trial. 12ml of arterial blood was drawn into 6ml EDTA tubes pre-operatively and at 3 and 24 hours post-operatively. IL-6, monocyte chemotactic protein-1 (MCP-1), granulocyte-colony stimulating factor (G-CSF), IL-8 and IL-1 receptor antagonist (IL-lra) were measured in patients' plasma using the Luminex. 100IS system for multiplex assays with reagents from R&D Systems (Minneapolis, MN) according to the manufacturer's directions. The Luminex system uses 5.6pm polystyrene beads which are internally dyed with red and infrared fluorophores. Each bead set is filled with different ratios of the two dyes so that each bead has a unique spectral signature based on the red/infrared ratio. Beads are then coated with antibodies specific to the protein of interest. A secondary antibody with a biotin tag is added, and finally avidin-tagged reporter dyes. Beads run through the capillary in the Luminex. 100IS system and lasers excite the internal dyes that identify each bead, and also any reporter dye captured during the assay. Fluorescent intensities from samples and controls were measured and used to quantify cytokine concentrations in patient serum samples using analytical software from R&D Systems (Minneapolis, MN). Blood Collection and Processing for Genomic DNA Discarded blood from routine clinical laboratory tests was collected for patients in the derivation and validation cohorts of critically ill patients and the cardiopulmonary bypass patient cohort. Blood samples were collected for 162 CPB patients as part of a separate randomized controlled trial. Samples were centrifuged at 400g for 15 minutes at 4°C. In all cases the buffy coat was collected and transferred into 2.0mL cryotubes and stored at -80°C. 121 DNA was extracted from the buffy coat using the Qiagen DNA Blood Mini Kit (Qiagen Inc. Mississauga, ON). Serum from the blood collected from the subset of cardiopulmonary bypass patients who had fresh blood drawn was withdrawn with sterile glass pipettes and transferred to 2.0mL cryotubes and stored at -80°C. Haplotypes and Selection of htSNPs We used unphased genotypic data from 23 Caucasians from the Coriell registry (from pga.mbt.washington.edu) to infer haplotypes of the IL-6 gene using PHASE software (Figure 4-1) 3 3 ' 4 1 . We then used MEGA2 software to infer a phylogenetic tree to identify major haplotype clades 4 2 (Figure 4-2). Haplotypes were sorted into clades according to this phylogenetic tree and this haplotype structure was inspected to choose a minimum set of "haplotype tag" single nucleotide polymorphisms (htSNPs) (Figure 4-1) 3 4 ' 3 5 . We chose 3 htSNPs that identified 4 major haplotype clades of IL-6 in Caucasians. One of these SNPs was the previously reported promoter polymorphism -174G/C (position, NCBI ID: 1510G/C, rsl800795). The second SNP was a C-to-G substitution at nucleotide 1753 relative to the start (3437C/G, rs2069840). The third SNP was a G-to-C substitution at nucleotide 2954 (4638G/C, rsl548216) (NCBI IL-6 accession number AF372214). These three SNPs were then genotyped in the derivation cohort of critically ill patients and PHASE was used to identify haplot^es in each patient. Genotyping The genotypic analysis was performed in a blinded fashion, without clinical information. In the original derivation cohort, patients' genotypes at positions 1753 and 2954 were determined by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection 122 System (Applied Biosystems, Inc., Foster City, CA) 4 3 . Five nanograms of patients' genomic DNA was used per genotyping reaction in a 384-well plate. DNA with known genotype from 23 lymphocyte cell lines from the Coriell Cell Repository demonstrated complete concordance with our genotyping at positions 1753 and 2954 of the IL-6 gene 3 3 . Genotyping of -174G/C was done using both a previously reported PCR-RFLP method and using a MALDItof approach (Qiagen Inc., Mississauga, Ontario, Canada) and again demonstrated complete concordance 4 4 ' 4 5 . Individuals in the validation cohort of critically ill patients and the cohort of cardiopulmonary bypass patients were genotyped at positions 1753 and 2954 by real-time polymerase chain reaction (PCR) assay using specific fluorescence-labelled hybridization probes in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Inc., Foster City, CA) 4 3 . The -174G/C htSNP failed initial primer design for the TaqMan genotyping system by Applied Biosystems Inc.; 614G/A (2298G/A, rs2069832) was chosen as an alternate htSNP using the haplotype clade structure of the IL-6 gene. The htSNP IL-6 614G/A is in complete linkage disequilibrium with IL-6 -174G/C (D'=0.96). Individuals in the validation cohort of critically ill patients and the cohort of cardiopulmonary bypass patients were genotyped at 614G/A by real-time PCR assay in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Inc., Foster City, CA) 4 3 . DNA with known genotype from 23 lymphocyte cell lines from the Coriell Cell Repository demonstrated complete concordance with our genotyping at positions 614, 1753 and 2954 of the IL-6 gene . Ten percent of the genotyping in the two latter cohorts was repeated for quality control. Statistical Analysis We used a cohort study design. In the initial derivation cohort of critically ill patients, rates of dichotomous outcomes (28-day mortality, sepsis and shock at onset of 123 SIRS) were compared between clades. The C/C/G, G/G/G and G/C/C clades appeared to be associated with increased 28-day mortality and organ dysfunction. Further comparisons were then made between patients carrying two haplotypes from within the C/C/G, G/G/G or G/C/C haplotype clades and patients carrying only one or no copies of a haplotype from within the C/C/G, G/G/G or G/C/C haplotype clades using a chi-square test, assuming a recessive model of inheritance. Differences in continuous outcome variables between patients with two haplotypes from within the haplotype clades C/C/G, G/G/G or G/C/C and patients carrying only one or no copies of a haplotype from within the C/C/G, G/G/G or G/C/C haplotype clades were tested using Student's t-test. 28-day mortality was further compared between the two groups of patients while adjusting for other confounders (age, sex, and medical vs. surgical diagnosis) using a Cox regression analysis as well as a Kaplan-Meier analysis of censored survival data. Haplotype clade relative risk was calculated. Genotype distributions were tested for Hardy-Weinberg equilibrium 4 6 . We report the mean and 95% confidence intervals. Statistical significance was set at p < 0.05. IL-6 haplotype clades that were found to be associated with clinical outcome in the derivation cohort were tested for association to the same clinical outcomes in the validation cohort. In the validation cohort the htSNP 614G was equivalent to the htSNP -174G that was tested in the derivation cohort. Statistics were performed as in the derivation cohort. In the cohort of cardiopulmonary bypass patients rates of dichotomous variables were compared using Chi square and continuous variables were compared between patients carrying the 614G/1753C/2954G (G/C/G) clade and patients with two haplotypes from within the haplotype clades A/C/G, G/G/G or G/C/C using Student's t-test. Pre- and post-operative median serum concentrations of IL-6, MCP-1, G-CSF, IL-8 and ILlra were 124 compared between patients carrying the 614G/1753C/2954G (G/C/G) clade and patients with two haplotypes from within the haplotype clades A/C/G, G/G/G or G/C/C using the Mann-Whitney U-test. The data was analyzed using SPSS 11.5 for Windows (SPSS Inc, Chicago, IL, 2003). 4.3 Results Derivation Cohort of Critically 111 Patients We were able to infer haplotypes from complete sequencing of IL-6 for 23 Caucasians in the Coriell Cell Repository using PHASE software, and identified four major haplotype clades using MEGA2 software (Figure 4-1) 4 1 ' 4 1 . These four clades could be resolved by genotyping three htSNPs; -174G/C, 1753C/G, and 2954G/C, in our 228 patient cohort 3 4. The -174C/1753C/2954G (C/C/G) haplotype clade occurred with a frequency of 43.2%, the G/G/G haplotype clade occurred with a frequency of 32.9%, the G/C/G haplotype clade occurred with a frequency of 20.6%, and the G/C/C haplotype clade occurred with a frequency of 3.1%. A fifth haplotype defined by -174C/1753G/2954G occurred only once (0.2%) and thus was not included in further presentation of this data. These frequencies were 33 similar to frequencies deduced from other available Caucasian data . For the 228 successfully genotyped individuals of the cohort of Caucasian patients who had at least 2 of 4 SIRS criteria, no haplotype clade of IL-6 was significantly associated with a difference in age, gender, medical vs. surgical diagnoses at admission, or severity of illness at the time of admission (as estimated by the A P A C H E II score) (Table 4-2). The IL-6 haplotype clades C/C/G, G/G/G and G/C/C were found to be associated with greater 28-day mortality rates than the G/C/G haplotype clade. Patients who carried any 125 combination of two haplotypes from within the haplotype clades C/C/G, G/G/G or G/C/C had significantly greater 28-day mortality than patients who carried at least one copy of the G/C/G haplotype clade (p=0.015) (Figure 4-3). Carrying two copies of the C/C/G, G/G/G or G/C/C haplotype clades conferred a relative risk of mortality of 1.27. After adjusting for other predictors of survival (age, sex, medical vs. surgical diagnosis at admission) using a Cox regression analysis, patients carrying two haplotypes from within the haplotype clades C/C/G, G/G/G or G/C/C still had significantly greater rates of 28-day mortality with a hazard ratio of 1.83 (p<0.02) (Table 4-3). Kaplan-Meier analysis of censored 28-day survival data confirmed that patients carrying 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had increased probability of mortality during the 28-day observation period (p=0.013) (Figure 4-4). Patients carrying 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had significantly fewer days alive and free of cardiovascular dysfunction (Figure 4-5 A) (p=0.006) and required more cardiovascular support as measured by fewer days alive and free of vasopressors (Figure 4-5A) (p=0.01). Patients carrying 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades also had significantly fewer days alive and free of Acute Lung Injury (ALI) (Figure 4-5B) (p=0.002) and fewer days alive and free of ventilation (Figure 4-5B) (p=0.04). The C/C/G, G/G/G and G/C/C haplotype clades were associated with fewer days alive and free of renal dysfunction (Figure 4-5C) (p=0.005) and a trend to fewer days alive and free of renal support (Figure 4-5C) (p=0.07). Furthermore, patients who carried 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had significantly more hematologic system dysfunction (18.7 days alive and free of hematologic dysfunction for patients with 2 copies of C/C/G, G/G/G or G/C/C vs. 21.0 days for patients with one or no copies, p=0.02). 126 Patients who carried 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had significantly fewer days alive and free of 4/4 (18.4 days alive and free vs. 21.1 days, p=0.025) and 3/4 SIRS criteria (13.9 days alive and free vs. 17.4 days, p=0.006) than patients who carried one or no copies of the C/C/G, G/G/G or G/C/C haplotype clades. In contrast to the strong association between the IL-6 C/C/G, G/G/G and G/C/C haplotype clades with adverse clinical outcome, there was no association between any of the individual SNPs and the clinical outcome variables. Specifically, there was no association between -174G/C and increased 28-day mortality. There was also no significant association between the -174G /C polymorphism and cardiovascular, hepatic, hematologic, or neurologic system dysfunction. The genotype distribution of-174G/C was similar to the frequencies reported in other critically ill Caucasians (Table 4-4) 2 8 . The sample population was in Hardy-Weinberg equilibrium for all genotyped SNPs 4 6 . Validation Cohort of Critically 111 Patients The four haplotype clades of IL-6 were resolved by genotyping 614G/A, 1753C/G, and 2954G/C in the 441 patient validation cohort 3 4 . The 614A/1753C/2954G (A/C/G) haplotype clade which was equivalent to the -174C/1753C/2954G (C/C/G) clade in the derivation cohort occurred with a frequency of 36.4%, the G/G/G haplotype clade occurred with a frequency of 35.1%, the G/C/G haplotype clade occurred with a frequency of 25.5%, and the G/C/C haplotype clade occurred with a frequency of 2.9%. The distribution of IL-6 haplotype clade frequencies was not significantly different between the derivation and validation cohorts (Table 4-5). The genotype and allele frequencies in the validation cohort for each IL-6 htSNP are displayed in Table 4-6. 127 No haplotype clade of IL-6 was significantly associated with a difference in age, gender, medical vs. surgical diagnoses at admission, or severity of illness at the time of admission (as estimated by the A P A C H E II score) in the validation cohort (Table 4-7). In contrast to our previous findings in the derivation cohort, there was no association between the IL-6 A/C/G, G/G/G and G/C/C haplotype clades and greater 28-day mortality in the validation cohort of critically ill patients. There was no difference in 28-day mortality rates between patients carrying 2 copies of the A/C/G, G/G/G or G/C/C haplotype clades and patients who carried at least one copy of the G/C/G haplotype clade (Figure 4-6), even after adjusting for other predictors of survival (age, sex, medical vs. surgical diagnosis at admission) using a Cox regression analysis (Table 4-8). Additionally, there was no significant difference in days alive and free of organ dysfunction and organ support between patients carrying 2 copies of the A/C/G, G/G/G or G/C/C haplotype clades and patients who carried at least one copy of the G/C/G haplotype clade. As in the derivation cohort there was no association between any of the individual htSNPs and outcome in the validation cohort. The genotype distribution of the htSNPs was similar to the frequencies reported in other critically ill Caucasians (Table 4-4) 2 8 and in the derivation cohort of critically ill Caucasians. The validation cohort was in Hardy-Weinberg equilibrium for all genotyped SNPs 4 6 . Cohort of Cardiopulmonary Bypass Patients 603 patients undergoing cardiac surgery with cardiopulmonary bypass were successfully genotyped for IL-6 614G/A, 1753C/G, and 2954G/C. The 614A/1753C/2954G (A/C/G) haplotype clade occurred with a frequency of 39%, the G/G/G haplotype clade 128 occurred with a frequency of 31%, the G/C/G haplotype clade occurred with a frequency of 28%, and the G/C/C haplotype clade occurred with a frequency of 2% (Table 4-9). These frequencies were similar to the frequencies of the equivalent haplotype clades tagged by -174G/C, 1753C/G, and 2954G/C in the derivation cohort of critically ill patients. Based on our findings in the derivation cohort of critically ill patients that haplotype clades of IL-6 were associated with greater 28-day mortality and more organ dysfunction in critically ill patients with SIRS we tested the hypothesis that cardiac surgery patients who carried any combination of two haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades would be at increased risk for developing a vasodilatory syndrome compared to patients carrying at least one copy of the G/C/G clade. There were no significant differences in age, gender, BMI, prevalence of diabetes mellitus, smoking, use of A C E inhibitors, P-blockers, Ca2+ channel blockers, or immunosuppressives between cardiac surgery patients carrying 2 haplotypes from within the A/C/G, G/G/G or G/C/C clades and patients carrying at least one copy of the G/C/G clade (Table 4-10). There were also no significant differences in duration of CPB, cross-clamp and surgery, or in the proportion of patients in each group who received intraoperative milrinone, aprotinin, or aminocaproic acid (amicar) (Table 4-11). No patient received intraoperative protamine. A significantly greater proportion of patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades experienced a vasodilatory syndrome than patients who carried 2 copies of the G/C/G haplotype clade. 56.2% of patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades had two consecutive 129 measurements of SVRI that were less than 1800 dyne*sec/cm5/m2, while only 48.6% of patients who carried 2 copies of the G/C/G haplotype clade had two consecutive measurements in this category (p<0.05) (Figure 4-7). In a subgroup of patients who were put on vasopressors post-operatively, 58.2% of patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades had a vasodilatory syndrome compared to only 47.0% of patients who carried 2 copies of the G/C/G haplotype clade (p<0.002) (Figure 4-8). Although the A/C/G, G/G/G and G/C/C haplotype clades were associated with occurrence of a vasodilatory syndrome in cardiac surgery patients, they were not associated with hours spent in the ICU following surgery (Figure 4-9). Serum concentrations of IL-6, MCP-1, G-CSF, IL-8 and IL-lra were measured in a subgroup of cardiac surgeon patients pre-operatively and at 3 and 24 hours post-operatively. Serum cytokine concentrations were not normally distributed and so were compared between patients using a Mann-Whitney U test. There was no difference in median serum concentrations between patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades and patients who carried 2 copies of the G/C/G haplotype clade (Table 4-12). 4.4 Discussion We found that the C/C/G, G/G/G and G/C/C haplotypes clades of IL-6 were strongly and significantly associated with increased 28-day mortality and increased organ dysfunction in a derivation cohort of critically ill Caucasians who had SIRS. Because none of the individual htSNPs used to define the haplotype clades of IL-6 were associated with survival, 130 haplotype analysis of IL-6 proved to be a more valuable tool in identifying genetic associations than individual SNP analysis in the derivation cohort. In a separate validation cohort of critically ill Caucasians with SIRS we were unable to validate the association of IL-6 haplotype clades with clinical outcome. In the validation cohort neither haplotype clades nor individual htSNPs of IL-6 were associated with 28-day mortality or increased organ dysfunction. In a separate cohort of cardiac surgery patients we did, however, detect an association between the A/C/G, G/G/G and G/C/C haplotypes clades and increased occurrence of a vasodilatory syndrome after coronary artery bypass graft surgery with cardiopulmonary bypass, suggesting that these haploytpe clades may be associated with increased inflammation after cardiac surgery. Haplotype clades of IL-6 were not associated, however, with altered levels of IL-6, MCP-1, G-CSF, IL-8 or IL-lra at any time after cardiac surgery with CPB. Although our results appear to have reached genetic equipoise, these findings suggest genetic variations of IL-6 may play a role in the pathogenesis of the systemic inflammatory response syndrome in some instances. There have been a number of previous studies examining the association of IL-6 genetic variation and the inflammatory response with contradictory findings 2 1 ' 2 5 ' 2 1 ' 2 % . These earlier contradictory studies, in addition to our own, may reflect the pluripotent role of IL-6 in inflammation. IL-6 is a pleiotropic cytokine involved in the regulation of the acute inflammatory response and in the modulation of specific immune responses 4 7 . IL-6 functions as both a pro- and anti-inflammatory cytokine 4 1 . It is produced by macrophages, dendritic cells and Kupffer cells as an intermediate inflammatory cytokine at approximately six hours after an injurious stimulus 4 1 ' 4 8 . Later, during the proliferative stage of the immune response, IL-6 is produced by CD8+ cytotoxic 131 T-lymphocytes and CD4+ T H 2 cells and drives humoral immunity but suppresses cell-mediated immunity 4 9 . IL-6 has many transcription factor binding sites in its promoter to control its many functions 4 1 . The wide range of roles that IL-6 plays in inflammation may limit our ability to determine the effect that genetic variants of IL-6 may have on outcome from inflammatory disease. In addition to the complex roles that IL-6 plays in inflammation, and the potential for this to limit our understanding of the influence that genetic variants of IL-6 have on clinical outcome in inflammatory disease, the design of our study may also have limited our ability to validate our original association of IL-6 haplotype clades with 28-day mortality. We used a prospective cohort study design as it is difficult to find an appropriate control group for studies in critical illness. Healthy adults recruited as controls may have the same genetic susceptibility to infection and poor outcome as patients in our cohort but may not have had the same exposure to infectious microorganisms. Because most of our patients are middle-aged or older we often do not have access to family members to perform family-based studies. However, it may have been useful to compare the cohort of patients with SIRS to a population without SIRS to test for differences in the frequency of genetic variants. Additionally, in retrospect we should not have divided our cohort into derivation and validation cohorts temporally, with the earliest patients recruited being put into the derivation cohort and the later patients being put into the validation cohort. A more robust method would have been to recruit all patients and then divide them randomly into derivation and validation cohorts. This would have limited the effect of changes in practice and patients demographics over time. 132 Identification of important associations between genotype and clinical outcome is important for clinicians caring for critically ill patients for two key reasons. First, there is a substantial genetic contribution to the degree of activation of the systemic inflammatory response in patients with sepsis. Association of a gene polymorphism with outcome implicates pathways involving that gene in the pathophysiology of the clinical outcome - and thereby identifies potential therapeutic targets. Second, and more importantly, the benefit of specific therapy will depend on genotype. For example, anti-inflammatory therapy is a double-edged sword globally modulating both appropriate and inappropriate inflammatory responses. Thus, a number of recently developed immunomodulatory therapies will be beneficial in patients with specific genotypes and not beneficial in others. This view is consistent with the observations that indiscriminate application (by genotype) of many promising immunomodulatory therapies failed in phase III randomized controlled trials. In the very near future it will become technologically feasible to screen patients at the bedside in order to predict this response and design patient-tailored therapy using markers of clinical outcome such as the IL-6 C/C/G (or A/C/G), G/G/G and G/C/C haplotype clades. Already there are a number of high-throughput DNA chips being developed for rapid bedside genotyping of clinically relevant polymorphisms 5 0 ' 5 1 . Genetic information will be used by the clinician to define clinical subtypes of disease, and to stratify patients according to their risk of poor outcome. Genotype will be used to determine the optimal drug and its dose for treatment of an individual, while minimizing adverse effects. This knowledge will be instrumental in both preventative medicine and in treatment decisions. Analysis of haplotypes of SNPs within a gene is a powerful approach to study 9 0 association of common genetic variants with susceptibility to common diseases Haplotype analysis is more effective in determining association of genotype with phenotype 133 than is individual SNP analysis 2 9 ' 3 U . Haplotypes serve as markers of unidentified 9 0 polymorphisms that may be the cause of phenotypic variation . The phylogenetic history of haplotypes within a population may be determined and used to group haplotypes into clades using cladistic analysis 3 1 ' 5 2 " 5 5 . Cladistic analysis has two unique strengths. First, grouping haplotypes into clades decreases the degrees of freedom, thereby increasing the statistical power to associate genotype with phenotype Second, grouping haplotypes into clades facilitates identification of causal SNPs. Another practical strength of our approach is that a small number of "haplotype tag" SNPs (htSNPs) can be used to distinguish haplotype clades, eliminating the need to genotype all SNPs within a gene 3 4 , 5 6 . We selected as htSNPs the previously reported SNP -174G/C, as well as 1753C/G, and 2954G/C to define the 4 major haplotype clades within the IL-6 gene. 614G/A was chosen as an alternative to -174G/C as they were in complete linkage disequilibrium. This is the first report of a significant association of the IL-6 C/C/G, G/G/G and G/C/C haplotype clades with 28-day mortality in a derivation cohort of critically ill patients with SIRS. None of the htSNPs were individually associated with increased 28-day mortality, demonstrating the power and utility of haplotypes in association studies of genotype with phenotype. Although we did not find an association between these clades and outcome in a validation cohort of critically ill patients, the IL-6 A/C/G, G/G/G and G/C/C clades were associated with increased vasodilation in a cohort of cardiopulmonary bypass patients, while none of the individual htSNPs were. The SNP that is responsible for the possibly deleterious effect of the C/C/G (or A/C/G), G/G/G and G/C/C haplotype clades in patients with SIRS and in cardiopulmonary bypass patients is most likely a SNP that is common to these three clades, and is different in the G/C/G haplotype clade. Thus, by using a haplotype-based approach to associate genetic variation within the IL-6 gene with survival 134 of critically ill patients, we substantially narrowed down the number of SNPs in the IL-6 gene that may cause increased mortality in septic patients and increased vasodilation in cardiopulmonary bypass patients. Previous studies of the association of polymorphisms of IL-6 with outcomes of inflammatory diseases yielded greatly conflicting results. Fishman et al. demonstrated that a promoter region construct containing the C allele of the IL-6 -174G/C polymorphism transiently transfected into HeLa cells yielded lower expression of IL-6 than the -174G construct, both at baseline and after stimulation with LPS or IL-1 2 1 . The C allele also occurs less frequently in patients with systemic-onset juvenile chronic arthritis (S-JCA) 2 1 . The IL-6 -174C allele has been associated with increased survival rates of patients with ARDS 5 7 . GG homozygosity at -174G/C has also been associated with greater plasma concentrations of IL-6 and greater duration of ICU and hospital stay following coronary artery bypass graft surgery 5 8 . Patients with pneumococcal disease who are homozygous for -174GG are more likely to develop extrapulmonary pneumococcal infection 5 9 . In contrast, there is no association of IL-6 -174G/C genotype with systemic lupus erythematosus (SLE) 6 0 , incidence of sepsis 2 8 , or with the IL-6 response to endotoxin in whole blood from septic patients 6 1 ' 6 2 . Surprisingly, in direct contradiction to earlier studies, the G allele was associated with improved survival in septic patients in one study28, while Brull et al. found that -174CC homozygotes developed higher peak IL-6 levels at 6 hours post cardiopulmonary bypass 6 3 . Our current findings do not help clarify the inconsistent prior literature. We found no association between IL-6 -174G/C genotype and clinical outcome in any of our three cohorts, and we found no association between IL-6 -174G/C genotype and serum IL-6 levels. HoweVer, the genotype of this SNP contributes to defining the IL-6 C/C/G, G/G/G and G/C/C haplotype clades that were associated with strikingly poorer outcome in the 135 derivation cohort of critically ill patients, though not in the validation cohort of critically ill patients. -174G/C is also in linkage disequilibrium with 614G/A SNP that marks the A / C / G haplotype clade that along with the G/G/G and G/C/C haplotype clades was associated with poorer outcome in a cohort of cardiopulmonary bypass patients. To date, the role of IL-6 genetic variation in inflammatory disease remains unclear. IL-6 genetic variants may need to be examined in very large groups of homogenous patient cohorts in conjunction with many other genes in order to understand how variations of IL-6 interact with variants of other key immune system genes to control inflammation. Because IL-6 has such an important role in so many aspects of the immune response studying genetic variants of IL-6 in isolation may not provide us with a clear picture of the influence of genetic variation of IL-6 in the inflammatory response. Limitations of genetic association studies should be addressed. Firstly, we used large cohorts of critically ill patients (n=228 and n=431) and cardiopulmonary bypass patients (n=603) to reduce the risk of Type I error (finding a spurious association) compared to other studies of smaller sample size (n=20 -100) 6 4 ~ 6 7 . Secondly, ethnic heterogeneity within a study population can also lead to false positive associations between genotype and phenotype . Therefore we included only Caucasians in our cohorts of critically ill adults and our cohort of cardiopulmonary bypass patients to decrease the risk of positive associations due to population stratification. Third, we further limited Type I error by using several secondary measures of clinical phenotype in addition to the primary outcome variable in the cohorts of critically ill patients. Thus, our association of the C/C/G, G/G/G and G/C/C haplotype clades with increased 28-day mortality in the derivation cohort of critically ill patients was supported by the association of the C/C/G, G/G/G and G/C/C haplotype clades with worsened secondary outcome variables. Furthermore, the association of the C/C/G, G/G/G 136 and G/C/C haplotype clades with multiple worsened secondary outcomes was also consistent with the association of the C/C/G, G/G/G and G/C/C haplotype clades with increased need for organ life support. Finally, we attempted to validate the association of IL-6 haplotype clades with clinical outcome from inflammatory disease in multiple independent cohorts of patients. Statistical power is important in genetic association studies. By grouping haplotypes into clades to test for association with clinical outcomes, we decreased degrees of freedom 30 and increased statistical power to detect associations . We used strategies to maximize the accuracy of the inferred haplotypes of IL-6. We used unphased Caucasian genotypic data (from pga.mbt.washington.edu) to infer haplotypes of the IL-6 gene using PHASE software 3 3 ' 4 1 . We did not have access to the genotypes of family members of the genotyped individuals, and therefore could only infer probable haplotypes of the IL-6 gene using statistical methods 4 1 . Importantly, Stephens et al. have shown that PHASE is highly accurate. Therefore, reconstruction of haplotypes experimentally or by genotyping additional family members may not provide further information 4 1 . Furthermore, PHASE provides estimates of the certainty of haplotype assignment. In view of the fairly simple haplotype structure of the IL-6 gene, the PHASE algorithm determined with absolute certainty 98% of the phase calls in our cohort, and estimated the certainty associated with the remaining 2% of phase calls to be 0.99. A limitation of haplotype association studies is that association of the C/C/G (or A/C/G), G/G/G and G/C/C haplotype clades with poor outcome does not identify an injurious SNP but importantly narrows the possibilities 2 9 . The association of haplotypes of IL-6 with increased mortality or increased vasodilation does not provide us with a pathological 137 mechanism for the increase in mortality or vasodilation. Given that the injurious SNP has not been identified our results do not address how the C/C/G (or A/C/G), G/G/G and G/C/C haplotype clades affect regulation and/or levels of IL-6. Additionally, because we have not measured RNA expression or serum levels of IL-6 in our cohort of critically ill patients, we are not able to conclude what functional consequences haplotypes of IL-6 may have in sepsis and the inflammatory response. Furthermore, we did not detect significant differences in post-cardiac surgery IL-6 serum concentrations among IL-6 haplotype clades so the mechanism by which they increase post-surgical vasodilation in cardiac surgery patients remains unclear. In summary, we found that the G-to-C polymorphism at position -174 is not associated with outcome in patients who have SIRS or in cardiopulmonary bypass patients. In contrast, the C/C/G (or A/C/G), G/G/G and G/C/C haplotype clades were significantly associated with increased 28-day mortality, fewer days alive and free of SIRS and fewer days alive and free of multi-system organ dysfunction in a derivation cohort of critically ill Caucasians who had SIRS, although this association was not validated in a validation cohort of critically ill Caucasians who had SIRS. However, the IL-6 A/C/G, G/G/G and G/C/C haplotype clades were associated with increased occurrence of a vasodilatory syndrome following cardiac surgery, although not with post-cardiac surgery serum levels of IL-6. We have reached genetic equipoise in our association studies of IL-6 and outcome from inflammatory states. 138 4.5 Tables and Figures Table 4-1. Brussels organ dysfunction scoring criteria ORGANS Free of Organ Dysfunction Clinically Significant Organ Dysfunction Normal . Mild Moderate Severe Extreme Cardiovascular >90 Systolic BP (mmHg) Pulmonary >400 Pa02/F!02 (mmHg) Renal <1.5 Creatinine (mg/dL) Hepatic <1.2 Bilirubin (mg/dL) Hematologic >120 Platelets (xl05/mm3) Neurologic (Glasgow Score) 15 >90 400-301 1.5-1.9 1.2-1.9 120-81 14-13 >90 >90 + 300-201 Acute lung injury 2.0-3.4 2.0-5.9 80-51 12-10 ARDS 50-21 9-6 Brussels, March 12-14,1994 37 >90 + Responsive Unresponsive pH >7.3 pH >7.2 to fluid to fluid 200-101 <100 Severe ARDS 3.5-4.9 >5.0 6.0-11.9 >12 <20 <5 Round Table Conference on Clinical Trials for the Treatment of Sepsis 139 Table 4-2. Baseline characteristics of the derivation cohort of 228 critically ill adults with SIRS Haplotype Clade No. Age (MeaniSD) Gender (% Female) Diagnosis (%Surgical) APACHE II (MeaniSD) C/C/G 197 60.0±16.2 37.6 27.7 21.0±7.0 G/G/G 150 58.4±17.1 31.3 27.9 21.6±8.7 G/C/G 94 60.2±14.8 39.4 33.3 21.2±8.7 G/C/C 14 53.1±19.1 28.6 35.7 16.2±7.6 P NS NS NS NS NS A P A C H E II: Acute Physiology and Chronic Health Evaluation score. Measures the severity of acute disease by quantifying the degree of abnormality of multiple physiologic measures. 140 Table 4-3. Cox Proportional Hazard Analysis - Hazard ratios for mortality by IL-6 haplotype clade in derivation cohort of critically ill patients Covariate Hazard Ratio 95% CI P Female sex 0.62 0.38-1.02 0.06 Age 1.02 1.00-1.03 0.02 Surgical Diagnosis 0.90 0.55-1.47 0.70 2 copies of C/C/G, G/G/G, or G/C/C 1.83 1.10-3.04 0.02 141 Table 4-4. Genotype frequencies and allele frequencies for IL-6 htSNPs -174G/C, 1753C/G, and 2954G/C in a derivation cohort of 228 critically ill adults with SIRS Genotype Frequencies Allele Frequencies p * G/G G/C C/C G C G-174C .343 .446 .211 .566 .434 .963 C/C C/G G/G C G C1753G .454 .442 T04 .675 .325 .963 G/G G/C C/C G C G2954C .939 0.061 0 .969 .031 1.00 exact test of Guo and Thompson to test for Hardy-Weinberg equilibrium 142 Table 4-5. Haplotype clade frequencies in the derivation cohort versus the validation cohort of critically ill patients IL-6 Haplotype Clade C/C/G (A/C/G) N (%) G/G/G N (%) G/C/G N (%) G/C/C N (%) Derivation (n=228) 197(43.2) 150(32.9) 94 (20.6) 14(3.1) Validation (n=441) 321 (36.4) 310(35.1) 225 (25.5) 26 (2.9) 143 Table 4-6. Genotype frequencies and allele frequencies for IL-6 htSNPs 614G/A, 1753C/G, and 2954G/C in validation cohort of 441 critically ill adults with SIRS Genotype Frequencies Allele Frequencies p * G/G G/A A/A G A G614A .420 .433 .147 .636 .364 .675 c/c C/G G/G C G C1753G .447 .404 .150 .649 .351 .993 G/G G/C C/C G C G2954C .943 .054 .002 .971 .029 .999 144 Table 4-7. Baseline characteristics of validation cohort of 441 critically ill adults with SIRS Haplotype Clade No. Age (Mean±SD) Gender (% Female) Diagnosis (%Surgicaf) A P A C H E II (Mean±SD) A / C / G 321 57.4±15.8 37.7 24.3 24.5±9.2 G/G/G 310 56.2±17.7 38.4 21.9 23.7±8.9 G/C/G 225 55.0±16.4 41.3 17.8 22.7±8.9 G/C/C 26 55.0±15.8 42.3 15.4 23±8.9 P NS NS NS NS APACHE II: Acute Physiology and Chronic Health Evaluation score. Measures the severity of acute disease by quantifying the degree of abnormality of multiple physiologic measures. 145 Table 4-8. Cox Proportional Hazard Analysis - Hazard ratios for mortality by IL-6 haplotype clade in validation cohort of critically ill patients Covariate Hazard Ratio 95% CI P Female sex 0.91 0.66-1.23 0.56 Age 1.03 1.01-1.04 3xl0~6 Surgical Diagnosis 1.57 1.05-2.35 0.03 2 copies of A/C/G, G/G/G, or G/C/C 0.94 0.78-1.45 0.72 146 Table 4-9. Frequency of IL-6 haplotype clades in a cohort of 603 cardiopulmonary bypass patients SNP 614 1753 2954 htSNPs Haplotype Clade Frequency G C C 614G/1753C/2954C G/C/C 0.02 G C G 614G/1753C/2954G G/C/G 0.28 G G G 614G/1753G/2954G G/G/G 0.31 A C G 614A/1753C/2954G A / C / G 0.39 147 Table 4-10. Baseline demographics of cardiopulmonary bypass patients by IL-6 haplotype clade Demographics Carriers of G/C/G M on-carriers G/C/G of *p Age (mean±SD) 65±11 65±10 0.6 Sex (% Female) 22% 25% 0.5 BMI (meaniSD) 27.9±4.7 27.9±407 0.9 Diabetes mellitus (%) 24% 24% 0.9 Smoking (%) 34% 33% 0.9 A C E inhibitors (%) 63% 61% 0.6 p-blockers (%) 52% 47% 0.3 C a 2 + channel blockers (%) 61% 54% 0.07 Immunosuppression (%) 56% 57% 0.9 148 Table 4-11. Surgical details of cardiopulmonary bypass patients by IL-6 haplotype clade Detail Carriers of G/C/G Non-carriers of G/C/G *p Duration of CBP 1.7±0..8 1.7±0.9 0.6 Duration of cross-clamp 1.3±0.7 1.3±0.7 0.7 Duration of surgery 4.5±1.2 4.5±1.2 1.0 Intraoperative milrinone 15% 16% 0.9 Intraoperative aprotinin 14% 17% 0.3 Intraoperative amicar 49% 42% 0.1 149 Table 4-12. Post-operative serum concentrations of IL-6, MCP-1, G-CSF, IL-8 and IL-lra in cardiac surgery patients Cytokine Serum Concentration 3 hour's Serum Concentration 24 hours post-operatively (pg/mL) post-operatively (pg/mL) G/C/G Clade A/C/G, G/G/G, G/C/C clades P G/C/G Clade A/C/G, G/G/G, G/C/C clades P IL-6 836±2776 257±202 . 0.1. 114±95 104±95 0.6 MCP-1 1150±1404 1254±1985 0.8 242±203 190±147 0.1 G-CSF 437±734 327±301 0.3 199±111 171±113 0.2 IL-8 96±147 87±119 0.7 30±30 29±21 0.9 IL-lra 34747±22283 34179±24065 • 0.9 1279±3750 1884±6713 0.6 * mean concentration (pg/mL) ± SD 150 * h t S N P * h t S N P h t S N P h t S N P Figure 4-1. Haplotype structure of the IL-6 gene. Columns are polymorphic sites. Rows are haplotypes of IL-6 ordered by phylogenetic relationship. Yellow boxes are minor alleles and blue boxes are major alleles. -174G/C, 1753C/G, and 2954G/C were chosen as htSNPs. 614G/A was later chosen as an alternative htSNP to -174G/C; * -174G/C and 614G/A are in complete linkage disequilibrium (D'=0.96). 151 0.02 G / C / G Figure 4-2. Phylogenetic relationship of IL-6 haplotypes. IL-6 Haplotypes inferred by PHASE were sorted into 4 clades (circled) according to this evolutionary tree structure generated by MEGA2. Haplotype clades on this unrooted phylogenetic tree are labelled according to alleles they carry at the htSNPS -174G/C, 1753C/G and 2954G/C of the IL-6 gene. Distances along branches of the tree indicate percent difference between the sequences of related haplotypes. The length of the scale bar corresponds to a 2% difference in sequence. 152 A 28 Day Mortality G/C/G C/C/G G/G/G IL-6 Haplotype Clade G/C/C 28 Day Mortality *p=0.015 _ L _ G/C/G C/C/G, G/G/G or G/C/C IL-6 Haplotype Clades Figure 4-3. 28-day mortality rates by IL-6 clade. (A) The G/C/G haplotype clade was associated with decreased 28-day mortality (p=0.02). (B) Critically ill patients with 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had greater 28-day mortality than those with 1 or no copies (p=0.015). 153 28 Day Mortality -5 ~i Figure 4-4. Kaplan-Meier mortality analysis by IL-6 clade. Kaplan-Meier analysis showed that patients with 2 copies of the C/C/G, G/G/G or G/C/C haplotype clades had greater mortality rates than patients with at least 1 copy of the G/C/G haplotype clade for the entire 28-day observation period (p=0.GT3). 154 A DAF CVS Dysfunction p=0.01 CVS Dysfunction Vasopressors Clinical Outcome Variable ' DAF Respiratory Dysfunction p=0.04 ALI Ventilation Clinical Outcome Variable Figure 4-5. Multiple-system organ dysfunction in critically ill patients. Patients with 2 copies of the C/C/G, G/G/G or G/C/C clades had (A) fewer days alive and free of C V S dysfunction (p=0.006) and vasopressors (p=0.01), (B) fewer days alive and free of ALI (p=0.002) and ventilation (p=0.04), and (C) fewer days alive and free of renal dysfunction (p=0.005). 155 Figure 4-6. Kaplan-Meier mortality analysis by IL-6 clade in the validation cohort of critically ill patients. Kaplan-Meier analysis showed that there was no difference in 28-day mortality by IL -6 haplotype clade in the validation cohort of critically ill Caucasians (p=0.7). 157 Figure 4-7. A/C/G, G/G/G and G/C/C haplotype clades are associated with increased occurrence of a vasodilatory syndrome following cardiac surgery. A higher proportion of patients who carried 2. haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades had two consecutive measures of SVRi<1800dyne*sec/cm5/m2 after cardiac surgery than patients who carried one or more copies of the G/C/G clade (p<0.02). 158 80 n c o 1 or 2 copies G/C/G A/C/G, G/G/G and G/C/C IL6 Haplotype Clades Figure 4-8. A/C/G, G/G/G and G/C/C haplotype clades are associated with increased occurrence of a vasodilatory syndrome following cardiac surgery in a subgroup of patients on vasopressors. A higher proportion of patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades had two consecutive measures of SVRi<1800dyne*sec/cm5/m2 despite the use of vasopressors after cardiac surgery than patients who carried one or more copies of the G/C/G clade (p<0.002). 159 Figure 4-9. IL-6 haplotype clades are not associated with hours spent in the CSICU after cardiac surgery. Patients who carried 2 haplotypes from within the A/C/G, G/G/G or G/C/C haplotype clades did not spend more time in the CSICU after cardiac surgery than patients who carried 1 or more copies of the G/C/G clade (p>0.7). 160 4.6 References 1. Bone RC. The sepsis syndrome. Definition and general approach to management. Clin Chest Med 1996; 17:175-81. 2. Majetschak M , Flohe S, Obertacke U et al. Relation of a TNF gene polymorphism to severe sepsis in trauma patients. Ann Surg 1999;230:207-14. 3. Mira JP, Cariou A, Grail F et al. Association of TNF2, a TNF-alpha promoter polymorphism, with septic shock susceptibility and mortality: a multicenter study [see comments]. Jama 1999;282:561-8. 4. Read RC, Camp NJ, di Giovine FS et al. An interleukin-1 genotype is associated with fatal outcome of meningococcal disease. Journal of Infectious Diseases 2000;182:1557-60. 5. Sorensen TI, Nielsen GG, Andersen PK, Teasdale TW. Genetic and environmental influences on premature death in adult adoptees. NEngl J Med 1988;318:727-32. 6. Wan S, LeClerc JL, Vincent JL. Inflammatory response to cardiopulmonary bypass: mechanisms involved and possible therapeutic strategies. Chest 1997;112:676-92. 7. Kharazmi A, Andersen LW, Baek L, Valerius NH, Laub M , Rasmussen JP. Endotoxemia and enhanced generation of oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98:381-5. 8. Sawa Y, Shimazaki Y, Kadoba K et al. Attenuation of cardiopulmonary bypass-derived inflammatory reactions reduces myocardial reperfusion injury in cardiac operations. J Thorac Cardiovasc Surg 1996;111:29-35. 9. Downing SW, Edmunds LH, Jr. Release of vasoactive substances during cardiopulmonary bypass. Ann Thorac Surg 1992;54:1236-43. 10. Kirklin JK. Prospects for understanding and eliminating the deleterious effects of cardiopulmonary bypass. Ann Thorac Surg 1991 ;51:529-31. 11. Wei M, Kuukasjarvi P, Laurikka J et al. Relation of cytokines to vasodilation after coronary artery bypass grafting. World J Surg 2003;27:1093-8. 12. Edmunds LH, Jr. Blood-surface interactions during cardiopulmonary bypass. J Card Surg 1993;8:404-10. 13. Colman RW, Scott CF, Schmaier A H , Wachtfogel YT, Pixley RA, Edmunds LH, Jr. Initiation of blood coagulation at artificial surfaces. Ann N Y Acad Sci 1987;516:253-67. 14. Chenoweth DE, Cooper SW, Hugh TE, Stewart RW, Blackstone EH, Kirklin JW. Complement activation during cardiopulmonary bypass: evidence for generation of C3a and C5a anaphylatoxins. NEngl J Med 1981 ;304:497-503. 15. Tatoulis J, Rice S, Davis P, Goldblatt JC, Marasco S. Patterns of postoperative systemic vascular resistance in a randomized trial of conventional on-pump versus off-pump coronary artery bypass graft surgery. Ann Thorac Surg 2006;82:1436-44. 16. Qing M , Woltje M , Schumacher K et al. The use of moderate hypothermia during cardiac surgery is associated with repression of tumour necrosis factor-alpha via inhibition of activating protein-1: an experimental study. Crit Care 2006;10:R57. 17. Hack CE, De Groot ER, Felt-Bersma RJ et al. Increased plasma levels of interleukin-6 in sepsis. Blood 1989;74:1704-10. 18. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T. The complex pattern of cytokines in serum from patients with meningococcal septic shock. Association between interleukin 6, interleukin 1, and fatal outcome. J Exp Med 1989;169:333-8. 161 19. Martin C, Boisson C, Haccoun M , Thomachot L, Mege JL. Patterns of cytokine evolution (tumor necrosis factor-alpha and interleukin-6) after septic shock, hemorrhagic shock, and severe trauma. Crit Care Med 1997;25:1813-9. 20. Meduri GU, Headley S, Kohler G et al. Persistent elevation of inflammatory cytokines predicts a poor outcome in ARDS. Plasma IL-1 beta and IL-6 levels are consistent and efficient predictors of outcome over time. Chest 1995;107:1062-73. : 21. Roth-Isigkeit A, Hasselbach L, Ocklitz E et al. Inter-individual differences in cytokine release in patients undergoing cardiac surgery with cardiopulmonary bypass. Clin Exp Immunol 2001 ;125:80-8. 22. Nathan N, Denizot Y, Cornu E, Jauberteau MO, Chauvreau C, Feiss P. Cytokine and lipid mediator blood concentrations after coronary artery surgery. Anesth Analg 1997;85:1240-6. 23. Tsang G M , Allen S, Pagano D, Wong C, Graham TR, Bonser RS. von Willebrand factor and urinary albumin excretion are possible indicators of endothelial dysfunction in cardiopulmonary bypass. Eur J Cardiothorac Surg 1998;13:385-91. 24. Holmes JHt, Connolly NC, Paull D L et al. Magnitude of the inflammatory response to cardiopulmonary bypass and its relation to adverse clinical outcomes. Inflamm Res 2002;51:579-86. 25. Gaudino M , Andreotti F, Zamparelli R et al. The -174G/C interleukin-6 polymorphism influences postoperative interleukin-6 levels and postoperative atrial fibrillation. Is atrial fibrillation an inflammatory complication? Circulation 2003; 108 Suppl 1:11195-9. 26. Cremer J, Martin M , Redl H et al. Systemic inflammatory response syndrome after cardiac operations. Ann Thorac Surg 1996;61:1714-20. - 27. Fishman D, Faulds G, Jeffery R et al. The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemic-onset juvenile chronic arthritis. J Clin Invest 1998;102:1369-76. 28. Schluter B, Raufhake C, Erren M et al. Effect of the interleukin-6 promoter polymorphism (-174 G/C) on the incidence and outcome of sepsis. Crit Care Med 2002;30:32-7. 29. Akey J, Jin L, Xiong M . Haplotypes vs single marker linkage disequilibrium tests: what do we gain? Eur J Hum Genet 2001;9:291-300: 30. Zhang K, Calabrese P, Nordborg M , Sun F. Haplotype block structure and its applications to association studies: power and study designs. Am J Hum Genet 2002;71:1386-94. 31. Templeton AR, Weiss K M , Nickerson DA, Boerwinkle E, Sing CF. Cladistic structure within the human lipoprotein lipase gene and its implications for phenotypic association studies. Genetics 2000;156:1259-75. 32. Keavney B. Genetic association studies in complex diseases. J Hum Hypertens 2000;14:361-7. 33. Rieder M CD, Chung M-W. Homo sapiens interleukin 6 (IL-6) gene, complete eds. ACCESSION AF372214. 2002 vol: SeattleSNPs. NHLBI Program for Genomic Applications, UW-FHCRC; 2001. 34. Johnson GC, Esposito L, Barratt BJ et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001 ;29:233-7. 35. Gabriel SB, Schaffner SF, Nguyen H et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225-9. 36. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet 2001;2:91-9. 162 37. Sibbald WJ, Vincent JL. Round table conference on clinical trials for the treatment of sepsis. Brussels, March 12-14, 1994. Intensive Care Med 1995 ;21:184-9. 38. Bernard GR, Wheeler AP, Arons M M et al. A trial of antioxidants N-acetylcysteine and procysteine in ARDS. The Antioxidant in ARDS Study Group. Chest 1997;112:164-72. 39. Knaus WA, Wagner DP, Draper EA et al. The A P A C H E III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100:1619-36. 40. Kristof AS, Magder S. Low systemic vascular resistance state in patients undergoing cardiopulmonary bypass. Crit Care Med 1999;27:1121-7. 41. Stephens M , Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001 ;68:978-89. 42. Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 2001 ;17:1244-5. 43. Livak KJ. Allelic discrimination using fluorogenic probes and the 5' nuclease assay. Genet Anal 1999;14:143-9. 44. Pascual M , Nieto A, Mataran L, Balsa A, Pascual-Salcedo D, Martin J. IL-6 promoter polymorphisms in rheumatoid arthritis. Genes Immun 2000;1:338-40. 45. Bray MS, Boerwinkle E, Doris PA. High-throughput multiplex SNP genotyping with MALDI-TOF mass spectrometry: practice, problems and promise. Hum Mutat 2001;17:296-304. 46. Guo SW, Thompson EA. Performing the exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics 1992;48:361-72. 47. Song M , Kellum JA. Interleukin-6. Crit Care Med 2005;33:S463-5. 48. Keel M, Trentz O. Pathophysiology of polytrauma. Injury 2005;36:691-709. 49. Ayala A, Chung CS, Grutkoski PS, Song GY. Mechanisms of immune resolution. Crit Care M?J2003;31:S558-71. 50. Zhong X B , Reynolds R, Kidd JR et al. Single-nucleotide polymorphism genotyping on optical thin-film biosensor chips. Proc Natl Acad Sci USA 2003;100:11159-64. 51. Nakai K, Habano W, Fujita T et al. Highly multiplexed genotyping of coronary artery disease-associated SNPs using MALDI-TOF mass spectrometry. Hum Mutat 2002;20:133-8. 52. Templeton AR, Crandall KA, Sing CF. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 1992;132:619-33. 53. Templeton AR. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping or DNA sequencing. V. Analysis of case/control sampling designs: Alzheimer's disease and the apoprotein E locus. Genetics 1995;140:403-9. 54. Long AD, Langley CH. The power of association studies to detect the contribution of candidate genetic loci to variation in complex traits. Genome Res 1999;9:720-31. 55. Braaten O, Rodningen OK, Nordal I, Leren TP. The genetic algorithm applied to haplotype data at the L D L receptor locus. Comput Methods Programs Biomed 2000;61:1-9. 56. Weiss K M , Clark AG. Linkage disequilibrium and the mapping of complex human traits. Trends Genet 2002; 18:19'-24. 163 57. Marshall RP, Webb S, Hill MR, Humphries SE, Laurent GJ. Genetic polymorphisms associated with susceptibility and outcome in ARDS. Chest 2002;121:68S-69S. 58. Burzotta F, lacoviello L, Di Castelnuovo A et al. Relation of the -174 G/C polymorphism of interleukin-6 to interleukin-6 plasma levels and to length of hospitalization after surgical coronary revascularization. Am J Cardiol 2001;88:1125-8. 59. Schaaf B, Rupp J, Muller-Steinhardt M et al. The interleukin-6 -174 promoter polymorphism is associated with extrapulmonary bacterial dissemination in Streptococcus pneumoniae infection. Cytokine 2005;31:324-8. 60. Linker-Israeli M , Wallace DJ, Prehn J et al. Association of IL-6 gene alleles with systemic lupus erythematosus (SLE) and with elevated IL-6 expression. Genes Immun 1999;1:45-52. 61. Heesen M , Bloemeke B, Heussen N, Kunz D. Can the interleukin-6 response to endotoxin be predicted? Studies of the influence of a promoter polymorphism of the interleukin-6 gene, gender, the density of the endotoxin receptor CD 14, and inflammatory cytokines.PG. Crit Care Med 2002;30:664-9. 62. Heesen M , Obertacke U, Schade FU, Bloemeke B, Majetschak M . The interleukin-6 G(-174)C polymorphism and the ex vivo interleukin-6 response to endotoxin in severely injured blunt trauma patients. Eur Cytokine Netw 2002; 13:72-7. 63. Brull DJ, Montgomery HE, Sanders J et al. Interleukin-6 gene -174g>c and -572g>c promoter polymorphisms are strong predictors of plasma interleukin-6 levels after coronary artery bypass surgery. Arterioscler Thromb Vase Biol 2001 ;21 T458-63. 64. Arnalich F, Lopez-Maderuelo D, Codoceo R et al. Interleukin-1 receptor antagonist gene polymorphism and mortality in patients with severe sepsis. Clin Exp Immunol 2002;127:331-6. 65. Gibot S, Cariou A, Drouet L, Rossignol M , Ripoll L. Association between a genomic polymorphism within the CD 14 locus and septic shock susceptibility and mortality rate. Crit Care Med 2002;30:969-73. 66. Lowe PR, Galley HF, Abdel-Fattah A, Webster NR. Influence of interleukin-10 polymorphisms on interleukin-10 expression and survival in critically ill patients. Crit Care Med 2003;31:34-8. 67. Schaaf B M , Boehmke F, Esnaashari H et al. Pneumococcal septic shock is associated with the interleukin-10 - 1082 gene promoter polymorphism. Am J Respir Crit Care Med 2003;168:476-80. 164 C H A P T E R 5: CONCLUSIONS 5.1 Summary of Main Findings 1 have found that allelic variants of key innate immunity (CD14, MBL, TLR2 and IRAK4) and inflammatory (IL-6) genes may be predictive of risk of infection (prevalence of positive microbiological cultures at admission to ICU) and outcome (mortality, organ dysfunction) in critically ill adults who have systemic inflammatory response syndrome. Specifically, I tested the previously described C-159T CD14, the X / Y and B, C, and D polymorphisms of MBL, and the novel haplotype-tag SNP T-16933A of TLR2 for association to prevalence of. positive microbiological cultures, sepsis and septic shock at admission to the ICU, and 28-day survival with genotype in a derivation cohort of 252 critically ill Caucasians with SIRS. I found that in the derivation cohort of critically ill patients CD14 -159TT was associated with increased rates of positive bacterial cultures on admission to the ICU, and that CD14 -159CT and -159TT were associated with increased prevalence of Gram-negative infections. In a larger validation cohort of critically ill patients I did not find that CD14 C-159T genotype was associated with prevalence of Gram-negative infections, although CD14 -159TT was associated, with significantly decreased 28-day survival. This polymorphism may increase risk of death in critically ill patients who have SIRS, possibly as a result of increased risk to Gram-negative nosocomial infections later in the course of their ICU stay. M B L haplotype pairs XO/O and O/O were also associated with increased prevalence of positive bacterial cultures but not with a specific organism class. TLR2 -16933AA was associated with increased incidence of sepsis and with Gram-positive bacteria. 165 I next investigated the association of haplotype clades of IRAK4, a central innate immunity signaling molecule, with susceptibility to and outcome from infection in a large cohort of critically ill adults. I found that the 23338C/24472T/29429A (C/T/A) haplotype clade tagged by the A allele of the htSNP G29429A (Ala428Thr) was associated with increased prevalence of Gram-positive infection at admission to ICU, and with decreased in vitro B-lymphocyte and fibroblast immune response to toll-like receptor ligands. IRAK4 Ala428Thr may alter the ability of Toll-like receptors to signal through IRAK4. Finally, I examined the association of haplotype clades of IL-6 with mortality and organ dysfunction in a derivation and validation cohort of critically ill patients. I also examined the association of these same haplotype clades with occurrence of a vasodilatory syndrome in cardiopulmonary bypass patients and with post-surgical cytokine levels in a subgroup of the same cardiopulmonary bypass patients. In the derivation cohort of critically ill patients I found that three haplotype clades of IL-6 were associated with increased mortality and more organ dysfunction. Forty percent of patients carrying 2 copies of the IL-6 haplotype clades defined by -174C/1753C/2954G (C/C/G clade), G/G/G and G/C/C died during the 28 day observation period, while patients who carried only one or none of these clades had a mortality rate of only 26.0% (p=0.013). Haplotype clades C/C/G, G/G/G and G/C/C were also associated with significantly less DAF of organ dysfunction and organ system support (p<0.05). These results were not corroborated in the validation cohort of critically ill patients. Our inability to repeat the association in the validation cohort may be a result of the way we divided patients into derivation and validation cohorts. We divided the cohorts temporally; the earliest set of patients recruited into the study became the derivation cohort, while later patients recruited into the study became the validation cohort. As the study extended from 2001 until late 2005 there may have been changes in standard of care 166 and hospital policy that affected the characteristics and outcome of patients entering the ICU at St. Paul's Hospital. If there were undetected differences between the cohorts these may have limited our ability to repeat results in the validation cohort. In the future a better study design would be to recruit all patients and later randomly split them into derivation and validation cohorts. This would enable us to eliminate the effects of changes in practice over time on our ablility to detect associations between genotype and clinical outcome. Although we did not detect a statistical difference betweeen patients of different IL-6 haplotype clades in the validation cohort, it does not necessarily mean there is no effect of IL-6 haplotype clades on clinical outcome in critical illness. It may simply not have been possible to detect a statistical difference because of issues in study design. In the cohort of cardiopulmonary bypass patients, however, the A / C / G (equivalent to C/C/G), G/G/G and G/C/C haplotype clades were associated with increased occurrence of a vasodilatory syndrome, which occurs as a result of inflammation caused by cardiopulmonary bypass. Haplotype clades of IL-6 were not, however, associated with cardiac surgery patients' post-surgical serum concentrations of inflammatory cytokines. Given previous conflicting literature regarding the association of genetic variants of IL-6 with inflammatory disease 1 4 and my own conflicting results, it would appear that we have reached a genetic equipoise concerning the role of IL-6 genetic variants in determining outcome from inflammatory disease. Genetic variants of IL-6 may influence outcome from inflammatory disease in some instances, but their influence may depend greatly on interaction with variants of many other genes. My findings show that allelic variants in the key innate immunity and inflammatory genes CD14, MBL, TLR2, IRAK4 and IL-6 may explain a component of variation in 167 individuals' responses to infection. Interestingly, SNPs in innate immunity genes appear to be associated with susceptibility to the organisms for which they are specific. This includes IRAK4, which is redundant for signaling recognition of Gram-negative bacteria 5" 1 0 , but not Gram-positive bacteria, and was associated with increased susceptibility to Gram-positive bacteria. Variants of CD 14 and other innate immunity genes were associated with increased incidence of bacterial cultures and infection. Only CD 14, however, appeared to be associated with increased late mortality, perhaps from nosocomial infections. Genetic variants of innate immunity receptors and signaling molecules may alter recognition and clearance of bacteria leading to an increased bacterial load and increased incidence of severe infections. My results regarding IL-6 were less conclusive, but this may simply reflect the pluripotent role of IL-6 in immunity and inflammation. 5.2 Strengths and Weaknesses The same issues of sample size, power, false positive and negative results, lack of reproducibility, ethnic heterogeneity and biologic plausibility that affect all genetic association studies affected my genetic association studies. Candidate Gene Approach I used a candidate gene approach to select genes for my association studies. There are potentially thousands of genes to test for association to outcome of a disease. By selecting candidate genes that are part of physiological processes known to influence the disease, we can narrow down our search for genetic variability that is associated with outcome u . The majority of patients in the ICU have a systemic inflammatory response as a result of infection, thus we believed that genetic variability in innate immunity receptors, which are the first to recognize invading pathogens, and innate immunity signaling 168 molecules, which transmit the signal from the receptor, would be important in determining inter-individual variability in the response to infection. The fact that innate immune receptors and signaling molecules were found to be associated with the risk of positive microbiological cultures and infection lends validity to my candidate gene approach. IL-6 is known to have a central role in the systemic inflammatory response 1 2 ' 1 3 , 1 4 and so it was logical to assume that genetic variability in the IL-6 gene would influence the magnitude and outcome of the inflammatory response. The risk of looking for candidate genes among pathways we already know is that we may miss key genes because of ignorance of other biological systems involved Sample Size, Power and False Negative Genetic Association Studies In all association studies, whether they are a case-control, cohort or family-based study design, adequate sample size is a crucial issue to diminish the risk of false negative results 1 5 ' 1 6 . Although the use of a case-control study design increases sample size, I used a prospective cohort study design to minimize difficulties in development of a proper control group 1 6 . The use of an improper control group can lead to false genetic associations because of differences of allelic frequencies between cases and controls due to evolutionary or migratory history, assortative mating, or gender and age differences l 6 , 1 7 . Regardless of study design, many studies are underpowered to eliminate false negative results. In addition, small sample size and underpowered studies are a major cause of lack of reproducibility of genetic association studies l 8 ' ' 9 . At the time of publication, my genetic association studies in a derivation cohort of more than 250 patients were the largest published in critical care. In my study of association of IL-6 haplotype clades with mortality in 263 critically ill patients I had greater than 80% power to detect a 10% difference (a<0.05) in mortality between patients who carried the G/C/G clade and those that did not. In the future, cohorts consisting 169 of thousands of patients may be necessary to tease out the small contributions of multiple genetic variants to complex disease. Multiple Testing and False Positive Genetic Association Studies In my genetic association studies more than one variable was examined simultaneously, and multiple SNPs were tested for association to multiple outcome measures. While multiple testing may result in false positive genetic association studies and contribute to lack of reproducibility of studies ' 6 , however it is necessary to generate hypotheses. Standard corrections for multiple testing such as the Bonferroni adjustment may be overly conservative in genetic association studies where many of the variables are not independent. Genes work in pathways and, therefore, are not independent. Some genes in the same pathway may even be physically linked on the same chromosome. A number of our outcome measures are similarly not independent. A patient who has more days alive and free of renal dysfunction is more likely to also have more days alive and free of renal support. As an alternative to overly conservative standard corrections for multiple testing I used the strategies of a validation cohort and biological experiments to increase the likelihood of identifying a true positive result. My initial association studies were derivation studies, I later attempted to validate my results in larger validation cohorts to correct for multiple comparisons and a somewhat heterogeneous patient cohort, as well as to ensure proper power to rule out false negative associations. I also attempted to lend biological plausibility to my positive genetic associations by measuring intermediate phenotypes and examining biological mechanism. A good study design for future genetic association studies would be one that includes both a hypothesis-generating derivation cohort and a hypothesis-testing validation cohort, in addition to molecular experiments to prove mechanism. 170 Ethnic Heterogeneity In order to decrease the potential confounding influence of population admixture secondary to ethnic diversity on associations between genotype and phenotype I only tested for associations of genotype with outcome from SIRS in Caucasians. This minimizes the risk of false associations. In the future, however, it will be necessary to test my associations of CD14, MBL, TLR2, IRAK4 and IL-6 polymorphisms with outcome from SIRS in other separate ethnic groups of patients. Haplotype Clade Approach to Association Studies As discussed in the preceding chapters, I used a haplotype clade approach to select haplotype tag SNPs for my association studies of TLR2, IRAK4 and IL-6. I believe that a haplotype-based approach to association studies is more useful than a single SNP approach because single SNP approaches for genetic association studies only work well when the examined SNP is causal or is in significant linkage disequilibrium with the causal SNP. In contrast, haplotypes serve as markers for all measured and unmeasured alleles within the haplotypes 2 0 A haplotype-based approach can narrow the search for causal SNPs and 20 2 1 * haplotypes can serve as predictors of disease severity ' . Multiple haplotypes exist for all studied genes, thus grouping haplotypes into haplotype clades (evolutionarily related groups of haplotypes) increases the statistical power to associate genotype with phenotype 2 2 . I believe the use of a haplotype-based approach allowed me to show that the IL-6 -174G/C SNP is not associated with mortality in critically ill patients, it is simply a marker of a haplotype clade in which the causal SNP may lie. While I have not found the causal SNP, it can be found by more closely examining common SNPs that lie in the IL-6 haplotype clades associated with increased mortality. Interestingly, the C/T/A clade of 1RAK4 is marked by a non-synonymous SNP, Ala428Thr. If I had used a functional SNP approach I most likely 171 would have selected this SNP as a candidate for association. However, Ala428Thr is in linkage disequilibrium with many other SNPs in the 1RAK4 gene, and if Ala428Thr had not been detected by the SeattleSNPs Program for Genomic Applications, it would still be possible to detect an association between the C/T/A haplotype clade and outcome. Plausible functional candidate SNPs within the haplotype clade could have then been found and tested for association to incidence of Gram-positive infection, which would have led to Ala428Thr. Biological Plausibility and Mechanisms of Variability A major limitation of our haplotype association studies is that association of haplotypes or haplotype clades with outcome does not identify an injurious SNP or provide a mechanism through which genetic variation alters phenotype 1 5 ' 2 0 . For example, the association of haplotypes of IL-6 with increased mortality does not provide a pathological mechanism to explain the increase in mortality. Even when a SNP associated with outcome is in a promoter region or exon and is found to alter gene expression or protein function in vitro, such as IRAK4 Ala428Thr, this does not prove its function in vivo. For these reasons it is important to measure intermediate phenotypes such as tissue RNA expression or serum protein levels to determine if a SNP alters gene transcription, translation or the protein stability of the gene product15. Fresh blood samples from the cohorts of critically ill patients were not available. As such, it was not possible to measure tissue RNA expression or protein levels of CD14, MBL, TLR2, IRAK4 or IL-6 in these patients. While RNA and protein expression of CD14 and M B L in patients of different CD14 -159C/T genotypes and M B L haplotypes were not measured in our studies, many previous reports show a marked difference in RNA and protein levels associated with these genetic variations 2 3 ~ 2 8 . TLR2 -16933T/A has not been • 172 previously tested for association to disease, and its function, if any, is unknown. It is not know if TLR2 -16933T/A is associated with a change in RNA or protein expression, and if it is I do not know if it is the causal SNP or whether it is simply in linkage with a causal SNP. Although tissue RNA expression or ex vivo intracellular concentrations of IRAK4 were not measured, biological plausibility of the IRAK4 genotype-clinical phenotype association has been demonstrated through in vitro experiments with B-lymphocytes and fibroblasts. Although these are not the ideal cells in which to be measuring differences in innate immune response, an association between the C/T/A haplotype clade and decreased immune activation these cells was found. These results suggest that the Ala428Thr polymorphism inhibits toll-like receptor signaling and lend biologic plausibility to the genotype-clinical phenotype association. No association was found between haplotype clades of IL-6 and post-surgical serum concentrations of cytokines in cardiopulmonary bypass patients. However, it was not possible to measure IL-6 in the serum of critically ill patients so it is not known how genetic variants of IL-6 affect serum cytokine concentrations. The C allele of IL-6 -174G/C has been associated with decreased levels of IL-6 expression, however this was in a luciferase reporter vector transiently transfected into HeLa cells, and does not necessarily reflect in vivo conditions •'. The same study found that in healthy controls the C allele was found to be associated with significantly lower levels of plasma IL-6 but did not examine patients with an infection '. These findings suggest that genetic variation in the IL-6 gene may cause alterations in RNA and protein expression, but do not definitively provide a mechanism for a difference in outcomes in patients with different IL-6 haplotype clades. 173 5.3 Significance and Relevance to Current Knowledge While this work is simply the beginning to our understanding of how genetic variability influences inter-individual differences in the response to injury and infection, my findings suggest that there are important associations between genotype of inflammatory and innate immunity genes and outcome from the systemic inflammatory response. Genetic association studies are an important first step on the road to personalized medicine. As more economical and rapid high-throughput genotyping methods are developed in parallel to more robust statistical analysis of genetic data, the idea of designing patient treatment based on an individual's genotype will become a reality. Every association we find between genotype and phenotype takes us a bit closer to this reality. My findings confirm or clarify previous association studies of polymorphisms of the innate immune receptors CD14, TLR2 and MBL and outcome from critical illness. CD14 -159C/T and the high and low producing haplotypes of M B L are well characterized 919^ 97 9Q polymorphisms ' ' . There have been a number of contradictory reports examining the association of the CD14 -159C/T polymorphism with risk of developing, and outcome from, severe sepsis and septic shock 3 0 " 3 6 . The CD14 C-159T polymorphism does not appear to be associated with risk of septic shock or mortality in Asian populations 3 5 ' 3 6 , and there have been conflicting reports in mixed ethnicity and Caucasian patient samples 2 3 ' 3 0 ~ 3 3 . The derivation and validation cohorts in which 1 tested CD14 C-159T for association to risk of and outcome from sepsis were limited to Caucasians, and are both larger than any association study of CD14 -159C/T with outcome in critical illness to date. Smaller negative genetic association studies, including my association study in the derivation cohort, may have been underpowered to detect a difference in mortality by CD14 C-159T genotype. 174 M B L haplotypes that have been shown to produce low levels of serum M B L have been associated with increased prevalence of sepsis and septic shock, and increased mortality rates in ICU patients 2 S ' 2 1 ' 2 9 . While we did not find that these haplotypes were associated with risk of septic shock or mortality, we did find that low M B L haplotypes were associated with increased rates of positive bacterial cultures on admission to the ICU, which are a marker of increased infection and sepsis. TLR2 -16933T/A has not been previously examined for association with outcome in critically ill patients. However, other SNPs in TLR2 have been associated with increased risk of Gram-positive infections, which is consistent with our finding of an association between TLR2 -16933AA and increased prevalence of Gram-positive cultures at admission to ICU 3 7 ' 3 8 . No other group has reported published results describing the association of IRAK4 genetic variants with risk of infection, sepsis or outcome from sepsis. It is interesting to note however, that pediatric patients with rare non-functional mutations of IRAK4 experience recurrent Gram-positive infections 9 ' 3 9 - 4 5 . This is consistent with my finding of an association between the IRAK4 C/T/A haplotype clade and risk of Gram-positive cultures at admission to ICU. Additionally, immune cells from these pediatric patients are hyporesponsive to toll-like receptor ligands 4 1 ' 4 4 , as are immune cells carrying the IRAK4 C/T/A clade or the Ala428Thr polymorphism that marks the C/T/A clade. My examination of the association of haplotype clades of IL-6 with outcome from inflammatory disease does not help to clarify previous conflicting genetic association studies regarding the association of the IL-6 -174 G/C polymorphism with outcome from SIRS, 175 sepsis and infection 2 ' 4 6 " 4 8 . I obtained inconsistent results in the derivation, validation and cardiopulmonary bypass patient cohorts. The inconsistent results in both the literature and in my own experiments may be a reflection of the pluripotent role of IL-6 in inflammation and immune modulation as discussed in Chapter 4. The wide range of roles that IL-6 plays in the immune response, and its multiple sites of regulation may limit our ability to determine the effects that genetic variants of IL-6 may have on outcome from the systemic inflammatory response. 5.4 Future Research My examination of the association of polymorphisms in CD14, MBL, TLR2, IRAK4 and IL-6 with outcome in critically ill patients with SIRS is only the beginning of a full understanding of the role genetic variants play in determining individuals' susceptibility to and ability to combat infection. Many questions regarding how these and other genetic variants influence patient outcome remain. Validation Studies One of the most important steps to take in a genetic association study is to confirm initial results in a validation cohort. I have attempted to confirm the association of CD14 and IL-6 genetic variants with clinical outcome measures in separate validation cohorts, but the results have not been entirely conclusive. It will be important in the future to re-examine my findings of an association between IL6 haplotype clades and mortality, and associations of polymorphisms in innate immune receptors and signaling molecules with prevalence of positive microbiological cultures and infection, in a very large validation cohort of critically ill Caucasians with SIRS. Furthermore, it is important that we test genotype-phenotype associations in cohorts of patients from other hospitals and other countries so we can be sure 176 that the genotype-phenotype associations are widely applicable. Our group is currently recruiting a new cohort of critically ill patients with SIRS from the ICU at St. Paul's Hospital. I believe that we must perform future genetic association studies in thousands of patients to be sure we have adequate power to detect even small differences in phenotype. By validating our findings in a very large cohort we will ensure that we have adequate power to limit the chance of false negative associations if we cannot replicate our previous findings. Recruiting a very large cohort of critically ill patients with SIRS will also allow us to refine our clinical phenotypes in order to define a less heterogeneous patient population. Since our original genetic association studies of 250 critically ill patients we have evaluated 1026 critically ill Caucasian patients from our ICU, 77 % of whom had sepsis, to better define our patient population/cohort. Of the 1026 patients, 794 (77%) had sepsis, and 567 (71%) of patients with sepsis were admitted to ICU for non-surgical diagnoses. We are currently recruiting an average of 5 new ICU patients per week into our study. With this large cohort we will be able to look at specific subgroups of patients, for example patients with community-acquired pneumonia, male patients versus female patients or patients infected with different organisms. We are recruiting patients of all ethnicities as well; as the cohort grows we will have adequate numbers of patients of different ethnicities to test for associations of genetic polymorphisms with outcome in these groups. A second group of patients in which associations of polymorphisms of inflammatory and innate immunity genes with outcome from SIRS can be validated is coronary artery bypass graft (CABG) patients undergoing surgery with cardiopulmonary bypass (CPB), as I did for IL-6. As discussed in Chapter 4, htSNPs found to be associated with clinical outcome in critically ill adults will be genotyped in a separate cohort of 1000 patients undergoing 177 elective coronary artery bypass surgery. Our primary outcome variable has been the occurrence of an indexed low systemic vascular resistance (SVRi) less than 1800 dyne*sec/cnr/m2 at 2 consecutive time points post-cardiopulmonary bypass 4 9 . In the future we can test for the association of genetic variants of CD 14, TLR2, M B L and IRAK4 with occurrence of this vasodilatory syndrome. We currently have DNA and clinical phenotype data for over 1200 cardiac surgery patients. Additionally, our group is currently recruiting an average of 10 new patients per week from whom we collect blood pre-operatively, and 0, 4, 12 and 24 hours post-operatively and in whom we are measuring neurocognitive dysfunction after surgery. It is believed that neurocognitive dysfunction after cardiopulmonary bypass surgery results from cerebral microemboli, global cerebral hypoperfusion, inflammation, and genetic susceptibility 5 0 " 5 2 . Neurocognitive function can be measured before and after surgery and any decline quantified. It may be a more sensitive clinical phenotype than occurrence of a vasodilatory syndrome to use in genetic association studies of inflammatory and innate immunity genes and outcome after cardiac surgery. I expect that haplotype clades and polymorphisms of innate immunity and inflammatory cytokine genes that are found to be associated with clinical outcomes in patients with SIRS will also be clinically relevant in patients undergoing cardiovascular surgery with cardiopulmonary bypass. Intermediate Phenotypes It is important to associate clinically relevant genetic variants with an in vivo change in protein concentration or other intermediate phenotype. However, intermediate phenotypes are difficult to measure in critically ill patients. Critically ill patients in the ICU are a very heterogeneous group. ICU patients develop SIRS as a result of a multitude of insults, infectious organisms, and sources of infection. Many ICU patients may have been sick for an extended period of time and so it is not known exactly when their inflammatory response 178 began. In contrast, C A B G patients are relatively healthy, and their inflammatory response is a direct result of their cardiac surgery with cardiopulmonary bypass. We know when the inflammatory stimulus occurred and what it was, and we can measure inflammatory proteins in their blood at precise time-points before and after their surgery. In this way, we can precisely measure the evolution of the systemic inflammatory response and compare this to cardiac surgery patients' genotypes. Thus, if we find an important association between genotype and outcome of critically ill patients with SIRS, we can test for an association of this important genotype with the production of cytokines and other important inflammatory mediators during the inflammatory response following cardiac surgery. While I did not find an association between haplotype clades of IL-6 and post-cardiac surgery serum concentrations of cytokines, we are currently recruiting more cardiac surgery patients and collecting blood at more time points and so we may find an association between post-cardiac surgery serum cytokine concentrations and genetic variants of future candidate genes. In addition to measuring the evolution of the inflammatory response in cardiac surgery patients, we will try in the future to measure intermediate phenotypes in our cohort of critically ill patients from the ICU. As discussed above, patients in the ICU are a very heterogeneous group, but as our cohort grows larger we will try to define specific subgroups of patients to limit variability. As the size of our validation cohort increases we will enter underlying diseases and primary diagnoses into a multivariate regression model to account for underlying diseases or separate patients into subgroups according to their diagnosis. We are now drawing fresh blood on the day that patients enter the ICU in order to measure plasma concentrations of cytokines such as IL-6 and other soluble inflammatory mediators such as sCD14 and M B L at what we judge as the beginning of their inflammatory response. We may also draw blood later in the course of their hospital stay in order to measure 179 differences in peak concentrations of cytokines, and greatest change from baseline concentrations, although this will be confounded by early deaths in the sickest groups of patients. Septic patients in the ICU will most likely have more day-to-day variability than cardiac surgery patients, but with a large validation cohort we will have adequate power to detect differences in cytokine concentrations by genotype. As well as measuring soluble factors in patients' peripheral blood, we could also measure the density of membrane-bound receptors such as TLR2 and CD 14 on the surface of peripheral blood mononuclear cells (PBMCs) isolated from patients' blood using flow cytometry. We could also measure the levels of intracellular proteins such as IRAK4 in PBMCs by Western blot. Determining whether genetic variants of candidate genes are associated with in vivo protein levels of their respective proteins is an important step to establish biological plausibility for genetic associations. Determination of Functional SNPs and Mechanistic Studies While the functional SNPs of CD14 and M B L have been well characterized in 23 25 27 29 previous studies ' ' , the functional SNPs responsible for the association of haplotype clades of TLR2 and IL-6 with clinical phenotype have yet to be determined. I believe the IRAK4 Ala428Thr is SNP responsible for the association of the IRAK4 C/T/A clade with prevalence of Gram-positive cultures, although further experiments are needed to prove this. In the future we should resequence candidate genes in a larger group of individuals in order to detect even very low frequency SNPs contained within haplotype clades associated with outcome. We could then test individual promoter or coding SNPs contained within a clinically important haplotype for association to outcome in a large cohort of patients in order to determine the functional SNP marked by the haplotype. 180 If probable functional SNPs are found for IL-6 and TLR2 we could then examine through what mechanism these SNPs and 1RAK4 Ala428Thr alter the inflammatory response to infection. We could first measure ex vivo IL-6, TLR2 and IRAK4 RNA expression in PBMCs isolated from critically ill patients and test for the association of RNA levels with candidate functional SNPs. Additionally, we could measure RNA expression of genes of different genotypes using reporter gene systems in transfection assays, or by stimulating lymphocyte cell lines with known genotype from the Coriell Cell Repositories with LPS, peptidoglycan and other important immune response stimulants and measure levels of IL-6, TLR2 and IRAK4 messenger RNA expression at specific time points using quantitative reverse-transcriptase PCR. In this way we could determine if probable functional SNPs affect transcription of their respective genes. Furthermore, if functional SNPs of IL-6, TLR2 or 1RAK4 were found in the promoter region of these genes we could use bioinformatics tools to determine what transcription factors bind these regions and measure the efficiency of transcription factor binding to different alleles using electrophoretic mobility shift assays to understand how the functional SNPs regulated RNA transcription. It is highly probable that IRAK4 Ala428Thr is the functional SNP responsible for the association of the IRAK4 C/T/A clade with increased prevalence of Gram-positive cultures. In order to elucidate its mechanism of action macrophage and lymphocyte cell lines of known IRAK4 Ala428Thr genotype could be stimulated with whole bacteria or bacterial products and activation and nuclear translocation of N F - K B quantified and compared among cell lines of different genotype. Phosphorylation of IKKa/p, IKKOI and I K K P could be detected by immunoblot. Kinase assays could be performed in the different stimulated cell lines to examine the activation of IRAK4 and also the downstream kinases p38, ERK, JNK, and Akt and their role in activating N F - K B following TLR binding by pathogen-associated 181 molecular patterns (PAMPs). Similar experiments could be performed in immune cells of different TLR2 genotype in order to measure differences in receptor signaling caused by a functional SNP. Additionally, we could measure whether the IRAK4 Ala428Thr polymorphism affects IRAK4 interaction with MyD88, IRAKI and MyD88-adapter-like/Toll-interleukin 1 receptor (TIR) domain-containing adapter protein (Mal/TIRAP) using a yeast 2-hybrid system. Considering that haplotype clades of innate immunity genes appear to be associated with the prevalence of positive microbiological cultures and prevalence of infection at admission to ICU, it would be interesting to develop a killing assay to measure the ability of immune cells of different genotype to kill or phagocytose live bacteria. Functional SNPs of innate immunity genes may alter the ability of effector cells to recognize and fight off an invading pathogen. Further Genetic Association Studies and Interactions Finally, the inflammatory and innate immune responses to infection are complex and involve hundreds of interacting genes and gene products. In the future it will be important to test htSNPs in other candidate genes in the inflammatory and innate immunity pathways for association to outcome from SIRS in thousands of patients, and to test how these and other SNPs interact to predict outcome. As gene-gene and gene-environment interactions are studied, we will require sample sizes in the thousands of patients to be adequately powered ' 19 182 5.5 References 1. Fishman D, Faulds G, Jeffery R et al. The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemic-onset juvenile chronic arthritis. J Clin Invest 1998;102:1369-76. 2. Schluter B, Raufhake C, Erren M et al. Effect of the interleukin-6 promoter polymorphism (-174 G/C) on the incidence and outcome of sepsis. Crit Care Med 2002;30:32-7. 3. Gaudino M , Andreotti F, Zamparelli R et al. The -174G/C interleukin-6 polymorphism influences postoperative interleukin-6 levels and postoperative atrial fibrillation. Is atrial fibrillation an inflammatory complication? Circulation 2003; 108 Suppl 1:11195-9. 4. Roth-Isigkeit A, Hasselbach L, Ocklitz E et al. Inter-individual differences in cytokine release in patients undergoing cardiac surgery with cardiopulmonary bypass. Clin Exp Immunol 2001;125:80-8. 5. Kawai T, Adachi O, Ogawa T, Takeda K, Akira S. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity 1999;11:115-22. 6. Kawai T, Takeuchi O, Fujita T et al. Lipopolysaccharide stimulates the MyD88-independent pathway and results in activation of IFN-regulatory factor 3 and the expression of a subset of lipopolysaccharide-inducible genes. J Immunol 2001;167:5887-94. 7. Girardin SE, Boneca IG, Carneiro LA et al. Nodi detects a unique muropeptide from gram-negative bacterial peptidoglycan. Science 2003;300:1584-7. 8. Chamaillard M , Hashimoto M , Horie Y et al. An essential role for NODI in host recognition of bacterial peptidoglycan containing diaminopimelic acid. Nat Immunol 2003;4:702-7. 9. Yang K, Puel A, Zhang S et al. Human TLR-7-, -8-, and -9-mediated induction of IFN-alpha/beta and -lambda Is IRAK-4 dependent and redundant for protective immunity to viruses. Immunity 2005;23:465-78. 10. Lee JS, Nauseef WM, Moeenrezakhanlou A et al. Monocyte pi 10{alpha} phosphatidylinositol 3-kinase regulates phagocytosis, the phagocyte oxidase, and cytokine production. JLeukoc Biol 2007;81:1548-61. 11. Vink JM, Boomsma DI. Gene finding strategies. Biol Psychol 2002;61:53-71. 12. Hack CE, De Groot ER, Felt-Bersma RJ et al. Increased plasma levels of interleukin-6 in sepsis. Blood 1989;74:1704-10. 13. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T. The complex pattern of cytokines in serum from patients with meningococcal septic shock. Association between interleukin 6, interleukin 1, and fatal outcome. J Exp Med 1989;169:333-8. 14. Martin C, Boisson C, Haccoun M , Thomachot L, Mege JL. Patterns of cytokine evolution (tumor necrosis factor-alpha and interleukin-6) after septic shock, hemorrhagic shock, and severe trauma. Crit Care Med 1997;25:1813-9. 15. Keavney B. Genetic association studies in complex diseases. J Hum Hypertens2000;\4:361-7. 16. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet 2001;2:91-9. 17. Cardon LR, Palmer LJ. Population stratification and spurious allelic association. Lancet 2003;361:598-604. 18. Wang S, Zhao H. Sample size needed to detect gene-gene interactions using association designs. Am J Epidemiol 2003;158:899-914. 183 19. Keavney B, McKenzie C, Parish S et al. Large-scale test of hypothesised associations between the angiotensin-converting-enzyme insertion/deletion polymorphism and myocardial infarction in about 5000 cases and 6000 controls. International Studies of Infarct Survival (ISIS) Collaborators. Lancet 2000;355:434-42. 20. Akey J, Jin L, Xiong M. Haplotypes vs single marker linkage disequilibrium tests: what do we gain? Eur J Hum Genet 2001;9:291-300. 21. Zhang K, Calabrese P, Nordborg M , Sun F. Haplotype block structure and its applications to association studies: power and study designs. Am J Hum Genet 2002;71:1386-94. 22. Templeton AR, Weiss K M , Nickerson DA, Boerwinkle E, Sing CF. Cladistic structure within the human lipoprotein lipase gene and its implications for phenotypic association studies. Genetics 2000;156:1259-75. 23. Hubacek JA, Rothe G, Pit'ha J et al. C(-260)->T polymorphism in the promoter of the CD14 monocyte receptor gene as a risk factor for myocardial infarction. Circulation 1999;99:3218-20. 24. Baldini M, Lohman IC, Halonen M, Erickson RP, Holt PG, Martinez FD. A Polymorphism* in the 5' flanking region of the CD14 gene is associated with circulating soluble CD14 levels and with total serum immunoglobulin E. Am J Respir Cell Mol Biol 1999;20:976-83. 25. Fidler KJ, Wilson P, Davies JC, Turner MW, Peters MJ, Klein NJ. Increased incidence and severity of the systemic inflammatory response syndrome in patients deficient in mannose-binding lectin. Intensive Care Med2004;30:1438-45. 26. Crosdale DJ, Oilier WE, Thomson W et al. Mannose binding lectin (MBL) genotype distributions with relation to serum levels in UK Caucasoids. Eur J Immunogenet 2000;27:111-7. 27. Madsen HO, Garred P, Thiel S et al. Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein. J Immunol 1995;155:3013-20. 28. Steffensen R, Thiel S, Vanning K, Jersild C, Jensenius JC. Detection of structural gene mutations and promoter polymorphisms in the mannan-binding lectin (MBL) gene by polymerase chain reaction with sequence-specific primers. / Immunol Methods 2000;241:33-42. 29. Garred P, J JS, Quist L, Taaning E, Madsen HO. Association of mannose-binding lectin polymorphisms with sepsis and fatal outcome, in patients with systemic inflammatory response syndrome. J Infect Dis,2003;\ 88:1394-403. 30. Gibot S, Cariou A, Drouet L, Rossignol M , Ripoll L. Association between a genomic polymorphism within the CD14 locus and septic shock susceptibility and mortality rate. Crit Care Med 2002;30:969-73. 31. Heesen M , Bloemeke B, Schade U, Obertacke U, Majetschak M. The -260 C ->T promoter polymorphism of the lipopolysaccharide receptor CD 14 and severe sepsis in trauma patients. Intensive Care Med 2002;28:1161 -3. 32. Agnese D M , Calvano JE, Hahm SJ et al. Human toll-like receptor 4 mutations but not CD14 polymorphisms are associated with an increased risk of gram-negative infections. J Infect Dis 2002;186:1522-5. 33. Barber RC, Chang L Y , Arnoldo BD et al. Innate Immunity SNPs are Associated with Risk for Severe Sepsis after Burn Injury. Clin Med Res 2006;4:250-5. 34. D'Avila LC, Albarus M H , Franco CR et al. Effect of CD14 -260C>T polymorphism on the mortality of critically ill patients. Immunol Cell Biol 2006;84:342-8. 184 35. Nakada TA, Hirasawa H, Oda S et al. Influence of toll-like receptor 4, CD14, tumor necrosis factor, and interleukine-10 gene polymorphisms on.clinical outcome in Japanese critically ill patients. J Surg Res 2005;129:322-8. 36. Zhang DL, Zheng H M , Yu BJ, Jiang ZW, Li JS. Association of polymorphisms of IL and CD14 genes with acute severe pancreatitis and septic shock. World J Gastroenterol 2005;11:4409-13. 37. Lorenz E , Mira JP, Cornish K L , Arbour NC, Schwartz DA. A novel polymorphism in the toll-like receptor 2 gene and its potential association with staphylococcal infection. Infect Immun 2000;68:6398-401. 38. Bochud PY, Hawn TR, Aderem A. Cutting edge: a Toll-like receptor 2 polymorphism that is associated with lepromatous leprosy is unable to mediate mycobacterial signaling. J Immunol 2003;170:3451-4. 39. Day N, Tangsinmankong N, Ochs H et al. Interleukin receptor-associated kinase (IRAK-4) deficiency associated with bacterial infections and failure to sustain antibody responses. J Pediatr 2004;144:524-6. 40. Davidson DJ, Currie AJ, Bowdish D M et al. IRAK-4 mutation (Q293X): rapid detection and characterization of defective post-transcriptional TLR/IL-1R responses in human myeloid and non-myeloid cells. J Immunol 2006;177:8202-11. 41. Currie AJ, Davidson DJ, Reid GS et al. Primary immunodeficiency to pneumococcal infection due to a defect in Toll-like receptor signaling. J Pediatr 2004;144:512-8. 42. Enders A, Pannicke U, Berner R et al. Two siblings with lethal pneumococcal meningitis in a family with a mutation in Interleukin-1 receptor-associated kinase 4. J Pediatr 2004;145:698-700. 43. Ku C L , Yang K, Bustamante J et al. Inherited disorders of human Toll-like receptor signaling: immunological implications. Immunol Rev 2005;203:10-20. 44. Medvedev A E , Lentschat A, Kuhns DB et al. Distinct mutations in IRAK-4 confer hyporesponsiveness to lipopolysaccharide and interleukin-1 in a patient with recurrent bacterial infections. J Exp Med 2003; 198:521-31. 45. Picard C, Puel A, Bonnet M et al. Pyogenic bacterial infections in humans with IRAK-4 deficiency. Science 2003;299:2076-9. Epub 2003 Mar 13. 46. Marshall RP, Webb S, Hill MR, Humphries SE, Laurent GJ. Genetic polymorphisms associated with susceptibility and outcome in ARDS. Chest 2002;121:68S-69S. 47. Heesen M , Bloemeke B, Heussen N, Kunz D. Can the interleukin-6 response to endotoxin be predicted? Studies of the influence of a promoter polymorphism of the interleukin-6 gene, gender, the density of the endotoxin receptor CD 14, and inflammatory cytokines.PG. Crit Care Med 2002;30:664-9. 48. Heesen M , Obertacke U, Schade FU, Bloemeke B, Majetschak M. The interleukin-6 G(-174)C polymorphism and the ex vivo interleukin-6 response to endotoxin in severely injured blunt trauma patients. Eur Cytokine Netw 2002; 13:72-7. 49. Kristof AS, Magder S. Low systemic vascular resistance state in patients undergoing cardiopulmonary bypass. Crit Care Med 1999;27:1121-7. 50. Jones RH, Hannan E L , Hammermeister K E et al. Identification of preoperative variables needed for risk adjustment of short-term mortality after coronary artery bypass graft surgery. The Working Group Panel on the Cooperative C A B G Database Project. JAm Coll Cardiol 1996;28:1478-87. 185 51. Abildstrom H, Christiansen M , Siersma V D , Rasmussen LS. Apolipoprotein E genotype and cognitive dysfunction after noncardiac surgery. Anesthesiology 2004;101:855-61. 52. Hsiung GY, Sadovnick AD, Feldman H. Apolipoprotein E epsilon4 genotype as a risk factor for cognitive decline and dementia: data from the Canadian Study of Health and Aging. Cmaj 2004;171:863-7. 186 APPENDIX 1: RECRUITMENT AND PHENOTYPING OF PATIENT COHORTS FOR GENETIC ASSOCIATION STUDIES Al . l Recruitment of Critically 111 Patients for Genetic Association Studies A prospective cohort of critically ill patients was recruited from the Intensive Care Unit at St. Paul's Hospital, Vancouver, British Columbia. St. Paul's Hospital is part of Providence Health Care and is a teaching hospital affiliated with the University of British Columbia. The ICU at St. Paul's Hospital is a 16 bed mixed medical-surgical ICU that has about 750-800 admissions per year. All patient recruitment protocols were approved by the Providence Health Care Research Ethics Board and by the University of British Columbia Research Ethics Board. Dr. James Russell and Dr. Keith Walley, who are both staff physicians at St. Paul's Hospital, were the Principal Investigators for all studies. All patients admitted to ICU at St. Paul's Hospital were included in the cohort on the day that they met 2 of 4 criteria for SIRS: 1) fever (>38°C) or hypothermia (<36 °C), 2) tachycardia (>90 beats/minute), 3) tachypnea (>20 breaths/minute), PaC02 <32 mm Hg, or need for mechanical ventilation, and 4) leukocytosis (total leukocyte count >12,000 mm3) or leukopenia (<4,000 mm3)1. The majority of patients (95%) met the at least 2 of 4 criteria for SIRS at admission to ICU. Patients were excluded from the cohort if they had 1) severe underlying chronic disease such as a malignancy that would limit 28-day survival, 2) severe underlying chronic obstructive lung disease (FEVi <500 rhL or on home oxygen) or severe chronic restrictive lung disease. The latter impairs weaning from ventilation and thus confounds determination of days alive and free of acute lung injury. For patients with repeat admissions to the ICU, the repeat admissions were excluded from the analysis. Medical care was provided as per the discretion of the ICU team. At the time of recruitment we were not required to obtain consent as all patients 187 identifiers were discarded and patient samples and data were assigned random six-digit identification numbers that could not be traced back to individual patients. Patients of all ethnicities were recruited to the study and their DNA and clinical data were collected and stored. However, at this time there are not sufficient numbers of patients in ethnic groups other than Caucasians to have adequate power to study non-Caucasians. Our future plan is to collaborate with other sites so that we can obtain data on additional ethnic groups. A1.2 Collection of Clinical Data for the Cohort of Critically III Patients After admission to the prospective cohort, patients were observed for 28 days. Twenty-eight days is the standard treatment/follow-up time, for clinical trials in sepsis and acute lung injury 2~ s. Additionally, most critically ill patients have either died or have 2 5 been successfully weaned from mechanical ventilation within 28 days " . Clinical data was gathered by patient chart review by trained Intensivists and recorded in chart review forms (CRTs) (Table A-l). Baseline demographic data, including age, gender, medical versus surgical diagnosis on admission (according to APACHE III diagnostic codes) 6 , co-morbidities as defined by the APACHE II Chronic Health Evaluation points , admission APACHE II score, and source and microbiology of positive cultures at admission were recorded 1 . Positive cultures were categorized as Gram-positive, Gram-negative, mixed, fungal or other. Sources of infection were categorized as respiratory (sputum), gastrointestinal, genitourinary (urine), endovascular (blood, lines, valves), or from skin, soft tissues or wounds. Cultures were categorized as contaminations or colonizations at the discretion of the attending physician. 188 Clinical data were recorded for each 24-hour period (8 am to 8 am) for 28 days or until hospital discharge to evaluate organ dysfunction. For each day of the 28 day observation period raw clinical and laboratory variables were recorded using the worst or most abnormal variable for each 24-hour period with the exception of Glasgow Coma Score (where the best possible score for the 24-hour period was recorded). Because data were not always available during each 24 hour period for each organ dysfunction variable, we used the "carry forward" assumption as defined previously 8 . Missing data on the day of admission was assigned a normal value and missing data after day 1 was substituted by carrying forward the previous day's value. If any variable was never measured, it was assumed to be normal. The CRF used to record data is included as Table A- l . When data collection for each patient was complete, all patient identifiers were removed from all records and the patient file was assigned a unique random number that was cross-referenced with the blood samples. The completed raw data file was converted to calculated descriptive and severity of illness scores using standard definitions according to the Bmssels Organ Dysfunction Scoring System (Table A-2). To assess duration of organ dysfunction (cardiovascular, respiratory, renal, hepatic, hematologic, and neurologic organ systems) and SIRS and to correct organ dysfunction and SIRS scoring for deaths in the 28-day observation period we calculated days alive and free (DAF) of organ dysfunction as previously reported 9 . During each 24 hour period (8 am to 8 am), for each variable, days alive and free was scored as 1 if the patient was alive and free of organ dysfunction (normal or mild dysfunction) (Table A-2). Days alive and free was scored as 0 if the patient had organ dysfunction (moderate or worse) or was not alive. Every day over the 28-day observation period after meeting inclusion criteria was scored in this way. Thus, the lowest score possible for each variable was zero and the highest score possible was 28. A low score indicates more organ dysfunction as there would be fewer days alive and free of organ dysfunction, 189 while a high score indicates less organ dysfunction. Days alive and free of organ support was also recorded. Days alive and free of vasopressors was recorded as a measure of cardiovascular support. Vasopressor use was defined as dopamine >5 ug/kg/min or any dose of norepinephrine, epinephrine, vasopressin, or phenylephrine. Days alive and free of mechanical ventilation was recorded as a measure of respiratory support. Mechanical ventilation was defined as the need for intubation and positive airway pressure. Spontaneous breathing while intubated with a T-piece and non-invasive ventilation was not considered mechanical ventilation. Renal support was defined as hemodialysis, peritoneal dialysis, or any continuous renal support mode (e.g. CWHD) . Patients were defined as having sepsis if they met at least 2 of 4 SIRS criteria and had a proven (positive microbiological culture) or suspected infection (clinician prescribed antibiotics). Patients were defined as having septic shock if they had sepsis in addition to hypotension. The presence or absence of sepsis and septic shock was recorded for each day of the 28 day observation period. A1.3 Clinical Outcomes in the Cohort of Critically 111 Patients The association of innate immunity genetic variants to risk, type and source of infection was tested. Genetic variants were tested for association to the prevalence of positive microbiological cultures at admission to ICU, the type of positive microbiological cultures (Gram-positive, Gram-negative, mixed, fungal or other), and the source of positive microbiological cultures (respiratory (sputum), gastrointestinal, genitourinary (urine), endovascular (blood, lines, valves), or from skin, soft tissues or wounds). Genetic variants were also tested for association to the prevalence of infection at admission to ICU, which was defined as the presence of a proven (positive microbiological culture) or suspected (clinician prescribes antibiotics) infection. 190 Twenty eight day mortality was the primary outcome variable. Secondary outcome variables were days alive and free of SIRS, days alive and free of sepsis, days alive and free of septic shock, and days alive and free of organ dysfunction and organ support (cardiovascular, respiratory, renal, hepatic, hematologic, and neurologic organ systems). Clinically significant organ dysfunction for each organ system was defined as present during a 24 hour period if there was evidence of at least moderate organ dysfunction using the Brussels criteria (Table A-2) 1 0 . A1 .4 Co l lect ion and Storage of C r i t i c a l l y 111 Pat ients ' D N A Discarded whole blood from routine clinical laboratory tests was collected for each patient in order to extract genomic DNA. Samples were centrifuged at 400g for 15 minutes at 4°C. The buffy coat (containing the leukocytes) was collected and transferred into 2.0mL cryotubes and stored at -80°C. The tubes were labeled with the six-digit identification number assigned to a particular patient. DNA was extracted from the buffy coat using the Qiagen DNA Blood Mini Kit (Qiagen Inc. Mississauga, ON) and quantified using a Picogreen assay. The genomic DNA was then transferred to 2.0mL cryotubes labeled with patient's six digit identification number or to 96 well plates for high-throughput genotyping. Genomic DNA concentrations and volumes were stored in the laboratory database computer. A1.5 Recru i tment o f Patients Undergo ing C o r o n a r y A r t e r y Bypass G r a f t w i th C a r d i o p u l m o n a r y Bypass for Genet ic Associat ion Studies A prospective cohort of patients undergoing elective coronary artery bypass graft (CABG) surgery with cardiopulmonary bypass was recruited as a secondary cohort. Patients were recruited from St. Paul's Hospital in Vancouver, British Columbia between February 2001 and December 2003. St. Paul's Hospital is the designated Heart Centre of the University of British Columbia. All patient recruitment protocols were approved by 191 the Providence Health Care Research Ethics Board and by the University of British Columbia Research Ethics Board. Dr. James Russell and Dr. Keith Walley, who are both staff physicians at St. Paul's Hospital, were the Principal Investigators for all studies. At the time of recruitment we were not required to obtain consent as all patients identifiers were discarded and patient samples and data were assigned random six-digit identification numbers that could not be traced back to individual patients. A subset of C A B G with cardiopulmonary bypass patients were recruited as part of a separate clinical trial, "Albumin ameliorates the inflammatory response to and organ dysfunction after cardiac surgery especially in patients with TNF, ILl-ra and IL-6 polymorphisms". These patients were approached for consent to participate in the study in the pre-admissions clinic of St. Paul's Hospital by a research coordinator. They read and signed and informed consent form. These patients had fresh blood drawn for cytokine measurements. Emergent patients were excluded from the cohort as these patients may already exhibit an inflammatory response or other triggers such as shock, making it more difficult to isolate the proportion of the response that can be attributed to genotype. Valve replacement and repeat C A B G patients were also excluded from the cohort as they may have different pre-operative pathophysiology and often have longer total surgical or CPB times. Off-pump patients were excluded from the cohort as we wanted to examine the role of genotype in determining the inflammatory response to cardiopulmonary bypass. At this time we limit our analysis to Caucasian patients as we currently do not have sufficient patients of other races to obtain statistically meaningful results. Our future plan is to collaborate with other sites so that we can obtain data on additional ethnic groups. 192 A1.6 Collection of Clinical Data for the Cohort of Cardiopulmonary Bypass Patients CABG patients' clinical data was gathered by patient chart review by medical residents and graduate students and recorded in CRFs (Table A-3). Baseline demographics recorded were age, gender, the presence of diabetes mellitus, body mass index, smoking, use of angiotensin-converting enzyme inhibitors, use of beta-blockers, use of calcium channel blockers and use of immunosuppressives. Surgeon, pump-time, cross-clamp time, total surgery time, use of intra-operative milrinone, aprotinin and amicar were recorded. Patient clinical data detailed in Table A-3 was collected pre-operatively, immediately post-operatively and 4, 12 and 24 hours post-operatively. Patient clinical data included the Glasgow Coma Scale, temperature, heart rate, whether the patient was paced, systolic blood pressure (BP), diastolic BP, mean arterial pressure (MAP), central venous pressure (CVP), pulmonary artery (PA) diastolic pressure, cardiac output, cardiac index, carry time, whether the patient was ventilated, respiratory rate, and what drugs they were receiving. Whether a patient received transfused blood products pre-operatively or post-operatively was recorded, as was patients' arterial blood gases and laboratory data pre-operatively and immediately and 4, 12 and 24 hours post-operatively. A1.7 Clinical Outcomes in the Cohort of CABG with Cardiopulmonary Bypass Patients Mortality from CABG with cardiopulmonary bypass is rare and so it was necessary to examine other outcome measures in the cardiac surgery cohort. Kristof and Magder 1 1 found that low systemic vascular resistance index (SVRI) was a particularly useful clinical manifestation of systemic inflammation. They found that vasodilatory syndrome identified by low SVRI was associated with important clinical parameters including longer cross-clamp times and lower post-CPB platelet count ". Low systemic 193 vascular resistance (SVR) as defined as an indexed low systemic vascular resistance (SVRi) less than 1800 dyne*sec/cm5/m2 at 2 consecutive time points post-operatively 1 1 was the primary outcome variable (SVRi = [(MAP-CVP)*80]/CI, where MAP is the mean arterial pressure, CVP is the central venous pressure and CI is the cardiac index). Because mortality after CPB is extremely low, we measured hours in cardiac surgery intensive care unit (CSICU) after surgery as a clinically relevant secondary outcome variable. Patients were followed until hospital discharge or death. In addition to post-operative vasodilation, extended length of stay in the cardiac surgery ICU is a common measure of adverse outcome in the literature 1 2 , 1 3 . We measured hours in cardiac surgery intensive care unit (CSICU) after surgery as a clinically relevant outcome variable. In St. Paul's Hospital the decision to discharge patients from the ICU following CPB surgery is protocol driven and genotype was unknown to care providers, making this measurement a reasonably unbiased outcome measurement. Patients were followed until hospital discharge or death. A1.8 Intermediate Phenotypes in the Cohort of CABG with Cardiopulmonary Bypass Patients Serum cytokines were measured in a subset of cardiac surgery patients who had blood drawn pre-operatively and at 3 and 24 hours post-operatively. The inflammatory response of cardiac surgery patients begins at a known time, the time of surgery, so that they are a more homogenous group of patients than critically ill patients in which to measure the development of the inflammatory response. In contrast, critically ill patients with SIRS may experience a systemic inflammatory response to a wide array of different stimuli, including infectious microorganisms, surgery or trauma. Most of the time we do not know precisely when the inflammatory stimulus occurred in a critically ill patient. 194 For these reasons it is more reliable to measure the production of cytokines in the cardiac surgery cohort to compare by genotype. Patients' serum samples were diluted four times and serum cytokine concentrations of Interleukin-6 (IL-6), Monocyte chemotactic protein-1 (MCP-1), Granulocyte-Colony Stimulating Factor (G-CSF), IL-8 and IL-1 receptor antagonist (IL-lra) were measured in the subset of cardiac surgery patients using the Luminex 100IS system for multiplex assays with reagents from R&D Systems (Minneapolis, MN) according to the manufacturer's directions. The Luminex system uses 5.6pm polystyrene beads with are internally dyed with red and infrared fluorophores. Each bead set is filled with different ratios of the two dyes so that each bead has a unique spectral signature based on the red/infrared ratio. Beads are then coated with antibodies specific to the protein of interest. A secondary antibody with a biotin tag is added, and finally avidin-tagged reporter dyes. Beads run through the capillary in the Luminex 100IS system and lasers excite the internal dyes that identify each bead, and also any reporter dye captured during the assay. A1.9 Collection and Storage of Cardiopulmonary Bypass Patients' DNA and serum Discarded whole blood from routine clinical laboratory tests was collected for each patient in order to extract genomic DNA. Samples were centrifuged at 400g for 15 minutes at 4°C. The buffy coat was collected and transferred into 2.0mL cryotubes and stored at -80°C. The tubes were labeled with the six-digit identification number assigned to a particular patient. DNA was extracted from the buffy coat using the Qiagen DNA Blood Mini Kit (Qiagen Inc. Mississauga, ON) and quantified using a Picogreen assay. The genomic DNA was then transferred to 2.0mL cryotubes labeled with patient's six digit identification number or to 96 well plates for high-throughput genotyping. Genomic DNA concentrations and volumes were stored in the laboratory database computer. 195 A subset of 162 cardiopulmonary bypass patients were recruited as part of a separate randomized control trial. These patients had fresh blood drawn pre-operatively and at 4 and 24 hours post-operatively. The fresh blood was centrifuged at 400g for 15 minutes at 4°C. The serum was drawn off with glass pipettes and stored in 2.0mL cyrotubes labeled with barcodes unique to the patient. The serum samples were stored at -80°C. The buffy coat was then collected and processed as above. 196 Table Al-1. Chart Review Form for collection of critically ill patients' data Patient ID SEX: female=1; male=0 AGE Aplll Dx Ethnicity Actual Day (dd/mm/yyyy) Day (1-28) 1 2 3 4 5 6 7 8 LOCATION: 1=ICU; 2=CSICU; 3=CCU; 4=HOSP WARD; 5=HOME: 6=DEAD: 7=OTHER HOSP STEROIDS: 1 = physiological replacement; 2=pulse high dose Warfarin/aPC : 1=warfarin; 2=aPC CXR - infiltrates: 1= focal (excluding LLL atelectasis); 2=diffuse or bilateral or pulmonary edema TEMP MOST ABNORMAL (HIGH OR LOW) HHR HIGHEST THAT DAY HR AT SAME TIME AS LOWEST MAP SAP LOWEST MAP LOWEST CVP AT SAME TIME AS LOWEST MAP GCS BEST THAT DAY NEURO SCORE BEST THAT DAY Dopamine MCG/KG/MIN - HIGHEST DOSE Dobutamine HIGHEST DOSE Epinephrine HIGHEST DOSE Milrinone HIGHEST DOSE Norepinephrine HIGHEST DOSE Vasopressin HIGHEST DOSE Sedation: 0=NONE; 1= INTERMITTENT; 2=CONTINUOUS INFUSION RR HIGHEST THAT DAY Table Al -1 . Chart Review Form for collection of critically ill patients' data UO / HR LOWEST FOR 2 CONSECUTIVE HOURS - RECORD ONLY WHEN 24 HOUR NOT AVAILABLE (IE DAY ONE) NUTRITION 0=NONE; 1=ENTERAL; 2=TPN - MUST BE FOR AT LEAST HALF OF DAY RENAL SUPPORT; 0=NONE; ^INTERMITTENT; 2=CONTINUOUS. FOR INTERMITTENT. RECORD FIRST DAY UO /DAY 24 HOUR TOTAL IF AVAILABLE VENT; 0=NO; 1=YES (RECORD 1 IF VENTILATED AT ALL) Fi02 RECORD VALUE ASSOCIATED WITH LOWEST P/F RATIO Sa02 RECORD ONLY IF PA02 NOT MEASURED Pa02: RECORD VALUE ASSOCIATED WITH LOWEST P/F RATIO Ph most abnormal pC02 same time as pH Glucose most abnormal Sodium most abnormal Potassium most abnormal HC03 most abnormal BUN most abnormal Creatinine most abnormal Lactate most abnormal Bilirubin record most abnormal if missing on day one, record value within 10 days of admission Albumen record most abnormal, if missing on day one, use closest value within 10 days WBC most abnormal HGB lowest Table Al-1. Chart Review Form for collection of critically i l l patients' data Hematocrit lowest Platelets lowest INR highest Fibrinogen lowest FDP's highest CULTURE: 1= Gram's Pos; 2=Gram's Neg; 3=Mixed; 4=Funqal; 5=Other SOURCE OF CULTURE: 1=Resp; 2=GI; 3=Skin, soft tissues, wounds; 4=GU; 5=Endovascular (blood,lines, valves); Organism 1 Organism 2 Clinical suspicion sepsis (ie Started antibiotics) Source of suspected sepsis 1=Resp; 2=GI; 3=Skin, soft tissues, wounds; 4=GU; 5=Endovascular (blood,lines, valves); 6=Neuro; 7=other Troponin Table Al-2. Brussels Organ Dysfunction Scoring System used for assessing organ dysfunction in critically ill patients ORGANS Free of Organ Dysfunction Clinically Significant Organ Dysfunction Normal Mild Moderate Severe Extreme Cardiovascular Systolic BP (mmHg) >90 < 90 Responsive to fluid <90 Unresponsive to fluid < 90 plus pH < 7.3 < 90 plus pH < 7.2 Pulmonary Pa02/FI02 (mmHg) >400 400-301 300-201 Acute . lung injury 200-101 ARDS < 100 Severe ARDS Renal Creatinine (mg/dL) <1.5 1.5-1.9 2.0-3.4 3.5-4.9 >5.0 Hepatic Bilirubin (mg/dL) <1.2 1.2-1.9 2.0-5.9 6.0-11.9 > 12 Hematologic Platelets (xl05/mm3) >120 120-81 80-51 50-21 <20 Neurologic (Glasgow Score) 15 14-13 12-10 9-6 <5 Round Table Conference on Clinical Tria March 12-14, 1994. s for the Treatment of Sepsis, Brussels, 200 Table Al -3 . Chart Review Form for cohort of CABG with cardiopulmonary bypass patients Background Information UNIQUE # Aspirin CSICU Trial Age ACE Inhibitor Placebo/Intervention Gender (M=0,F=1) P-Blocker From ICU? Ethnicity Ca + +-Blocker To ICU? Height (m) Steroid Use CSICU Outcome Weight (kg) Total Cholesterol Hospital Outcome Smoker HDL Cholesterol Hypertension LDL Cholesterol Diabetes Triglycerides Insulin-Dependent Type DM Cardiac Surgery Details Procedure Cardioplegia Temp. CSICU total Surgical Class Cardioplegia Type Hospital Total ASA Class (1-5) IntraOp Aprotinin O.R. Total Surgeon IntraOp Milrinone Bypass Total PreOp EF IntraOp Amicar X-Clamp Total PreOp Assessment IntraOp Protamine Pump? Time Point PreOp PostOp 4hr. 12 hr. 24 hr. Patient Clinical Data Glasgow Coma Scale Temperature Heart Rate Paced Systolic BP Diastolic BP MAP CVP PA Diastolic Cardiac Output Cardiac Index Carry Time Ventilated Respiratory Rate Dopamine Dobutamine Epinephrine Norepinephrine Phenylephrine Milrinone Nitroglycerine Nitroprusside Otherl 201 Table Al -3 . Chart Review Form for cohort of CABG with cardiopulmonary bypass patients UNIQUE # 0 Time Point PreOp PostOp 4hr. 12 hr. 24 hr. Transfused Blood Products Packed RBCs Platelets Albumin FFP Cryoprecipitate Other2 Arterial Blood Gases Measured Sat. F i0 2 Pa02 PaC0 2 PH Carry Time Patient Lab Data WBCs Hemoglobin MCV Platelets Neutrophils Lymphocytes Glucose Sodium Potassium Chloride Bicarbonate Creatinine Magnesium Troponin INR Carry Time 202 A l . l l References 1. Bone RC. The sepsis syndrome. Definition and general approach to management. Clin Chest Med 1996; 17:175-81. 2. Schoenfeld DA, Bernard GR. Statistical evaluation of ventilator-free days as an efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med 2002;30:1772-7. 3. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000;342:1301-8. 4. Bernard GR, Vincent JL, Laterre PF et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699-709. 5. Munford RS. Chapter 254. Severe Sepsis and Septic Shock. In: Dennis L. Kasper EB, Anthony S. Fauci, Stephen L. Hauser, Dan L. Longo, J. Larry Jameson, and Kurt J. Isselbacher, ed. Harrison's Principles of Internal Medicine, 16th Edition: McGraw-Hill Medical; 2005. 6. Knaus WA, Wagner DP, Draper EA et al. The A P A C H E III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100:1619-36. 7. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. A P A C H E II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 8. Bernard GR, Wheeler AP, Arons M M et al. A trial of antioxidants N-acetylcysteine and procysteine in ARDS. The Antioxidant in ARDS Study Group. Chest 1997;112:164-72. 9. Bernard G, Wheeler A, Arons M et al. A trial of antioxidants N-.acetylcysteine and procysteine in ARDS. The Antioxidant in ARDS Study Group. Chest. 112 vol; 1997:164-172. 10. Sibbald WJ, Vincent JL. Round table conference on clinical trials for the treatment of sepsis. Brussels, March 12-14, 1994. Intensive Care Med 1995;21:184-9. 11. Kristof AS, Magder S. Low systemic vascular resistance state in patients undergoing cardiopulmonary bypass. Crit Care Med 1999;27:1121-7. 12. Nakasuji M , Matsushita M , Asada A. Risk factors for prolonged ICU stay in patients following coronary artery bypass grafting with a long duration of cardiopulmonary bypass. J Anesth 2005; 19:118-23. 13. Lawrence DR, Valencia O, Smith EE, Murday A, Treasure T. Parsonnet score is a good predictor of the duration of intensive care unit stay following cardiac surgery. Heart 2000;83:429-32. 203 APPENDIX 2 : ETHICS CERTIFICATES OF APPROVAL 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100762/manifest

Comment

Related Items