UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Candidate genes other than the CFTR gene as possible modifiers of pulmonary disease severity in cystic… Frangolias, Despina Daisy 2008

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2008_spring_frangolias_despina.pdf [ 1.76MB ]
Metadata
JSON: 24-1.0066304.json
JSON-LD: 24-1.0066304-ld.json
RDF/XML (Pretty): 24-1.0066304-rdf.xml
RDF/JSON: 24-1.0066304-rdf.json
Turtle: 24-1.0066304-turtle.txt
N-Triples: 24-1.0066304-rdf-ntriples.txt
Original Record: 24-1.0066304-source.json
Full Text
24-1.0066304-fulltext.txt
Citation
24-1.0066304.ris

Full Text

CANDIDATE GENES OTHER THAN THE CFTR GENE AS POSSIBLE MODIFIERS OF PULMONARY DISEASE SEVERITY IN CYSTIC FIBROSIS BY DESPINA DAISY FRANGOLIAS  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE FACULTY OF GRADUATE STUDIES (EXPERIMENTAL MEDICINE)  THE UNIVERSITY OF BRITISH COLUMBIA (VANCOUVER) FEBRUARY 2008  © DESPINA DAISY FRANGOLIAS, 2008  ABSTRACT Cystic fibrosis (CF) is a single gene Mendelian disorder characterized by pulmonary disease and pancreatic insufficiency. Pulmonary disease is the major cause of death in CF patients. Although some cystic fibrosis transmembrane conductance regulator (CFTR) genotypes are associated with less severe disease, patients possessing the same genotype show great variation in pulmonary disease severity and progression. Genes involved in modulating the inflammatory response and genes increasing susceptibility to infection are proposed as modifiers of pulmonary disease severity. Polymorphisms selected for based on evidence that they affect the function of the gene and prevalence of the putative risk allele: 1) antiprotease gene alpha-1-antitrypsin (α1-AT), 2) innate immunity genes: mannose binding lectin (MBL2) (promoter [G→C] at -221 and codon 52 (Arg52Cys, D allele), 54 (Gly54Asp, B allele), and 57 (Gly57Glu, C allele), and pulmonary surfactant genes SPA-1 (Arg219Trp), SPA-2 (Thr9Asn, Lys223Gln) and SPD (Thr11Met), 3) antioxidant genes GSTM1 and T1 (gene deletion polymorphisms), GSTP1 (Ile105Val) and GCLC repeats, 4) mucin genes (MUC2 and MUC5B). Pulmonary disease progression and survival in patients with chronic Burkholderia cepacia complex (BCC) infection were also investigated controlling for genomovar and RAPD type of the organism. BCC infection was associated with more severe pulmonary disease progression and worse survival. α1-AT genotype was not a major contributor to variability of pulmonary disease severity, but the results suggest that α1-AT plasma levels during pulmonary infections may be affected by poor nutritional status. We showed similar pulmonary disease progression and MBL2 genotype. Contrary to the previous literature, wild-type MBL2 genotype was associated with steeper decline in pulmonary disease over time following chronic infection with BCC, but genotype was not associated with increased susceptibility to BCC infection. We showed inconsistant results for the pulmonary surfactant gene polymorphisms, GSTM1, T1 and GSTP1 polymorphisms, and number of repeats for GCLC and MUC5B depending on the phenotype investigated. We conclude that some of the variability in pulmonary disease severity and progression in CF is explained by polymorphisms in secondary genes.  ii  TABLE OF CONTENTS Abstract………………………………………………………………………....... Table of Contents……………………………………………………………........ List of Tables……………………………………………………………………... List of Figures………………………………………………………………….…. List of Abbreviations……………………………………………………….…….. Acknowledgements…………………………………………………….…………. Dedication………………………………………………………….……………... Co-authorship Statement.......................................................................................... Chapter 1: Overview and objectives……………………………….……………. 1.0 Introduction……………………….………………………………….. 1.2 The CF gene…………………………………………...……………… 1.2.1 Discovery of the Cystic Fibrosis gene…………………….......... 1.2.2 Structure, function and localization of the CFTR gene……..….. 1.2.3 Classification of the CFTR mutations…….……………………. 1.2.4 Cellular mechanisms of ion transport by normal and CF airway epithelia…………………………………………………………. 1.2.5 Model for the development of airway disease in Cystic Fibrosis.. 1.2.6 Determinants of pulmonary disease in Cystic Fibrosis…………. 1.3 Candidate modifier genes investigated…………………..……………. 1.3.1 Alpha-1-antitrypsin………………………………………..…….. 1.3.2 Innate immunity genes………………………….………..……… 1.3.3 Antioxidant genes……………………………………..………… 1.3.4 Pro-inflammatory and anti-inflammatory mediators …….…….. 1.3.5 Additional modifier genes……………………….……………… 1.4 Statement of the problem………………………..…………………….. 1.5 Main study hypothesis…………..…………………………………….. 1.6 Significance……………………………………………….…………... 1.7 Bibliography………………………………………….…….…………. Chapter 2: Study methodology…………………………………..……………….. 2.0 Methods and procedures………………………...…………………….. 2.1 Subject and eligibility criteria……………...………………………….. 2.2 Laboratory techniques…………………………………………………. 2.2.1 Extraction of genomic DNA from human whole blood ……..… 2.2.2 DNA quantification………………………..…………………… 2.3 Genotyping methods……………………………………………….…. 2.3.1 Amplification of DNA using polymerase chain reaction……..… 2.3.2 Genotyping using restriction fragment length polymorphisms.… 2.3.3 Amplification of DNA using site directed mutagenesis polymerase chain reaction followed by restriction digestion…… 2.3.4 Amplification of DNA using allele-specific oligonucleotide PCR or sequence specific priming polymerase chain reaction…. 2.4 Gene polymorphisms investigated and genotyping methods…….…… 2.4.1 Gene polymorphisms investigated and genotyping techniques…. 2.4.2 Alpha-1-antitrypsin gene………………………………..………. 2.4.3 Mannose-binding lectin 2 gene…………..……………………… 2.4.4 Pulmonary surfactant protein-A1 gene………..………………… 2.4.5 Pulmonary surfactant protein-A2 gene…………..……………… iii  ii iii vii x xi xiv xv xvi 1 2 3 3 6 8 9 10 12 13 14 16 18 19 20 21 22 23 24 35 36 36 37 37 38 38 38 39 39 40 40 40 41 42 44 45  2.4.6 Pulmonary surfactant protein-D gene………………..………..… 2.4.7 Other genes………………………………………………....…… 2.4.7.1 MUC2 gene……………………………………………... 2.4.7.2 MUC5B gene…………………………………………… 2.4.7.3 GSTP1 gene…………………………………………….. 2.4.7.4 GSTM1 and GSTT1 gene………………………………. 2.4.7.5 GCLC gene……………………………………………... 2.5 Phenotypic data collection……………………………………….......... 2.5.1 Pulmonary function……………………………………………… 2.5.2 Cross-sectional and longitudinal data …………………...……… 2.5.3 Definition of stable clinical status and pulmonary exacerbation in CF…………………………………..………………………… 2.5.4. Identification and typing of B. Cepacia complex……….…..….. 2.5.5. Sub-study data collection: Measurement of alpha-1-antitrypsin levels during a pulmonary exacerbation episode……………….. 2.6 Genetic analyses……………………………..………………………... 2.6.1 Haplotype construction……………………………………...…... 2.6.2 Calculation of linkage disequilibrium and Hardy Weinberg equilibrium……………………………………………………… 2.7 Statistical analysis…………………………………...………………… 2.7.1 Description of outcome variables………………...……………... 2.7.2 Description of independent variables…………...………………. 2.7.3 Analysis of pulmonary disease progression……..……………… 2.7.4 Survival…………………………………………...……………... 2.7.5 Age of first infection and chronic infection with P. aeruginosa………………………………………………………. 2.7.6 Pulmonary infections requiring therapy with intravenous antibiotics………………………………………….……………. 2.8 Hypotheses tested………………….………………………………….. Alpha-1-antitrypsin gene………….…………………………………… MBL2 gene…………………………………………………..………… BCC infection, genomovar and RAPD type grouping…………..…….. SPA-1 gene…………………………………………………………..… SPA-2 gene…………………………………………………………….. SPD gene……………………………………………………………….. GST genes and GCLC gene………………………………………......... MUC2 and MUC5B genes……………………………………………... 2.9 Bibliography…………………………………………………………... Chapter 3: Alpha-1-antitrypsin deficiency alleles in Cystic Fibrosis lung disease…………………………………………………………………. 3.0 Introduction…………………………………………………..………... 3.1 Rationale and main hypothesis……………………...………………… 3.2 Results………………………………………………………………..... 3.3 Discussion……………………………………………………………... 3.4 Conclusions………………………………………………………........ 3.5 Bibliography…………………………………………………………... Chapter 4: Innate immunity genes as potential modifier loci in Cystic Fibrosis………………………………………………………………… 4.0 Introduction……………………………………………...…………….. iv  46 47 47 47 48 48 49 49 49 50 52 53 53 54 54 54 55 55 56 58 61 61 62 62 62 63 65 66 67 68 68 69 93 95 96 96 98 101 105 112 115 116  4.1 Rationale for the investigation of innate immunity genes as potential modifiers in CF………………………………...……………………… 4.1.1 Tissue distribution of innate immunity proteins……...…………. 4.1.2 Characteristics of innate immunity proteins……………...……... 4.1.3 Structure and function of innate immunity proteins……...……... 4.1.4 Complement activation pathways………………………...……... 4.1.5 Review of the literature for MBL2, SPA and SPD: Clinical correlates…………………………………….…………………... 4.1.6 Review of the literature for Burkholderia cepacia complex infection in CF…………………………………………………... 4.2 Results……………………………………………………..…………... 4.2.1 Hardy Weinberg equilibrium and linkage disequilibrium……..… 4.2.2 Descriptive data results and study cohort grouping……………… 4.2.3 Pulmonary disease progression: mixed effects regression on %predFEV1………………………………………..……………… 4.2.4 Survival analysis……………………………………..…………... 4.2.5 Effect of modifier genes on P. aeruginosa infection status…...…. 4.2.6 Frequency of pulmonary infections requiring intravenous antibiotic therapy………………………………………………. 4.2.7 BCC infection and pulmonary disease progression and survival……………………………....………………………… 4.2.7.1 BCC infection and pulmonary disease progression…....... 4.2.7.2 BCC and P. aeruginosa infection and pulmonary disease progression…………………..................……………….. 4.2.7.3 BCC infection and survival……………....…………….. 4.2.7.4 Frequency of pulmonary infections requiring IV therapy 12 and 24 months pre- and post-colonization with BCC.. 4.3 Discussion…………………………………………......………………. 4.3.1 Innate immunity genes: MBL2, SPA and SPD…......…………… 4.3.2 BCC infection in CF...................................................................... 4.4 Conclusions…………………………………………………….......….. 4.5 Bibliography………………………………………………………....... Chapter 5: Glutathione metabolism associated genes as potential modifier genes in Cystic Fibrosis……................…………………………………….. 5.0 Introduction……......…………………………………………………... 5.1 Rationale for the investigation of innate immunity genes as potential modifiers in CF………………….......................……………………… 5.1.1 Tissue distribution and function of GSH…......…………………. 5.1.2 GSH deficiency is a common characteristic in CF......………….. 5.1.3 Modifier genes to explain the heterogeneity in pulmonary and liver disease in CF…............……………………………………. 5.1.3.1 Pathway for glutathione synthesis and metabolism and potential modifier genes for pulmonary disease severity in CF… 5.1.3.2 Liver disease in CF and modifier genes……......……….. 5.1.3.3 Summary of literature findings…………………......…... 5.2 Results……………………………………………………………......... 5.2.1 Hardy Weinberg equilibrium………………………………......... 5.2.2 Descriptive data results and study cohort grouping………......…. 5.2.3 Pulmonary disease progression: mixed effects regression on v  116 117 118 118 122 124 129 132 132 133 134 142 143 143 145 145 147 147 149 150 150 161 165 198 215 216 216 218 219 222 222 225 226 228 228 228  %predFEV1……………………......…………………………….. 5.2.4 Pulmonary disease severity: mixed effects regression on current %predFEV1…………………..................……………………….. 5.2.5 Survival analysis……………………......……………………….. 5.2.6 Effect of modifier genes on P. aeruginosa infection status…....... 5.2.7 Liver disease and CF…………………………………………...... 5.3 Discussion…………………………………………………………....... 5.4 Conclusions………………………………………………………......... 5.5 Bibliography………………………………………………………....... Chapter 6: Respiratory mucin genes as potential modifiers genes in Cystic Fibrosis……………………….......….………………………………… 6.0 Introduction………………………......………………………………... 6.1 Rationale for the investigation of mucin genes as potential modifier genes in CF………………………………......................……….……... 6.1.1 Composition of airway secretions and mucociliary clearance........ 6.1.2 Rheological properties in CF airway mucus and mucociliary transport……………………......………………………………... 6.1.3 Structure of mucin……………….....……………………………. 6.1.4 Origin of mucin secreting cells and mucin genes…………......…. 6.1.5 Mucin genes as modifiers of CF pulmonary disease……….......... 6.2 Results……………………………………………………………......... 6.2.1 Hardy Weinberg equilibrium………………………………......... 6.2.2 Descriptive data results and study cohort grouping…………....... 6.2.3 Pulmonary disease progression: mixed effects regression on %predFEV1………………......………………………………….. 6.2.4 Survival analysis……….......……………………………………. 6.2.5 Effect of modifier genes on P. aeruginosa infection status.......… 6.3 Discussion…………………………………………………………....... 6.4 Conclusions………………………………………………………......... 6.5 Bibliography………………………………………………………....... Chapter 7: Concluding chapter…………………………………......…………….. 7.0 Overall conclusions…………………………………........……………. 7.1 Conclusions for modifier genes investigated……...........…………....... 7.2 Conclusions for BCC infection sub-study…………….........…........…. 7.3 Recent relavent research published during this write-up...…………..... 7.4 Strengths and weaknesses of the thesis study………........……………. 7.4.1 Study strengths…………………………………….......………..... 7.4.2 Study weaknesses………………………………………......….. .. 7.5 Applications of our research findings……………………………......... 7.6 Future directions…………………………………………………......... 7.7 Bibliography………………………………………………………....... Appendix A: Ethics approval forms…………………………………………........  vi  228 230 231 232 232 234 238 255 266 267 267 268 269 271 273 274 277 277 277 277 280 280 282 286 302 309 310 310 315 317 322 322 324 325 325 327 331  LIST OF TABLES Table 2.1. A description of participating clinics and contribution to sample size for modifier gene analysis in the study…………...…………….……... Table 2.2. Summary of genes and polymorphisms studied and genotyping method utilized……………………………………………………....… Table 2.3. Demographic spreadsheet variables (Part 1): Clinical parameters…….  72 73  Table 2.4. Demographic spreadsheet variables (Part 2): Pathogen infection parameters and current status (deceased/alive)…...…………...……….  74  Table 2.5. Longitudinal data spreadsheet…………………………………………  75  Table 2.6. Models for statistical analysis of pulmonary disease severity and progression of α-AT polymorphisms for chapter 3………………...... Table 2.7. Models for statistical analysis of pulmonary disease severity and progression of MBL2 gene polymorphisms and the effects of chronic BCC infection (on pre and post BCC acquisition) on pulmonary disease progression for chapter 4...…………………………….……… Table 2.8. Models for statistical analysis of pulmonary disease severity and progression of SPA-1 gene polymorphism for chapter 4 ………......… Table 2.9. Models for statistical analysis of pulmonary disease severity and progression of SPA-2 gene polymorphism for chapter 4……………... Table 2.10. Models for statistical analysis of pulmonary disease severity and progression of SPD gene polymorphism for chapter 4..……………… Table 2.11. Models for statistical analysis of pulmonary disease severity and progression of GSTs and GCLC polymorphisms for chapter 5……..… Table 2.12. Models for statistical analysis of pulmonary disease severity and progression for MUC2 and MUC5B polymorphisms for chapter 6 .... Table 3.1. Clinical characteristics of study subjects stratified by α1-AT S and Z genotypes…………………………………………………….……… Table 3.2. Clinical characteristics of study subjects stratified by α1-AT 3’ G1237→A genotype…….………………………………...…………. Table 3.3. Characteristics of the sub-group of patients in the acute phase α1-AT level study…………………………………………………………… Table 4.1. The chromosomal organization of the MBL2 and the pulmonary surfactant genes (SPA-1, SPA-2, and SPD) on chromosome 10……. Table 4.2. Summary of pathogens that MBL2, SPA and SPD proteins have been shown to bind to……………………..................................………….... Table 4.3. SPA-1-SPA-2 haplotypes are presented as reported in DiAngelo and associates (73)……………...……...………………………….………..  vii  71  76  77 78 79 80 81 82 106 107 108 166 167 168  Table 4.4. Distribution of genotypes, allele frequencies and Hardy Weinberg equilibrium for MBL2, SPA-1, SPA-2 and SPD gene polymophisms...  169  Table 4.5. Distribution of genotypes, allele frequencies and Hardy Weinberg equilibrium for MBL2 gene polymorphisms based on age grouping (i.e., <25 and >25 years of age)…………………………………….…..  171  Table 4.6. Distribution of genotypes, allele frequencies and Hardy Weinberg equilibrium for SPA-1, SPA-2 and SPD gene polymorphisms based on age grouping (i.e., <25 and >25 years of age)……….…………………  173  Table 4.7. Pairwise linkage disequilibrium results for MBL2 and pulmonary surfactant gene polymorphisms………………………………....……...  175  Table 4.8. List of haplotypes for MBL2 gene polymorphisms……………………  177  Table 4.9. List of haplotypes for SPA-1, SPA-2 and SPD gene polymorphisms found using PHASE……………………….....…………………….….  178  Table 4.10. List of haplotypes for SPA-1 and SPA-2 gene polymorphisms found using PHASE………………………………………………………..  179  Table 4.11. List of haplotypes for SPA-2 polymorphisms found using PHASE….  180  Table 4.12. Clinical characteristics of study cohort by MBL2 genotype grouping.  181  Table 4.13. Clinical characteristics of study cohort by SPA-1 genotype grouping. Table 4.14. Clinical characteristics of study cohort by SPA-2 genotype grouping.  182  Table 4.15. Clinical characteristics of study cohort by SPD genotype grouping…  183 184  Table 4.16. Mixed effects models for pulmonary disease progression considering MBL2 deficiency and BCC infection status………………….…...…  185  Table 4.17. Survival analysis of the time to event (death or lung transplantation) for CF patients………………...……………..………………………. Table 4.18. Effect of innate immunity gene polymorphisms on age of first P. aeruginosa infection…….………………………………...………… Table 4.19. Effect of innate immunity gene polymorphisms on age of chronic P. aeruginosa infection…….……………………………………...…… Table 4.20. Clinical characteristics of study subjects used for investigating BCC genomovar and pulmonary disease progression (chronically and transiently infected with BCC and the control group)……… …..…... Table 4.21. Linear mixed effects models for BCC infection and pulmonary disease progression………………….……………………………….. Table 4.22. Clinical characteristics of study subjects used for investigating BCC genomovar and MBL2 deficiency and pulmonary disease progression............................................................................................ Table 5.1. Frequency of alleles and Hardy Weinberg equilibrium for GSTP1 gene polymorphism in the cross-sectional and longitudinal study cohort.......................................................................................................  viii  186 188 190  192 193  195  240  Table 5.2. Frequency of alleles and Hardy Weinberg equilibrium for GCLC gene polymorphism in the cross-sectional study and longitudinal study cohort.......................................................................................................  241  Table 5.3. Clinical characteristics of study cohort by GSTM1 and T1 grouping based on having zero, 1 or 2 null alleles across the 2 gene polymorphisms…………………………………………………………  242  Table 5.4. Clinical characteristics of study cohort by GCLC ligase gene grouping based on number of GAC repeats genotype………..…………………..  244  Table 5.5. Clinical characteristics of study cohort by GSTP1 (Ile105Val) gene polymorphism grouping based on genotype…………………...…...…. Table 5.6. Survival analysis of the time to event (death or lung transplantation) for CF patients……….………………...…………...………………….. Table 5.7. Age of first infection with pathogen P. aeruginosa…………………… Table 5.8. Age of chronic infection with pathogen P. aeruginosa……………….. Table 5.9. Clinical characteristics of CF patients based on liver disease status….. Table 6.1. Clinical characteristics of study cohort by MUC2 as grouped for analyses. The grouping was based on having one or 2 long alleles.….  246 248 250 252 254  Table 6.2. Clinical characteristics of study cohort by MUC5B genotype………...  293 294  Table 6.3. Survival analysis of the time to event (death or lung transplantation) for CF patients classified by MUC2 and MUC5B grouping….……......  295  Table 6.4. MUC2 and MUC5B do not influence susceptibility to chronic infection with P. aeruginosa infection…………..…………………..…  296  Table 6.5. MUC2 and MUC5B do not influence susceptibility to chronic infection with P. aeruginosa or BCC infection……….………………..  297  Table 6.6. Effect of MUC2 and MUC5B polymorphisms on age of first P. aeruginosa infection. ……………………..………………....................  298  Table 6.7. Effect of MUC2 and MUC5B polymorphisms on age of chronic P. aeruginosa infection……………....………...…………………………  300  ix  LIST OF FIGURES  Figure 2.1. α1-AT S and Z alleles; TaqI digestion and visualization of PCR product on an agarose gel………………………………………... Figure 2.2. α1-AT 3 prime polymorphism; TaqI digestion and visualization of the PCR product on an agarose gel……………………………….. Figure 2.3. MBL2 gene B and C alleles; the B (codon 54) and C (codon 57) alleles for MBL2 gene were detected by restriction enzymes BanI and MboII, respectively…………………………………………… Figure 2.4. The D allele and the XY promoter polymorphisms for MBL2 gene…. Figure 2.5. Site directed mutagenesis PCR/RFLP for the SPA-1 polymorphism… Figure 2.6. Exon 2 and 4 polymorphisms investigated in the SPA-2 gene……….. Figure 2.7. Site directed mutagenesis PCR/RFLP of SPD polymorphism in exon 4……………………………………………………………………… Figure 2.8. MUC2 polymorphism was examined by PCR using primers which amplified the repetitive threonine/serine/proline-rich domain……. Figure 2.9. MUC5B polymorphism was examined by PCR using primers which amplified the repetitive threonine/serine/proline-rich sudomain…… Figure 2.10. Polymorphisms investigated in GST genes P1, T1 and M1………… Figure 3.1. Comparison of pulmonary disease severity and Z and S alleles of the α1-AT gene…………………………………………………………... Figure 3.2. Comparison of pulmonary disease severity and the 3’ G1237→A mutation of the α1-AT gene…………………………………………. Figure 3.3. Alpha-1-antitrypsin levels during a pulmonary exacerbation and post-exacerbation levels during stable clinical status……………….. Figure 4.1. Schematic overview of the three pathways that activate complement.. Figure 5.1. Pathway of glutathione synthesis and metabolism…………………… Figure 6.1. The effect of normal and abnormal ion transport on airway mucociliary clearance………………………………………………... Figure 6.2. Schematic diagram of the four major protein domains in secreted gelforming mucins……………………………………………………… Figure 6.3. Schematic diagram of a mucin molecule……………………………... Figure 6.4. Distribution of genotypes for the MUC2 gene polymorphism……….. Figure 6.5. Distribution of genotypes for the MUC5B gene polymorphism……...  x  83 84 85 86 87 88 89 90 91 92 109 110 111 196 239 288 289 290 291 292  LIST OF ABBREVIATIONS  AIDS  Acquired immune deficiency syndrome  ANOVA  Univariate analysis of variance  ASP PCR  Allele-specific oligonucleotide PCR  α1-AT  Alpha-1-antitrypsin  AMP  Adenosine monophosphate  ATP  Adenosine triphosphate  ASLF  Airway surface lining fluid  ATS  American Thoracic Society  BAL fluid  Bronchoalveolar lavage fluid  BCC  Burkholderia cepacia complex  BMI  Body mass index measured as Weight/Height2 (units: kg/m2)  CEPH  Centre d’Etude du Polymorphisme Humain  CHO  Carbohydrates  COPD  Chronic obstructive lung disease  cAMP  Cyclic adenosine monophosphate  CF  Cystic fibrosis  CFTR  CF transmembrane conductance regulator  CI  Confidence interval  DNA  Deoxyribonucleic acid  dNTP  Deoxynucleoside triphosphates  EM  Expectation-maximization  ER  Endoplasmic reticulum  FEV1  Forced expiratory volume in one second  FVC  Forced vital capacity  GEN  Genomovar  GSH  Reduced glutathione  GSSG  Glutathione disulfide  GCLC  Glutamate cysteine ligase catalytic subunit  GST  Glutathione S-transferase  H. influenza  Haemophilus influenza xi  HWE  Hardy Weinberg equilibrium  Ig  Immunoglobulin  IGEPAL  Octylphenyl-polyethylene glycol  IV  Intravenous  IFN-γ  Interferon gamma  IL  Interleukin  kb  Kilobase pair  kDa  Kilo Dalton  LD  Linkage disequilibrium  LPS  Lipopolysacharide  MASP  MBL2-associated protein  MHC  Major histocompatibility complex  MBL2  Mannose binding lectin  mRNA  Messenger ribonucleic acid  ml  Milliliters  μl  Microliter  MUC  Mucin (eg, MUC2 mucin 2 gene)  NBF  Nucleotide binding fold  NADPH  Nicotinamide adenine dinucleotide phosphate  NE  neutrophil elastase  PAMP  pathogen-associated molecular patterns  %predFEV1  % of predicted forced expiratory volume in 1 second  OD  Optical density  PA infection status  Pseudomonas aeruginosa infection status  PHASE  Inferred haplotype probabilities  PCR  Polymerase chain reaction  PSS  Pancreatic sufficiency status  P. aeruginosa  Pseudomonas aeruginosa  RDS  Respiratory distress syndrome  ROS  Reactive oxygen species  S-K score  Schwachman-Kulczycki score  SSP PCR  sequence-specific priming PCR xii  SEM  Standard error of the mean  Staph. aureus  Staphylococcus aureus  SDM PCR  Site Directed Mutagenesis Polymerase Chain Reaction  SP  Pulmonary surfactant  SNP  Single nucleotide polymorphism  RAPD  Random amplification of polymorphic DNA  RFLP  Restriction fragment length polymorphism  RPM  Revolutions per minute  RNA  Ribonucleic acid  TGF-β1  Transforming growth factor-β1  TMD  Transmembrane domains  TNFα  Tumor necrosis factor alpha  UV  Ultraviolet light  VNTR  Variable number tandem repeat  yr  Year  ≠  Not equal (different)  xiii  ACKNOWLEDGEMENTS  This study was funded by research grants by the Canadian Cysic Fibrosis Foundation, Bayer/Red Cross and B.C. Lung Foundation. I acknowledge and thank the Canadian CF Foundation and Michael Smith Foundation for their financial support. A very special thank-you to: Dr. Jiang Ruan and Ms. Tracey Weir for their technical support with the laboratory work. Ms. Eugenia Wu for her statistical help and extensive assistance in the development of the pulmonary infections statistical models for analysis. Mr. Ryan Woods and Ms. Kelly Burkett for their initial statistical help on the project. Dr. Speert’s Laboratory and Ms. Debora Henry for the BCC genotyping data. I would like to to express my sincere gratitude and appreciation to the staff and directors of the CF clinics involved in the study and the CF patients who volunteered their time and consented to participate in this study. I extend my sincere appreciation to my committee members (Drs. Joel Singer, Keith Walley, John Hill) and supervisor Dr. Peter Pare for their help and support. A special thank-you to Drs. Andy Sandford and Pearce Wilcox (committee members) and Drs. George Davidson, David Speert and Yves Berthiume for their guidance. I would also like to acknowledge and thank my husband’s company Nortrak who lent me and serviced a laptop for the first year of my clinical data collection period. And I should not forget my husband Brian Charles (P.Eng.) for coming along for the ride and ‘tolerating’ my absurd work hours and me. Thank-you for sticking by me.  xiv  DEDICATION  This project is dedictated to the CF community and their commitment to research and advancement of quality of life for the patient and their families. There are certain CF patients who have participated in studies I have been involved in that have impressed me with their spirit. This thesis is dedicated to them and their families: Lloyd, Collette and Bernadette, Helena, Bob and Sheldon. A special dedication to Andrea Herzer and her family. I met Andrea when she transitioned to the Adult clinic. I got to know her very well through her participation in numerous research projects. She was a delite to be around. I would sometimes pick her up from her home for scheduled research project testing and had the opportunity to meet her mother as well. I last saw her as she was to leave for the U.S. for her lung transplant and I was expecting my first baby. We sat in her hospital room and talked about our upcoming life changing events and remember saying our good byes with a hug telling each other to take care of ourselves and promised we see each other after it all. I had Stevie after some complications, but ultimately both of us were fine. Andrea’s transplant was not successful; she was too young when she lost her battle with CF. Lastly, I dedicate this work to my children (Steven and Thomas) who I acquired during this process and who at this time may have no clue what mommy has been doing on this little black toy (laptop) which they seem to have made their mission to take possession.  xv  CO-AUTHORSHIP STATEMENT  I, Despina Frangolias, performed the following tasks on the 2003 publication presented in chapter 3 entitled: ‘Frangolias D.D., Ruan J., Wilcox P.G., Berthiaume Y., Davidson G., Hennessey R., Corey M., Tullis E., Zielenski J., Wilson W.M., Freitag A., Sandford A.. Alpha-1-antitrypsin deficiency alleles in cystic fibrosis lung disease. American Journal of Respiratory Cell and Molecular Biology. 29:390-396, 2003.’ . 1. Identification of gene to investigate was collaboration between Drs. Peter Pare, Andrew Sandford and myself, Despina Frangolias. 2. Designed the study and composed and submitted ethics forms for study approval 3. Recruited Vancouver site subjects and recruited additional centers across Canada for the study, except for Toronto (recruited by my supervisor Dr. Peter Pare). 4. Defined the study variables, created the clinical data collection sheets and collected the clinical data for Vancouver centers and entered hardcopy data sent from centers Hamilton and Montreal. 5. Performed approximatey 90% of the laboratory experiments (i.e., extraction of DNA and genotyping for the alpha-1-antitrypsin gene polymorphisms). 6. Performed 50% of statistical analyses (longitudinal data analyses and analysis of substudy data on alpha-1-antitrypsin) and Mr. Ryan Woods performed the survival analyses and confirmed my statistical analyses findings. 7.  Wrote the manuscript, which was then passed on to the co-authors for their review prior to submission.  I have written and performed all the work for Chapters 4, 5 and 6 with the advice and consulatation of my supervisor Dr. Peter Pare and committee members Drs. Andrew Sandford and Dr. Wilcox.  xvi  CHAPTER 1: OVERVIEW AND OBJECTIVES  1  1.0 INTRODUCTION Cystic fibrosis (CF) affects approximately 1 in 2500 live births in Canada and the United States (1, 2). Cystic fibrosis is a systemic disorder transmitted as an autosomal recessive trait (1). In the late 1980s the gene responsible for the CF phenotype was identified (3-6). The CF gene codes for a cellular membrane protein, the CF transmembrane conductance regulator (CFTR), which forms a chloride channel and also regulates other channel proteins. The CF gene is expressed in the epithelial cells of many organs including the sweat glands, pancreas, lungs, gastrointestinal, and reproductive tract. The mutation disrupts exocrine function by causing abnormal regulation of epithelial ion transport. Chronic and recurrent pulmonary infections and digestive disorders, infertility, and ‘salty sweat’ characterize the cystic fibrosis phenotype. The classic CF phenotype includes pulmonary disease, pancreatic exocrine insufficiency and abnormal sweat gland function. The onset of pulmonary disease in CF patients is associated with colonization of the airways, most commonly by the pathogen Pseudomonas aeruginosa (7, 8). This provokes a vigorous local inflammatory response that prevents the spread of the infection beyond the lung, however by doing so it likely contributes to lung destruction(7-9). Precisely how the CF electrolyte transport defect leads to persistent lung infection and inflammation is still not well understood. With advances in treatment, patients who have CF are living beyond their adolescent years into adulthood. The median age of survival has increased to well into the third decade. The large variability in disease progression and degree of organ involvement has become even more evident with the increased survival. Only a portion of this variability can be explained by the individual’s CFTR genotype. Although some CFTR genotypes are associated with less severe disease, patients possessing the same genotype show great variation in disease severity and progression (1). Over 900 mutations have been identified in the CFTR gene, but ΔF508 is the most common(10). Environment plays a significant role in explaining some of the variability in disease severity and progression. Environmental influences include the availability of therapeutic modalities as well as the psychosocial issues such as the patient’s compliance to therapy and the patient’s coping mechanisms with their illness. However, other genes may play a significant role in explaining this wide range of clinical outcomes and disease progression. Variants in other secondary genes may act in concert to positively 2  or negatively contribute to disease severity and progression in CF. An understanding of the complex interaction of other genes with the CFTR gene may help us gain a better understanding of the cystic fibrosis phenotype and develop specific therapies. This chapter will focus on a brief historical review of the discovery of the cystic fibrosis gene, and a description of the gene’s structure, function, and the effect of the CFTR defect on the main exocrine organ it affects, the lung. A review of the literature on possible modifier genes in CF will be explored and this section will conclude with the study purpose, rationale, and main and secondary hypotheses. 1.2 THE CF GENE 1.2.1 Discovery of the Cystic Fibrosis Gene From early on it was known from population studies that cystic fibrosis was an autosomal recessive disorder and that it affected a number of organs, including the lung airways, pancreas and sweat glands. Knowledge that this disorder was recessive came from the observation that offspring who had CF were born from apparently unaffected parents. The basic defect is associated with decreased chloride ion conductance across the apical membrane of epithelial cells and the defect persisted in cultured cells derived from numerous epithelial tissues suggesting that the CF gene was expressed in these cells. It was also known that the defect lay in epithelial cells due to the observation of increased chloride in sweat, exocrine pancreatic obstruction leading to pancreatic insufficiency and intestinal obstruction (meconium ileus in newborns). Unlike other genetic disorders where the amino acid sequence of the protein that was defective was known (e.g. haemoglobinopathies, phenylketonuria and alpha-1-antitrypsin), in CF the basic protein defect resulting in the disorder was still unknown until the gene was discovered using positional cloning. From as early as 1986 it was hypothesized that the defect was probably due to a failure of an outwardly rectifying anion channel. Research in these early years focused on candidate genes. These included a sodium inhibitory factor identified from CF saliva (11), ciliary dyskinesia factor (12, 13), CF antigen on chromosome one and HLA (14). These studies failed to show any positive associations. In the late 1980s linkage analysis was used to search for the CF gene. The first step to identify the primary defect in CF was to establish the chromosomal localization of the 3  disease locus. Development and refinement of polymerase chain reaction during this era facilitated this search; restriction fragment length polymorphism (RFLP) markers were used to screen DNA that was obtained from families with 2 or more affected individuals to. The relatively straightforward CF phenotype, knowledge that it was a recessive disorder (since affected offspring were born to normal parents), and the large number of CF families facilitated this work. Early work did not yield candidates. In the mid 1980’s Eiberg and associates reported linkage between the CF locus and a polymorphic locus which controlled activity of the serum aryl esterase paraoxonase (PON), however although the location of the CF gene was narrowed down to one third of the genome, the chromosomal location of PON was not known (15). The location of the CF gene was narrowed down to one percent of the genome by Tsui and associates(3) who showed that a DNA marker called D0CRI-917 was genetically linked in a set of 35 families and was also linked to the PON locus, which by independent evidence had been linked to the CF locus. The search for the CF gene had been narrowed down to 30 million base pairs; the estimates of the genetic distances were calculated to be 5 centimorgans between the DNA marker and PON and 15 centimorgans between the DNA marker and the CF locus(3). In 1985 linkage was established between CF and polymorphic markers (i.e., D7S15, MET, D7S8, COL1A2), all known to be located between bands 22 and 31 on the long arm of chromosome 7 (3-6, 16-21). This still represented a very large region and two approaches were used to pinpoint the CF gene. One approach was to proceed to isolate the CF gene by directly looking for sequences that were preferentially expressed by epithelial cells and the second approach was to define the physical map of the region. Genetic and physical mapping studies showed the order and distance of the four markers that had been identified (i.e., MET-D7S340-D7S122-D7S8 with distances between them of 500, 10 and 980 kb, respectively) (18). Allelic and haplotype associations were shown between the CF locus and closely linked DNA markers. Beaudet and associates investigated recombination events between CF and these linked markers in 100 CF families with 2 or more affected children, the order between these markers and the CF gene was established and flanking markers were MET and pj3.11(D7S8) (22). Pancreatic insufficiency is a common pancreatic abnormality shared by many CF patients but not all. Kerem and associates used two clinical subgroups of CF 4  patients based on pancreatic sufficiency status (i.e., pancreatic insufficient (PI) and pancreatic sufficient (PS)) and showed that pancreatic dysfunction could be explained by different mutations on the basis of family studies and haplotype data (23, 24). PI CF patients (85%) were more homogeneous than PS patients (15%); however these studies were based on linked DNA markers whose exact relation to CF was not entirely certain (24). Chromosome walking and jumping were used to link DNA molecules at great distances from one another (400-500kb). This process was facilitated by the development of pulsefield gel electrophoresis and the discovery of additional restriction enzymes. A restriction map of the region was constructed which localized the CF locus with a number of closely linked markers. Each one of these closely linked markers was used as a starting point for a series of chromosomal walking and jumping experiments in order to clone and sequence large DNA regions (19). These cloned regions (which were overlapping) were used to look for candidate coding sequences that would be conserved in other animal species (rodent, bovine, mouse, and chicken). They found one such sequence that was conserved but it was fairly small (only 113bp long) and consequently represented only a small section of the gene. Ultimately this coding sequence was identified as exon 1 of the CF gene. By successive screening with cDNA libraries generated from a number of cell lines (i.e., a cell carcinoma line, normal and CF sweat gland cells, pancreas and adult lung), they were able to isolate an additional 18 clones. Ultimately Dr. Tsui and his collaborators were able to deduce the coding region of the CF locus from the overlapping DNA clones (3). Together these clones spanned 6.1 kb and encoded a protein which was 1480 amino acids long (3). To visualize the transcript of the CF gene, mRNA samples were prepared from various tissues. They found high mRNA levels in tissues which were affected by CF (e.g. nasal polyps, pancreas, lung, sweat glands, colon, placenta, liver, parotid gland) and no detectable mRNA expression in tissues not affected by the disease, such as the brain and adrenal gland. They consequently surmised that the expression of the CF gene occurred in many of the tissues examined, with higher levels in those tissues severely affected by CF (20). RFLPs associated with the CF locus were used to establish the relation to CF of the DNA segments isolated from the chromosome walking and jumping experiments and family studies. Families where cross-over events between CF and other flanking DNA markers had previously been discovered were used (24). The recombination breakpoints were localized 5  in 2 families that were informative for the DNA markers tested and the CF gene localized to a region between 2 markers (i.e., KM19 and D7S424) (25). Linkage disequilibrium was detected for markers that were close to the CF gene (markers within 300kb interval) and not for the markers further away. A similar conclusion was made by the investigators when they utilized haplotype analysis (24). For these analyses, investigators used PI patients because they appeared to be more homogeneous genetically. Most of these patients would later be shown to be carriers of the most common and severe mutation in CF, deltaF508 (∆F508). Comparisons between cDNA clone sequences derived from CF and unaffected individuals showed the most striking difference was a 3 base pair deletion, which resulted in the loss of a phenyalanine residue (∆F508 mutation). Sixty eight percent of CF chromosomes in the general patient population had this 3 base pair deletion. In contrast none of the unaffected individuals had this deletion. Extended haplotypes based on 23 DNA markers were generated for the CF and wild-type (i.e., unaffected) chromosomes in the collection of families previously used for linkage analysis. Five major groups of wild-type and CF haplotypes were identified within the region of the CF gene, with one of them associating with the most frequent CFTR mutation, the ∆F508 mutation. The protein structure was next characterized and from the cDNA clones the characteristic features of this protein were described to be two repeat motifs which included a transmembrane spanning domain and sequence resembling ATP-binding folds. From these characteristics and similarities with other membrane associated proteins it was predicted that the CF gene product was likely involved in the transport of ions across a membrane, although it was unclear how CFTR was involved in the regulation of ion conductance across the apical membrane of epithelial cells. Validation that the gene identified was the correct gene responsible for the disease came with the in vitro work showing the association of mutant CFTR and defective chloride transport(26-28). 1.2.2 Structure, function and localization of the CFTR gene The CFTR gene is a 250kb gene and has 27 exons that are transcribed into a 6.5kb mRNA, which is translated to a 1480 amino acid product. The protein is a member of the ATP binding cassette family. The protein is composed of two repeat motifs, each with a 6 transmembrane (loops) domains (TMD) and an intracellular nucleotide binding fold (NBF) 6  separated by an intracellular hydrophilic regulatory domain (R). The protein is a chloride channel activated by cyclic AMP mediated protein kinase A phosphorylation of the R domain and ATP hydrolysis by the NBFs. The mature CFTR protein forms an apical epithelial channel. The two transmembrane domains form the pore of the channel. The regulatory domain and the NBFs are situated intracellularly. The regulatory domain, which is encoded by exon 13, contains phosphorylation sites for protein kinase A and for protein kinase C. Phosphorylation of the regulatory domain results in a conformational change in this domain which moves it away from the pore of the channel and thus allows the flow of chloride through the channel (i.e., phosphorylation of the regulatory domain results in the opening of the channel). ATP hydrolysis occurs on the NBFs; hydrolysis of NBF-1 opens the CFTR channel and ATP hydrolysis of by NBF-2 closes the CFTR channel. The activity of the channel is regulated by intracellular cAMP levels. Epithelial cells contain receptors and respond to beta-adrenergic agonists, prostaglandins, adenosine and vasoactive intestinal peptide. Cellular calcium concentrations are controlled by bradykinin, substance P, leukotrienes and nucleotides such as ATP. These extracellular signals regulate the two main second messengers: cAMP and calcium. Increase in cAMP activates cAMP dependent protein kinase, which phosphorylates and thus activates CFTR chloride channels, basolateral Na/K/Cl cotransporters and likely some of the basolateral K channels. In contrast to the cAMP mediated chloride secretion, intracellular events that mediate chloride secretion stimulated by increased cytosolic calcium are less clear. The increase in membrane anion conductance by the calcium activated chloride channel is hypothesized to be mediated by a calcium/calmodulin dependent protein kinase II (2, 29). One possible mechanism described for the channel in intestinal epithelial cells is that increases in cytosolic calcium are stimulated by acetylcholine through the M3 muscarinic receptor which increases conductance of a basolateral membrane potassium channel (distinct from the cAMPregulated potassium channel) which results in membrane hyperpolarization and creates an increased driving force for chloride exit through these calcium activated (alternative) chloride channels (2, 30). It is also thought that CFTR functions in ATP efflux and concomitant regulation of the calcium activated chloride channel and as a cAMP dependent negative regulator of the 7  epithelial Na channel. Loss of this regulatory function may account for the abnormal Na transport characteristic of tissues which are affected by CF; however it is not clear if this is a consequence of direct CFTR interaction with the epithelial sodium channel or a secondary effect mediated by other regulatory proteins. The promoter region of the CFTR gene is 3.5kb and is a GC rich region with a major transcription start site and multiple minor transcription start sites. A number of SP-1 and SP-2 binding sites have been identified as well as sequences for cAMP and glucocorticoid response elements. 1.2.3 Classification of CFTR mutations Close to 1000 mutations in the CF gene have been identified (10, 31). Defective CFTR can result in defective protein production, defective processing and degradation in the endoplasmic reticulum (ER), or a defective channel pore or gating properties. CFTR mutations are grouped into 5 mechanistic classes based on demonstrated or predicted molecular dysfunction. Class 1: Includes mutations which cause defective synthesis. These mutations prevent transcription into a stable full length mRNA, leading to defective protein products. There is little or no full length protein that is produced and this causes loss of CFTR function. Class II: Includes mutations with defective maturation of the protein (i.e., protein processing). CFTR mRNA is formed, but the protein fails to mature and does not traffic to the cell membrane. ∆F508 belongs to this class since this mutation disrupts the proper folding of the CFTR protein such that the protein is rapidly degraded before leaving the endoplasmic reticulum. At sub-physiological temperatures (23-30degrees Celsius) or upon chemical modification (e.g. with glycerol) ∆F508 CFTR is folded and processed and is functional at the apical membrane. Class III: Includes mutations that have a defect in the channel regulation (blocked activation). CFTR protein is produced and traffics to the cell membrane but fails to respond to cyclic AMP stimulation. CFTR mutants are fully processed and properly localized but are not activated by cAMP. A number of mutations in the NBF1 and NBF2 belong to this class and these mutations disrupt binding and hydrolysis. Class III mutations can also affect phosphorylation sites in the R domain, but these sites are thought to be redundant and the domain confirmation may be sufficiently flexible to accommodate a mutation in the R domain without significant functional consequences. 8  Class IV: Mutations in this category are defective in conductance through the channel and represent milder mutations. CFTR protein is produced and traffics to the apical cell membrane, however mutations result in decreased chloride conductance due to altered ion conductance. Some of the mutations in this class (such as R117H, R334W, R347P) occur in the TMDs that form the channel pore and reduce the amount of chloride current by altering the rate of ion flow (R347P) or by changing the amount of time the channel remains open (R117H). The mutation P574H occurs in NBF1 and also alters the duration of channel opening. Class V: These mutations are the result of abnormal splicing (decreased abundance). Mutations influence the quantity of full length mRNA transcript and protein required for normal function. This class includes promoter mutations, mutations that contribute to alternate splicing or mutations that cause inefficient protein processing and consequently reduced levels of functional protein. 1.2.4 Cellular mechanism of ion transport by normal and CF airway epithelia There are at least three channels operating on the apical epithelial membrane: the epithelial sodium channel, the CFTR channel and the calcium activated chloride channel. The following channels are found on the basolateral membrane of airway epithelial cells: the sodium-potassium ATPase (Na+-K+-ATPase) pump, the sodium-chloride-potassium (Na+/2Cl-/K+) co-transporter, and the potassium pump. In normal airway epithelial cells sodium is absorbed through the sodium channels on the apical membrane and extruded through the basolateral membrane by the Na+-K+-ATPase pump. This transport of ions results in the movement of water by osmosis, thus this cycle of Na+ absorption also results in water absorption. These processes operate to limit the volume of the periciliary fluid layer.  Chloride also passes paracellularly following sodium  movement in an attempt to maintain electrical neutrality. Ion transport mechanisms also exist to re-hydrate the airway surface. Low intracellular Na+ concentration created by the Na+-K+-ATPase triggers movement of Na+ into the cell from the basolateral surface through the sodium-chloride-potassium (Na+/2Cl-/K+) co-transporters. Chloride thus enters the cell down a favorable Na+ gradient created by the ATPase and exits the cell down this created gradient toward the luminal surface through chloride channels. Water follows the chloride movement enabling rehydration of the airway surface. Chloride transport into the airway 9  lumen occurs via the CFTR channel, but this channel also stimulates extrusion through another apical chloride channel named the calcium activated chloride channel. The CFTR channel also controls activity of the epithelial sodium channel and inhibits sodium movement through the epithelial sodium channel by switching the response to cAMP from an increasing to a decreasing channel opening probability. Under basal conditions sodium absorption predominates, but during periods of activity during which the potential for airway dehydration exists, chloride secretion can be activated to maintain periciliary fluid volume. In CF this regulatory process is lost thus producing the increased sodium absorption; the epithelial sodium channels have a greater open probability that results in increased net sodium absorption, but also results in increased water absorption via osmosis and increased paracellular movement of chloride. Sodium exits the cell through the basolateral Na+-K+ATPase pump. During periods of activity, chloride enters the cell through the basolateral Na+/2Cl-/K+ co-transporters; however the Cl- cannot be extruded into the airway through the defective CFTR channel. Sodium absorption (via epithelial sodium channel) and chloride extrusion (via calcium activated chloride channel) are both also compromised in CF. 1.2.5 Model for the development of airway disease in cystic fibrosis CFTR controls fluid secretion from the apical surface of epithelial cells since the movement of chloride and sodium determine the osmotic forces of water movement. There are two competing theories to explain how these changes in airway surface lining fluid cause the abnormalities which are responsible for the pulmonary phenotype in CF. Boucher’s group (7, 8) postulates that the volume of the airway surface fluid is too little (or insufficient) for two reasons: ƒ  The defective CFTR cannot properly regulate fluid production.  ƒ  Fluid absorption is enhanced by the unregulated epithelial sodium channel.  Smith and associates (32-34) postulate that the volume of the airway surface lining fluid is the same in CF and non-CF patients, but that sodium and chloride concentrations are what differ. In their studies, measurements made on cultured normal and CF epithelial cells indicated differences in the sodium and chloride concentrations between normal and CF 10  cells, with higher salt concentration in CF cells. It is this high salt environment that is postulated to inactivate naturally occurring antibiotics. The hypothesis most supported is the depletion of the airway surface lining fluid layer. It is hypothesized that airway destruction and obstruction in CF is due to volume depletion, decreased mucocilliary clearance and resultant infection (2, 35). The cause of the reduced ASL fluid is hypothesized to be due to the CFTR defect which results in accelerated sodium transport, failure to regulate cAMP dependent chloride secretion which leads to reduced periciliary lining fluid and failure of mechanical mucus clearance. On normal airway epithelia a thin layer of mucus resides on top of the periciliary fluid layer. The mucous layer is capable of donating and accepting water and when there is depletion of liquid from the airway surface the mucous layer donates fluid until its height/volume ratio is reduced to approximately 50 percent and thereafter liquid is donated by the periciliary liquid layer(2, 36, 37). In the early stages of volume depletion, water is donated without it being replenished from the mucous layer to the periciliary liquid layer. A point is reached when the mucous layer can no longer donate liquid to the periciliary liquid layer and the periciliary liquid layer is absorbed. Depletion of the periciliary liquid layer has a number of effects. Depletion of the periciliary liquid layer prevents the cilia from extending normally and beating and therefore compromises ciliary-dependent mucus clearance. The concentrated mucins in the mucous layer become more adhesive due to the volume depletion and this increased viscosity of the mucous layer also contributes to hindering ciliary-dependent transport. As a result of the periciliary liquid depletion there is adhesion of the previously mobile mucous layer to the cell surface and this allows the contact of mucus with the cell surface glycocalyx which is believed to effectively adhere the previously mobile mucous layer to the airway surfaces(2, 8, 36). This has two effects (36): ƒ  It abolishes mucus clearance due to failure of the cilia to extend and beat.  ƒ  It degrades the ability of cough to remove mucus due to the physical adhesion of mucus to airway epithelia surfaces.  11  There is continued mucus secretion by goblet cells into immobile airway surface mucous plaques. The net effect of this accumulation is to extend the height of the mucous plaques and/or create mucous plugs. These mucous plaques in CF airways are characterized by: ƒ  Increased mucus viscosity, which decreases the ability of soluble antimicrobial factors (e.g., lysozyme, lactoferin) to diffuse within the mucous plaque.  ƒ  Increased viscosity that degrades the capacity of neutrophils to penetrate this niche.  ƒ  Increased height of the mucous plaques on airway surfaces coupled with accelerated oxygen consumption by CF airway epithelia that fuels an increased sodium transport and creates hypoxic niches at the bases of the mucous plaques.  Boucher and associates have shown that motile Pseudomonas aeruginosa is particularly suited to this environment and the pathogen adheres to respiratory mucus (8). The increase in CF airway epithelial oxygen consumption is due to the absence of the CFTR’s normal inhibitory activity on the epithelial sodium channel. The consequential accelerated sodium absorption on the apical membrane is fueled by an increased turnover rate of the ATP consuming Na+-K+-ATPase pump leading to the two- or three-fold increase in epithelial oxygen consumption(8). The persistent mucin secretion into stationary mucus generates these thick airway plaques and/or plugs. The combination of thickened mucus and raised oxygen consumption of CF airway epithelium generates steep oxygen gradients within adherent mucin. Bacteria deposited on the thickened mucus can penetrate into these hypoxic zones. The excessive volume absorption having obliterated normal rotational mucus transport allows motile P. aeruginosa to penetrate these static plaques and grow in hypoxic/anaerobic CF mucus. P. aeruginosa responds to hypoxic mucus with alginate production and biofilm-like macrocolony formation, which is the dominant phenotype of P. aeruginosa in CF airways (2). Clearance of mucus is likely a key feature of innate lung defense and a fundamental defect leading to chronic CF pulmonary infections is failure to clear mucus that contains bound bacteria from the airway lumen. 1.2.6 Determinants of pulmonary disease severity in cystic fibrosis It would seem that at some point early in a CF patient’s life infection sets up an inflammatory response. Precisely how the CF electrolyte transport defect leads to persistent 12  lung infection and inflammation is still not well understood. The most common CFTR mutation (i.e., ΔF508) is generally associated with more severe pulmonary disease (38, 39). However, even among ΔF508 homozygous individuals there is a wide variation in disease severity (38) suggesting that there are other genetic and/or environmental factors that influence pulmonary function. In contrast, a good correlation has been shown between pancreatic sufficiency and CFTR mutation (40). A series of mutations have been identified which are classified as mild for pancreatic status (i.e., pancreatic sufficiency) (41) and these mutations are predominately missense mutations involving exons 5, 7, and 17 of the CFTR gene (42, 43) which result in amino acid changes in the TMDs (44). The lack of a clear association between CFTR genotype and pulmonary phenotype suggests that additional factors are involved and these could be genetic modifiers. Santis et al. investigated the level of lung function (FEV1 % predicted) in pairs of siblings who had CF and showed significant association in %predFEV1 (r=0.84) between the siblings (45). Concordance of pulmonary disease severity in first-degree relatives is suggestive evidence for shared genetic factors. In addition, residual chloride secretion in intestinal tissue was more concordant in monozygous than in dizygous twin pairs (46). Some of the heterogeneity observed in pulmonary disease can be explained by CFTR mutations; that is variability in disease severity based on the class of the CFTR mutation. There are two other factors which also influence pulmonary disease severity. One such factor is the environment, specifically availability of therapeutic modalities to the patient, patient compliance to therapy recomendations and Burkholderia cepacia complex (BCC) cross-infection. The other factor is genes other than CFTR that may contribute to pulmonary disease progression. The concept of modifier genes as predictors of pulmonary disease heterogeneity and recent modifier gene studies will be reviewed below. 1.3 CANDIDATE MODIFIER GENES INVESTIGATED Currently, one modifier gene locus has been identified for meconium ileus. In a genomewide search for modifier genes in a CFTR-deficient mouse model of meconium ileus, a major modifier locus was detected near the centromere of mouse chromosome 7 (47). The orthologous region in humans is chromosome 19q13. A multicenter study of 197 CF sibling 13  pairs confirmed linkage of meconium ileus to this region (10). The gene responsible for this linkage has yet to be identified. This locus was not shown to modulate pulmonary disease severity either in mouse models or in humans. 1.3.1 Alpha-1-antitrypsin gene Alpha-1-antitrypsin (α1-AT) was one of the first candidate modifier genes investigated in CF. Studies had shown extremely high levels of neutrophil elastase (NE) in the airways of CF patients, which indicated that there is an imbalance between α1-AT and elastase in the airways of patients with CF. The inflammatory process in response to pulmonary infection in CF airways is characterized by a massive influx of neutrophils (48).  Neutrophils  represent less than 5% of the cells recovered in bronchoalveolar lavage fluid in normal individuals but in adults and children (1-5 years of age) who have CF, neutrophils may comprise up to 95% of the cell population (49). Neutrophils contain a number of proteolytic enzymes one of which, neutrophil elastase (NE), has been implicated in excessive pulmonary damage observed in cigarette smokers and in CF patients. Elevated levels of NE have been reported in the sputum of patients who have CF (50, 51). NE is capable of causing direct lung damage by hydrolyzing all the major connective tissue proteins that make up the lung and airway matrix. NE also affects adherence of P. aeruginosa to the airway epithelium (52-54), impairs complement-mediated phagocytosis (particularly of P. aeruginosa) (55-58), stimulates interleukin-8 (IL-8) secretion (59) and stimulates mucus production (58, 60, 61). In normal hosts, the actions of NE are prevented primarily by alpha1-antitrypsin (α1-AT), a serine protease inhibitor that binds to NE and inhibits the breakdown of elastic tissue in the lung. Normal to elevated levels of α1-AT have been reported in the airway secretions (49) and plasma (49, 54) of CF patients. Elevated levels of α1-AT have been reported during pulmonary infections in this patient population (53). The α1-AT MZ genotype had been shown to be a risk factor for COPD (62). This candidate gene has been investigated as a possible modifier gene in CF in the last decade with conflicting results concerning the role that α1-AT may play in pulmonary disease progression in CF. It is reasonable to hypothesize that individuals who have lower than normal levels of α1-AT may be at increased risk for lung damage. Several mutations of the  14  α1-AT gene result in a deficiency of this antiprotease. There is also evidence that α1-AT genotype influences the acute phase response (55). Doring and colleagues (57) showed no association between α1-AT S and Z alleles and pulmonary disease severity in CF but they did show an earlier age of onset of P. aeruginosa infection in CF subjects with these deficiency alleles (6 out of a total sample of 215). Mahadeva and associates questioned this association and in fact showed that CF patients who were heterozygous for the S and Z alleles (19 out of a total sample of 147) had higher levels of pulmonary function than wild type individuals (63). In another study, the same authors showed that the α1-AT Z and S deficiency alleles were not more prevalent in those CF patients with severe pulmonary disease (dead or lung transplanted CF patients)(64). Meyer and associates more recently showed similar age of onset of P. aeruginosa infection in MZ (N=5) or MS (N=16) CF patients compared with matched wild-type CF patients (65). Another polymorphism in the α1-AT gene studied in pulmonary diseases is the A allele of the polymorphism (G1237A) in the 3' region of the gene. Kalsheker and associates showed that the 3’ mutation was associated with COPD (66). Morgan and colleagues provided in vitro evidence that the association with COPD may be due to deficiency in the α1-AT acute phase response (55).  Sandford and associates (67), however, did not find that the 3’  mutation attenuated the acute phase rise in α1-AT in their study of patients undergoing openheart surgery. Similarly, Madadeva and associates showed that the 3' mutation had no effect on α1-AT levels in CF patients (63), whereas Henry and associates (68) showed less severe pulmonary disease and fewer infective pulmonary exacerbations over 2 years in CF patients who were heterozygous for the A allele. These data suggest that heterozygotes may have a slower disease progression. In CF cohorts, the association of α1-AT genotype and pulmonary disease severity is unclear (57, 63, 64), the main limitation of these studies being their small sample sizes (number of subjects with the deficiency alleles 6-20) and therefore the high possibility of type 2 error (false negative)(69).  15  1.3.2 Innate immunity genes The majority of children diagnosed as having CF develop chronic pulmonary infection, most often with P. aeruginosa, in late childhood. There is a positive association between age of first colonization with this pathogen and ultimate respiratory failure in CF patients. Genes regulating the first defense to bacterial and viral infections, especially during the first years of life can have a major impact on survival. There are at least four potential genes, which may increase the risk for developing bacterial and viral infections during infancy. Homozygosity or compound heterozygosity for mannose binding lectin (MBL2) variants has been shown to predispose young children to pulmonary infections with pathogens commonly cultured from CF patients (70, 71). MBL2 is synthesized by the liver and released upon IL-6 stimulation; it provides protection against bacterial and viral infections especially in infancy before adaptive immunity becomes established and provides nonspecific defense against pathogens continually. MBL2 protein binds to mannose containing proteins or carbohydrates on bacterial and viral surfaces which are then recognized by alveolar macrophages which induce phagocytosis and release inflammatory cytokines. The MBL2 gene is located on chromosome 10 and there are 3 missense polymorphisms at codons 54, 57 and 52 of exon 1. Further variation due to a polymorphism in the promoter region at position –221 has been shown to cause down regulation of gene expression and consequently lower serum MBL2 concentrations (72). Another polymorphism in the promoter region (at position at –550) is also associated with reduced MBL2 expression (73). Garred and associates (72) showed reduced serum MBL2 levels in heterozygotes for the coding polymorphisms, no MBL2 production in homozygotes for the variant alleles and lower gene expression with the promoter region polymorphism in a CF cohort. They also showed reduced lung function, earlier infection with P. aeruginosa, and also showed a trend suggesting that CF patients who had the variant structural and promoter alleles were more likely to become colonized with BCC. A closer look at their data however shows that the main difference in pulmonary function over the eight year longitudinal interval studied was that the MBL2 deficient group had a lower average lung function and one could not see a progressive decline in pulmonary function, in fact the graphs in their study showed a steady increase in pulmonary function in both groups which the investigators contributed to changes in pulmonary function testing equipment. Further evidence of increased susceptibility to BCC infection and MBL2 deficiency has come from Davies and colleagues 16  (74) who showed that MBL2 binds to BCC and activates complement. Specifically, they obtained BCC isolates from 10 CF patients and showed increased binding (61%) of MBL2 to BCC compared with non-mucoid strains of P. aeruginosa (2.9% binding). Gabolde and associates (75) in their case-control study of CF patients, matched for age and sex with normal healthy controls, showed a significant decline in lung function in those homozygous or heterozygous for the variant structural alleles. In all studies, sample size limited the power of their studies. Three other innate immunity genes which are proposed as candidate modifier genes are located in close proximity to MBL2; pulmonary surfactant protein genes SP-A1, SP-A2, and SP-D are located in succession on the long arm of chromosome 10. The SP-A and SP-D genes have been localized at 10q22-q23 and MBL2 placed at 10q21. As these genes are located in close proximity to one another, it cannot be dismissed that any association found between disease severity and MBL2 may be due to linkage disequilibrium with the pulmonary surfactant genes. The SP-A1 and SP-A2 genes are separated by an SP-A pseudogene. The SP-A1 gene is flanked by the SP-D gene. The pulmonary surfactant proteins SP-A (SP-A1 and SP-A2) and SP-D activate complement through the alternative pathway of complement activation. Unlike MBL2, SP-A and SP-D are synthesized within the lung predominately by alveolar type II cells and Clara cells. The pulmonary surfactant genes function as opsonins by binding to specific receptors on macrophages (76) and are also suggested to have both pro- and anti-inflammatory properties (77, 78). These surfactant proteins have been shown to bind to a large number of pathogens, including common pathogens found in CF sputum, although they differ in their interaction with the pathogens. Vandivier and associates (79) showed impaired clearance of apoptotic cells from CF airways and reduced clearance of apoptotic cells in SP-D knockout mice. The authors speculated that the severity of the inflammatory response may be enhanced by defective clearance due to SP-A or SP-D deficiency. Results from knockout mice have shown that SP-A deficiency and SP-D deficiency are associated with pulmonary infection (80-82). Polymorphisms within the SP-A locus have been associated with pulmonary disease including respiratory distress syndrome (83) and chronic obstructive pulmonary disease (84). SP-A and SP-D 17  polymorphisms (9 polymorphisms in the SP-A genes and 2 in the SP-D gene) have shown negative results in association studies of acute respiratory distress syndrome (85). In summary the association of MBL2 genotype and pulmonary disease severity is unclear in CF cohorts (72, 74, 75, 86). The pulmonary surfactant genes are good candidates as modifier genes in CF. 1.3.3 Antioxidant genes Glutathione (GSH) is an important factor in the prevention of oxidant-induced lung injury. Reduced GSH has been shown in a number of pulmonary diseases including idiopathic pulmonary fibrosis (87-89), acute respiratory distress syndrome (90, 91), COPD (92), idiopathic respiratory distress syndrome (93) and in HIV-positive patients (94-97). Decreased GSH levels have also been shown in CF patients (98) and in the CF mouse (99). There are two groups of enzymes in the GSH synthesis pathway which have been investigated as affecting GSH function. The glutathione S-transferases (GSTs) are involved in detoxification of hydroperoxides by conjugating them with glutathione (H202; the reaction is: 2 GSH + H202 →GSSG + 2H20) (100). There are several families of GSTs which have been identified in humans and are referred to as α, μ, π, and θ (101). GSTM1, GSTP1 and GSTT1 genes are polymorphic. The μ class GST, GSTM1, is located on chromosome 1 and the π class, GSTP1 is located on chromosome 11, and both are synthesized by the liver (102). GSTT1 is located on chromosome 22 and is synthesized by the liver and red blood cells. A rate limiting step in the glutathione synthesis pathway is the enzyme glutathione cysteine ligase. The catalytic subunit of the gene contains a polymorphic GAG trinucleotide repeat. In humans three alleles containing 7, 8, or 9 repeats have been identified. Roum and associates (103) showed reduced levels of GSH in airway surface lining fluid, but also showed reduced levels of oxidized GSH, thereby suggesting that the low GSH was likely due to low GSH production and not excessive GSH oxidation. The CFTR is permeable to oxidized GSH, and defective CFTR results in altered permeability to GSH. Aerosolized GSH has been shown to reduce airway inflammation in CF (98). Transfection of normal CFTR or synthetic chloride channels to CF epithelial cells restores GSH secretion (104, 105). 18  There are two functionally active GSTM1 alleles and a null allele which produces no protein. Those homozygous for the null allele have been shown to have increased risk for some cancers, pulmonary emphysema and chronic bronchitis.  Hull and Thomson  (106)showed that CF patients who have the null allele exhibited somewhat worse lung function, clinical scores of disease severity and nutritional status. These patients were also more likely to be positive for P. aeruginosa infection(106).  Baranov and associates  (107)earlier showed CF patients who were homozygous for the null allele were more likely to die before the age of five years. Of CF patients who were pancreatic sufficient and manifested pulmonary symptoms in their study 51.1% were homozygous for the null allele (total sample N=194)(107). Henrion-Caude and colleagues (108) showed a significant association between CF liver disease and a GSTP1 polymorphism (exon 5 Ile105Val) in 106 CF children. No association with liver disease was shown for GSTM1 for this CF cohort. 1.3.4 Pro-inflammatory and anti-inflammatory mediators Another group of studies have focused on investigating inflammatory and anti-inflammatory genes as candidate modifier genes in CF. CF is characterized by chronic inflammation focused on the airway lumen. Arkwright and associates (109) investigated inflammatory (tumor necrosis factor alpha (TNFα), interferon gamma (IFN-γ)), and anti-inflammatory (interleukin 10 (IL-10), transforming growth factor-β1 (TGF-β1)) cytokine genes as candidate modifier genes in 261 CF patients. They showed the high producer TGF-β1 (+915) genotypes to be associated with more severe pulmonary disease progression and earlier age of colonization with P.aeruginosa (109). In an earlier study Arkwright and associates (110) showed more rapid pulmonary disease progression in CF patients homozygous for the ΔF508 mutation (N=171) and another TGF-β1 polymorphism (Leu+869Pro). This same polymorphism was investigated in the group’s more recent study but no association was shown with pulmonary disease decline (109). Arkwright and associates did not show a significant association for TNF-α and decline in pulmonary function (109), which has been previously reported in a cohort of 53 CF children (106). Major histocompatibility complex class II genes have been investigated as possible modifier genes in CF. Aron and colleagues (111)investigated the contribution of this group of genes to the inflammation seen in this population. MHC class II antigen has been shown to be 19  expressed in CF nasal polyps (112). The investigators reported a significant association of HLA-DR7 with the presence of P. aeruginosa infection implying that the HLA class II region may modulate susceptibility to this pathogen (112). However, they did not show an association of HLA-DR alleles with level of lung function. In summary, study sample sizes are small and further study to replicate results on independent study sample of larger sample size is necessary. 1.3.5 Additional modifier genes Another family of genes which we propose as candidate modifier genes are the mucin genes. Twenty mucin genes that have been characterized, six (MUC2, MUC4, MUC5AC, MUC5B, MUC8 and MUC18) are known to be expressed in the respiratory tract (113-115). The main characteristic of all mucin genes is the presence of a central domain composed of a variable number of tandem repeats. Each repeat unit contains multiple sites for O-linked glycosylation. There is marked inter-individual variation in the number of glycosylated side chains on the mucus proteins MUC4 and MUC5AC due to inherited variability in the number of repeats. The MUC4 gene is one of the most polymorphic and the number of tandem repeats varies from 145-395 between individuals (116). MUC5AC is also highly polymorphic (117) and the protein forms a major constituent of respiratory mucin (118). Although correlation of the degree of glycosylation and the visco-elastic properties of mucus from individual subjects has not been studied it is likely that such a relationship does exist. Increased glycosylation would be expected to increase mucus visco-elasticity and slow mucociliary clearance of purulent broncho-pulmonary secretions. There is increased secretion of mucus in CF which is stimulated by inflammation. There is also hypertrophy of submucosal glands and goblet cell hyperplasia and metaplasia (extension of goblet cells into the bronchioles where they normally do not occur). These phenomena also account for some of the increase of mucus secretion. The increased viscosity of the mucus can be partially explained by the CFTR defect and the effects of chronic inflammation. The mucin 2 gene (MUC2) contains a variable number of tandem 69bp repeats (119, 120). These repeats contain multiple glycosylation sites and were previously believed to be monomorphic (119), but preliminary work in our laboratory has shown that this region is polymorphic. This may have functional consequences since a mucin with a higher number of repeats will have a higher number of glycosylation sites and this may alter the viscoelastic properties of the 20  molecule. The MUC5B gene contains highly variable number of tandem repeats in intron 36 (121). 1.4 STATEMENT OF THE PROBLEM Variability in pulmonary disease progression in CF cannot be explained by the CFTR defect alone. The CFTR defect in association with gene defects in other genes involved in modulating the severity of the inflammatory response and response to infection may contribute to more severe and rapid phenotypic progression of pulmonary disease in CF. The main questions that this study addressed were: ƒ  Are genes other than the CFTR gene responsible for the variability in pulmonary disease severity and pulmonary disease progression seen in CF?  ƒ  Are genes other than the CFTR gene responsible for differential survival (i.e., clinical outcome; i.e., that is early lung transplantation or death) in CF?  ƒ  Are genes other than the CFTR gene responsible for susceptibility to recurrent pulmonary infections?  ƒ  Are genes other than the CFTR gene responsible for age of first and chronic infection with the respiratory pathogen P. aeruginosa?  ƒ  Are genes other that the CFTR gene responsible for the variability in frequency of pulmonary infections requiring intravenous therapy seen in CF patients?  ƒ  Are genes (i.e., MBL2) other than the CFTR gene responsible for the variability in pulmonary disease severity and pulmonary disease progression seen in CF patients chronically infected with BCC?  ƒ  Are genes (i.e., MBL2) other than the CFTR gene responsible for differential survival in CF patients chronically infected with BCC?  The main secondary questions that this study addressed were: ƒ  Is there a difference in pulmonary disease severity and progression following chronic infection with BCC based on genomovar group and random amplification of polymorphic DNA (RAPD)-type?  21  ƒ  Is there a difference in pulmonary disease severity and progression in CF patients infected with BCC only versus infected only with P. aeruginosa or infected with both pathogens?  ƒ  Is there a difference in survival following chronic infection with BCC based on genomovar group and RAPD-type?  ƒ  Is there a difference in survival among CF patients who are chronically infected with BCC versus P. aeruginosa or infected with both pathogens and neither pathogen?  ƒ  Are genes (i.e., MBL2) other than the CFTR gene responsible for increased susceptibility to BCC colonization in CF? 1.5 MAIN STUDY HYPOTHESIS  Polymorphisms in genes other than CFTR influence CF disease severity and progression and are associated with: ƒ  Worse pulmonary disease.  ƒ  Accelerated pulmonary disease progression  ƒ  Earlier infection with P.aeruginosa.  ƒ  Susceptibility to B. cepacia complex colonization  ƒ  Worse clinical outcome (early lung transplantation, death).  Specific Aims: Primary aims: 1. Determine whether alpha-1-antitrypsin deficiency is associated with more severe pulmonary disease and pulmonary disease progression. 2. Determine whether MBL2 deficiency is associated with more severe pulmonary disease and pulmonary disease progression, earlier chronic colonization with P.aeruginosa and increased susceptibility to B. cepacia colonization. 3. Perform haplotypes analysis for MBL2 and determine if MBL2 is in linkage disequilibrium with the adjacent pulmonary surfactant genes. 4. Determine whether the pulmonary surfactant genes are in linkage disequilibrium. 5. Determine whether pulmonary surfactant gene variants are associated more severe pulmonary disease and pulmonary disease progression and earlier chronic colonization with P. aeruginosa. 22  6. Determine whether there is a relationship between length of the MUC2 and MUC5AC tandem repeats and the severity of pulmonary disease in CF. 7. Determine whether the GST gene polymorphisms and the catalytic subunit of the glutathione cysteine ligase gene are associated with more severe pulmonary disease, pulmonary disease progression and earlier chronic colonization with P.aeruginosa. Secondary aims: 1. Determine whether BCC colonization is associated with a more rapid decline in pulmonary function and more severe clinical outcome taking into consideration genomovar grouping of BCC. 2. Determine whether MBL2 deficiency and BCC colonization with genomovar III versus II are associated with more severe pulmonary disease progression following colonization. 3.  Determine whether the pulmonary surfactant gene polymorphisms are associated with susceptibility to P.aeruginosa colonization and with more severe pulmonary disease progression following colonization.  4. Determine whether MBL2 deficiency and the pulmonary surfactant gene polymorphisms are associated with differences in the frequencies of pulmonary infections requiring IV therapy.  1.6 SIGNIFICANCE The proposed study will extend current knowledge of secondary genetic factors in pulmonary disease progression in CF. Polymorphisms in genes which modulate the severity of inflammation and increase susceptibility to infection could contribute to a more rapid progression of pulmonary disease if such polymorphism can be identified and this information could be used in patients’ care. Our larger relative sample will help correct for confounding genetic and environmental parameters, which have been a problem in previous studies.  23  1.7 BIBLIOGRAPHY  1.Zielenski, J., and L. C. Tsui. 1995. Cystic fibrosis: genotypic and phenotypic variations. Annu Rev Genet 29:777-807. 2.Boucher, R. C. 2004. New concepts of the pathogenesis of cystic fibrosis lung disease. Eur Respir J 23(1):146-58. 3.Tsui, L. C., M. Buchwald, D. Barker, J. C. Braman, R. Knowlton, J. W. Schumm, H. Eiberg, J. Mohr, D. Kennedy, N. Plavsic, and et al. 1985. Cystic fibrosis locus defined by a genetically linked polymorphic DNA marker. Science 230(4729):1054-7. 4.Knowlton, R. G., O. Cohen-Haguenauer, N. Van Cong, J. Frezal, V. A. Brown, D. Barker, J. C. Braman, J. W. Schumm, L. C. Tsui, M. Buchwald, and et al. 1985. A polymorphic DNA marker linked to cystic fibrosis is located on chromosome 7. Nature 318(6044):380-2. 5.Wainwright, B. J., P. J. Scambler, J. Schmidtke, E. A. Watson, H. Y. Law, M. Farrall, H. J. Cooke, H. Eiberg, and R. Williamson. 1985. Localization of cystic fibrosis locus to human chromosome 7cen-q22. Nature 318(6044):384-5. 6.White, R., S. Woodward, M. Leppert, P. O'Connell, M. Hoff, J. Herbst, J. M. Lalouel, M. Dean, and G. Vande Woude. 1985. A closely linked genetic marker for cystic fibrosis. Nature 318(6044):382-4. 7.Boucher, R. C. 2001. Pathogenesis of cystic fibrosis airways disease. Trans Am Clin Climatol Assoc 112:99-107. 8.Boucher, R. C. 2002. An overview of the pathogenesis of cystic fibrosis lung disease. Adv Drug Deliv Rev 54(11):1359-71. 9.Hoiby, N., E. W. Flensborg, B. Beck, B. Friis, S. V. Jacobsen, and L. Jacobsen. 1977. Pseudomonas aeruginosa infection in cystic fibrosis. Diagnostic and prognostic significance of  Pseudomonas  aeruginosa  precipitins  determined  by  means  of  crossed  immunoelectrophoresis. Scand J Respir Dis 58(2):65-79. 10. Zielenski, J. 2000. Genotype and phenotype in cystic fibrosis. Respiration 67(2):117-33. 11. Mangos, J. A., N. R. McSherry, and P. J. Benke. 1967. A sodium transport inhibitory factor in the saliva of patients with cystic fibrosis of the pancreas. Pediatr Res 1(6):436-42. 12. Spock, A., H. M. Heick, H. Cress, and W. S. Logan. 1967. Abnormal serum factor in patients with cystic fibrosis of the pancreas. Pediatr Res 1(3):173-7. 24  13. Spock, A., H. M. Heick, H. Cress, and W. Logan. 1967. In vitro study of ciliary motility to detect individuals with active cystic fibrosis and carriers of the disease. Bibl Paediatr 86:200-6. 14. Goodchild, M. C., J. H. Edwards, K. P. Glenn, C. Grindey, R. Harris, P. Mackintosh, and J. Wentzel. 1976. A search for linkage in cystic fibrosis. J Med Genet 13(6):417-9. 15. Eiberg, H., J. Mohr, K. Schmiegelow, L. S. Nielsen, and R. Williamson. 1985. Linkage relationships of paraoxonase (PON) with other markers: indication of PON-cystic fibrosis synteny. Clin Genet 28(4):265-71. 16. Schmiegelow, K., H. Eiberg, L. C. Tsui, M. Buchwald, P. D. Phelan, R. Williamson, W. Warwick, E. Niebuhr, J. Mohr, M. Schwartz, and et al. 1986. Linkage between the loci for cystic fibrosis and paraoxonase. Clin Genet 29(5):374-7. 17. Tsui, L. C., M. Zsiga, D. Kennedy, N. Plavsic, D. Markiewicz, and M. Buchwald. 1985. Cystic fibrosis: progress in mapping the disease locus using polymorphic DNA markers. I. Cytogenet Cell Genet 39(4):299-301. 18. Rommens, J. M., S. Zengerling-Lentes, B. Kerem, G. Melmer, M. Buchwald, and L. C. Tsui. 1989. Physical localization of two DNA markers closely linked to the cystic fibrosis locus by pulsed-field gel electrophoresis. Am J Hum Genet 45(6):932-41. 19. Rommens, J. M., M. C. Iannuzzi, B. Kerem, M. L. Drumm, G. Melmer, M. Dean, R. Rozmahel, J. L. Cole, D. Kennedy, N. Hidaka, and et al. 1989. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science 245(4922):1059-65. 20. Riordan, J. R., J. M. Rommens, B. Kerem, N. Alon, R. Rozmahel, Z. Grzelczak, J. Zielenski, S. Lok, N. Plavsic, J. L. Chou, and et al. 1989. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245(4922):1066-73. 21. Eiberg, H., K. Schmiegelow, C. Koch, J. Mohr, M. Schwartz, and E. Niebuhr. 1985. Cystic fibrosis; hint of linkage with F13B. Clin Genet 27(2):206. 22. Beaudet, A., A. Bowcock, M. Buchwald, L. Cavalli-Sforza, M. Farrall, M. C. King, K. Klinger, J. M. Lalouel, G. Lathrop, S. Naylor, and et al. 1986. Linkage of cystic fibrosis to two tightly linked DNA markers: joint report from a collaborative study. Am J Hum Genet 39(6):681-93. 23. Kerem, B. S., J. A. Buchanan, P. Durie, M. L. Corey, H. Levison, J. M. Rommens, M. Buchwald, and L. C. Tsui. 1989. DNA marker haplotype association with pancreatic sufficiency in cystic fibrosis. Am J Hum Genet 44(6):827-34. 25  24. Kerem, B., J. M. Rommens, J. A. Buchanan, D. Markiewicz, T. K. Cox, A. Chakravarti, M. Buchwald, and L. C. Tsui. 1989. Identification of the cystic fibrosis gene: genetic analysis. Science 245(4922):1073-80. 25. Farrall, M., B. J. Wainwright, G. L. Feldman, A. Beaudet, Z. Sretenovic, D. Halley, M. Simon, L. Dickerman, M. Devoto, and G. Romeo. 1988. Recombinations between IRP and cystic fibrosis. Am J Hum Genet 43(4):471-5. 26. Rich, D. P., M. P. Anderson, R. J. Gregory, S. H. Cheng, S. Paul, D. M. Jefferson, J. D. McCann, K. W. Klinger, A. E. Smith, and M. J. Welsh. 1990. Expression of cystic fibrosis transmembrane conductance regulator corrects defective chloride channel regulation in cystic fibrosis airway epithelial cells. Nature 347(6291):358-63. 27. Gregory, R. J., S. H. Cheng, D. P. Rich, J. Marshall, S. Paul, K. Hehir, L. Ostedgaard, K. W. Klinger, M. J. Welsh, and A. E. Smith. 1990. Expression and characterization of the cystic fibrosis transmembrane conductance regulator. Nature 347(6291):382-6. 28. Drumm, M. L., H. A. Pope, W. H. Cliff, J. M. Rommens, S. A. Marvin, L. C. Tsui, F. S. Collins, R. A. Frizzell, and J. M. Wilson. 1990. Correction of the cystic fibrosis defect in vitro by retrovirus-mediated gene transfer. Cell 62(6):1227-33. 29. Boucher, R. C. 2004. Relationship of airway epithelial ion transport to chronic bronchitis. Proc Am Thorac Soc 1(1):66-70. 30. Dharmsathaphorn, K., J. Cohn, and G. Beuerlein. 1989. Multiple calcium-mediated effector mechanisms regulate chloride secretory responses in T84-cells. Am J Physiol 256(6 Pt 1):C1224-30. 31. Zielenski, J., P. Patrizio, D. Markiewicz, R. H. Asch, and L. C. Tsui. 1997. Identification of two mutations (S50Y and 4173delC) in the CFTR gene from patients with congenital bilateral absence of vas deferens (CBAVD). Hum Mutat 9(2):183-4. 32. Smith, J. J., P. H. Karp, and M. J. Welsh. 1994. Defective fluid transport by cystic fibrosis airway epithelia. J Clin Invest 93(3):1307-11. 33. Smith, J. J., and M. J. Welsh. 1993. Fluid and electrolyte transport by cultured human airway epithelia. J Clin Invest 91(4):1590-7. 34. Smith, J. J., S. M. Travis, E. P. Greenberg, and M. J. Welsh. 1996. Cystic fibrosis airway epithelia fail to kill bacteria because of abnormal airway surface fluid. Cell 85(2):229-36. 35. Knowles, M. R., J. M. Robinson, R. E. Wood, C. A. Pue, W. M. Mentz, G. C. Wager, J. T. Gatzy, and R. C. Boucher. 1997. Ion composition of airway surface liquid of patients with 26  cystic fibrosis as compared with normal and disease-control subjects. J Clin Invest 100(10):2588-95. 36. Matsui, H., B. R. Grubb, R. Tarran, S. H. Randell, J. T. Gatzy, C. W. Davis, and R. C. Boucher. 1998. Evidence for periciliary liquid layer depletion, not abnormal ion composition, in the pathogenesis of cystic fibrosis airways disease. Cell 95(7):1005-15. 37. Matsui, H., S. H. Randell, S. W. Peretti, C. W. Davis, and R. C. Boucher. 1998. Coordinated clearance of periciliary liquid and mucus from airway surfaces. J Clin Invest 102(6):1125-31. 38. Burke, W., M. L. Aitken, S. H. Chen, and C. R. Scott. 1992. Variable severity of pulmonary disease in adults with identical cystic fibrosis mutations. Chest 102(2):506-9. 39. Wilmott, R. W., S. L. Tyson, and D. J. Matthew. 1985. Cystic fibrosis survival rates. The influences of allergy and Pseudomonas aeruginosa. Am J Dis Child 139(7):669-71. 40. 1993. Correlation between genotype and phenotype in patients with cystic fibrosis. The Cystic Fibrosis Genotype-Phenotype Consortium. N Engl J Med 329(18):1308-13. 41. Kerem, E., M. Corey, B. Kerem, P. Durie, L. C. Tsui, and H. Levison. 1989. Clinical and genetic comparisons of patients with cystic fibrosis, with or without meconium ileus. J Pediatr 114(5):767-73. 42. Ferrari, M., and L. Cremonesi. 1996. Genotype-phenotype correlation in cystic fibrosis patients. Ann Biol Clin (Paris) 54(6):235-41. 43. Welsh, M. J., and A. E. Smith. 1993. Molecular mechanisms of CFTR chloride channel dysfunction in cystic fibrosis. Cell 73(7):1251-4. 44. Brancolini, V., L. Cremonesi, E. Belloni, E. Pappalardo, R. Bordoni, M. Seia, S. Russo, R. Padoan, A. Giunta, and M. Ferrari. 1995. Search for mutations in pancreatic sufficient cystic fibrosis Italian patients: detection of 90% of molecular defects and identification of three novel mutations. Hum Genet 96(3):312-8. 45. Santis, G., L. Osborne, R. A. Knight, and M. E. Hodson. 1990. Independent genetic determinants of pancreatic and pulmonary status in cystic fibrosis. Lancet 336(8723):10814. 46. Bronsveld, I., F. Mekus, J. Bijman, M. Ballmann, J. Greipel, J. Hundrieser, D. J. Halley, U. Laabs, R. Busche, H. R. De Jonge, B. Tummler, and H. J. Veeze. 2000. Residual chloride secretion in intestinal tissue of deltaF508 homozygous twins and siblings with cystic fibrosis. The European CF Twin and Sibling Study Consortium. Gastroenterology 119(1):32-40. 27  47. Zielenski, J., M. Corey, R. Rozmahel, D. Markiewicz, I. Aznarez, T. Casals, S. Larriba, B. Mercier, G. R. Cutting, A. Krebsova, M. Macek, Jr., E. Langfelder-Schwind, B. C. Marshall, J. DeCelie-Germana, M. Claustres, A. Palacio, J. Bal, A. Nowakowska, C. Ferec, X. Estivill, P. Durie, and L. C. Tsui. 1999. Detection of a cystic fibrosis modifier locus for meconium ileus on human chromosome 19q13. Nat Genet 22(2):128-9. 48. Cantin, A. M., S. Lafrenaye, and R. O. Begin. 1991. Antineutrophil elastase activity in cystic fibrosis serum. Pediatric Pulmonology 11(3):249-53. 49. Birrer, P., N. G. McElvaney, A. Rudeberg, C. W. Sommer, S. Liechti-Gallati, R. Kraemer, R. Hubbard, and R. G. Crystal. 1994. Protease-antiprotease imbalance in the lungs of children with cystic fibrosis. American Journal of Respiratory & Critical Care Medicine 150(1):207-13. 50. O'Connor, C. M., K. Gaffney, J. Keane, A. Southey, N. Byrne, S. O'Mahoney, and M. X. Fitzgerald. 1993. alpha 1-Proteinase inhibitor, elastase activity, and lung disease severity in cystic fibrosis. Am Rev Respir Dis 148(6 Pt 1):1665-70. 51. Goldstein, W., and G. Doring. 1986. Lysosomal enzymes from polymorphonuclear leukocytes and proteinase inhibitors in patients with cystic fibrosis. American Review of Respiratory Disease 134(1):49-56. 52. Suter, S. 1986. Imbalance between polymorphonuclear leukocyte proteases and antiproteases in chronic pyogenic infections and its relation to the proteolysis of complement component C3. Complement 3(1):1-24. 53. Suter, S., U. B. Schaad, H. Tegner, K. Ohlsson, D. Desgrandchamps, and F. A. Waldvogel. 1986. Levels of free granulocyte elastase in bronchial secretions from patients with cystic fibrosis: effect of antimicrobial treatment against Pseudomonas aeruginosa. Journal of Infectious Diseases 153(5):902-9. 54. Brown, M. A., W. J. Morgan, P. R. Finley, and P. Scuderi. 1991. Circulating levels of tumor necrosis factor and interleukin-1 in cystic fibrosis. Pediatric Pulmonology 10(2):8691. 55. Morgan, K., G. Scobie, P. Marsters, and N. A. Kalsheker. 1997. Mutation in an alpha1antitrypsin enhancer results in an interleukin-6 deficient acute-phase response due to loss of cooperativity between transcription factors. Biochim Biophys Acta 1362(1):67-76. 56. Hilman, B. C. 1997. Genetic and immunologic aspects of cystic fibrosis. Ann Allergy Asthma Immunol 79(5):379-90; quiz 391-4. 28  57. Doring, G., H. Krogh-Johansen, S. Weidinger, and N. Hoiby. 1994. Allotypes of alpha 1antitrypsin in patients with cystic fibrosis, homozygous and heterozygous for deltaF508. Pediatric Pulmonology 18(1):3-7. 58. Plotkowski, M. C., G. Beck, J. M. Tournier, M. Bernardo-Filho, E. A. Marques, and E. Puchelle. 1989. Adherence of Pseudomonas aeruginosa to respiratory epithelium and the effect of leucocyte elastase. J Med Microbiol 30(4):285-93. 59. Santis, G., L. Osborne, R. Knight, M. E. Hodson, and M. Ramsay. 1990. Genetic influences on pulmonary severity in cystic fibrosis. Lancet 335(8684):294. 60. Santis, G., L. Osborne, R. Knight, M. Ramsay, R. Williamson, and M. Hodson. 1990. Cystic fibrosis haplotype association and the delta F508 mutation in adult British CF patients. Hum Genet 85(4):424-5. 61. Plotkowski, M. C., A. M. Saliba, S. H. Pereira, M. P. Cervante, and O. Bajolet-Laudinat. 1994. Pseudomonas aeruginosa selective adherence to and entry into human endothelial cells. Infect Immun 62(12):5456-63. 62. McElvaney, N. G., R. C. Hubbard, P. Birrer, M. S. Chernick, D. B. Caplan, M. M. Frank, and R. G. Crystal. 1991. Aerosol alpha 1-antitrypsin treatment for cystic fibrosis. Lancet 337(8738):392-4. 63. Mahadeva, R., R. C. Westerbeek, D. J. Perry, J. U. Lovegrove, D. B. Whitehouse, N. R. Carroll, R. I. Ross-Russell, A. K. Webb, D. Bilton, and D. A. Lomas. 1998. Alpha1antitrypsin deficiency alleles and the Taq-I G-->A allele in cystic fibrosis lung disease. Eur Respir J 11(4):873-9. 64. Mahadeva, R., S. Stewart, D. Bilton, and D. A. Lomas. 1998. Alpha-1 antitrypsin deficiency alleles and severe cystic fibrosis lung disease. Thorax 53(12):1022-4. 65. Meyer, P., A. Braun, and A. A. Roscher. 2002. Analysis of the two common alpha-1antitrypsin deficiency alleles PiMS and PiMZ as modifiers of Pseudomonas aeruginosa susceptibility in cystic fibrosis. Clin Genet 62(4):325-7. 66. Kalsheker, N. A., G. L. Watkins, S. Hill, K. Morgan, R. A. Stockley, and R. B. Fick. 1990. Independent mutations in the flanking sequence of the alpha-1- antitrypsin gene are associated with chronic obstructive airways disease. Dis Markers 8(3):151-7. 67. Sandford, A. J., T. Chagani, J. J. Spinelli, and P. D. Pare. 1999. alpha1-antitrypsin genotypes and the acute-phase response to open heart surgery. Am J Respir Crit Care Med 159(5 Pt 1):1624-8. 29  68. Henry, M. T., S. Cave, J. Rendall, C. M. O'Connor, K. Morgan, M. X. FitzGerald, and N. Kalsheker. 2001. An alpha(1)-antitrypsin enhancer polymorphism is a genetic modifier of pulmonary outcome in cystic fibrosis. Eur J Hum Genet 9(4):273-8. 69. Wacholder, S., S. Chanock, M. Garcia-Closas, L. El Ghormli, and N. Rothman. 2004. Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst 96(6):434-42. 70. Koch, A., M. Melbye, P. Sorensen, P. Homoe, H. O. Madsen, K. Molbak, C. H. Hansen, L. H. Andersen, G. W. Hahn, and P. Garred. 2001. Acute respiratory tract infections and mannose-binding lectin insufficiency during early childhood. Jama 285(10):1316-21. 71. Summerfield, J. A., M. Sumiya, M. Levin, and M. W. Turner. 1997. Association of mutations in mannose binding protein gene with childhood infection in consecutive hospital series. Bmj 314(7089):1229-32. 72. Garred, P., T. Pressler, H. O. Madsen, B. Frederiksen, A. Svejgaard, N. Hoiby, M. Schwartz, and C. Koch. 1999. Association of mannose-binding lectin gene heterogeneity with severity of lung disease and survival in cystic fibrosis. J Clin Invest 104(4):431-7. 73. Madsen, H. O., P. Garred, S. Thiel, J. A. Kurtzhals, L. U. Lamm, L. P. Ryder, and A. Svejgaard. 1995. Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein. J Immunol 155(6):3013-20. 74. Davies, J., O. Neth, E. Alton, N. Klein, and M. Turner. 2000. Differential binding of mannose-binding lectin to respiratory pathogens in cystic fibrosis. Lancet 355(9218):1885-6. 75. Gabolde, M., M. Guilloud-Bataille, J. Feingold, and C. Besmond. 1999. Association of variant alleles of mannose binding lectin with severity of pulmonary disease in cystic fibrosis: cohort study. Bmj 319(7218):1166-7. 76. Tenner, A. J. 1999. Membrane receptors for soluble defense collagens. Curr Opin Immunol 11(1):34-41. 77. Eggleton, P., and K. B. Reid. 1999. Lung surfactant proteins involved in innate immunity. Curr Opin Immunol 11(1):28-33. 78. Crouch, E. C. 1998. Structure, biologic properties, and expression of surfactant protein D (SP-D). Biochim Biophys Acta 1408(2-3):278-89. 79. Vandivier, R. W., V. A. Fadok, C. A. Ogden, P. R. Hoffmann, J. D. Brain, F. J. Accurso, J. H. Fisher, K. E. Greene, and P. M. Henson. 2002. Impaired clearance of apoptotic cells from cystic fibrosis airways. Chest 121(3 Suppl):89S. 30  80. Wert, S. E., M. Yoshida, A. M. LeVine, M. Ikegami, T. Jones, G. F. Ross, J. H. Fisher, T. R. Korfhagen, and J. A. Whitsett. 2000. Increased metalloprotenase activity, oxidant production, and emphysema in surfactant protein D gene-inactivated mice. Proceeds of the National Academy of Science 97:5972-. 81. LeVine, A. M., K. E. Kurak, M. D. Bruno, J. M. Stark, J. A. Whitsett, and T. R. Korfhagen. 1998. Surfactant protein-A-deficient mice are succeptible to Pseudomonas aeruginosa infection. Am J Respir Cell Mol Biol 19:700-708. 82. LeVine, A. M., M. D. Bruno, K. M. Huelsman, G. F. Ross, J. A. Whitsett, and T. R. Korfhagen. 1997. Surfactant protein A-deficient mice are susceptible to group B streptococcal infection. J Immunol 158(9):4336-40. 83. Kala, P., T. Ten Have, H. Nielsen, M. Dunn, and J. Floros. 1998. Association of pulmonary surfactant protein A (SP-A) gene and respiratory distress syndrome: interaction with SP-B. Pediatr Res 43(2):169-77. 84. Guo, X., H. M. Lin, Z. Lin, M. Montano, R. Sansores, G. Wang, S. DiAngelo, A. Pardo, M. Selman, and J. Floros. 2001. Surfactant protein gene A, B, and D marker alleles in chronic obstructive pulmonary disease of a Mexican population. Eur Respir J 18(3):482-90. 85. Lin, Z., C. Pearson, V. Chinchilli, S. M. Pietschmann, J. Luo, U. Pison, and J. Floros. 2000. Polymorphisms of human SP-A, SP-B, and SP-D genes: association of SP-B Thr131Ile with ARDS. Clin Genet 58(3):181-91. 86. Davies, J. C., M. W. Turner, N. Klein, and t. L. M. S. Group. 2004. Impaired pulmonary status in cystic fibrosis adults with two mutated MBL-2 alleles. Eur Respir J 24:798-804. 87. MacNee, W., and I. Rahman. 1995. Oxidants/antioxidants in idiopathic pulmonary fibrosis. Thorax 50 Suppl 1:S53-8. 88. Cantin, A. M., R. C. Hubbard, and R. G. Crystal. 1989. Glutathione deficiency in the epithelial lining fluid of the lower respiratory tract in idiopathic pulmonary fibrosis. Am Rev Respir Dis 139(2):370-2. 89. Rahman, I., E. Skwarska, M. Henry, M. Davis, C. M. O'Connor, M. X. FitzGerald, A. Greening, and W. MacNee. 1999. Systemic and pulmonary oxidative stress in idiopathic pulmonary fibrosis. Free Radic Biol Med 27(1-2):60-8. 90. Pacht, E. R., A. P. Timerman, M. G. Lykens, and A. J. Merola. 1991. Deficiency of alveolar fluid glutathione in patients with sepsis and the adult respiratory distress syndrome. Chest 100(5):1397-403. 31  91. Moss, M., D. M. Guidot, M. Wong-Lambertina, T. Ten Hoor, R. L. Perez, and L. A. Brown. 2000. The effects of chronic alcohol abuse on pulmonary glutathione homeostasis. Am J Respir Crit Care Med 161(2 Pt 1):414-9. 92. Rahman, I., S. K. Biswas, and A. Kode. 2006. Oxidant and antioxidant balance in the airways and airway diseases. Eur J Pharmacol 533(1-3):222-39. 93. Boda, D., I. Nemeth, and S. Pinter. 1998. Surface tension, glutathione content and redox ratio of the tracheal aspirate fluid of premature infants with IRDS. Biol Neonate 74(4):2818. 94. Droge, W., and E. Holm. 1997. Role of cysteine and glutathione in HIV infection and other diseases associated with muscle wasting and immunological dysfunction. Faseb J 11(13):1077-89. 95. Mihm, S., J. Ennen, U. Pessara, R. Kurth, and W. Droge. 1991. Inhibition of HIV-1 replication and NF-kappa B activity by cysteine and cysteine derivatives. Aids 5(5):497-503. 96. Pacht, E. R., P. Diaz, T. Clanton, J. Hart, and J. E. Gadek. 1997. Alveolar fluid glutathione decreases in asymptomatic HIV-seropositive subjects over time. Chest 112(3):785-8. 97. Staal, F. J., S. W. Ela, M. Roederer, M. T. Anderson, and L. A. Herzenberg. 1992. Glutathione  deficiency  and  human  immunodeficiency  virus  infection.  Lancet  339(8798):909-12. 98. Roum, J. H., Z. Borok, N. G. McElvaney, G. J. Grimes, A. D. Bokser, R. Buhl, and R. G. Crystal. 1999. Glutathione aerosol suppresses lung epithelial surface inflammatory cellderived oxidants in cystic fibrosis. J Appl Physiol 87(1):438-43. 99. Velsor, L. W., A. van Heeckeren, and B. J. Day. 2001. Antioxidant imbalance in the lungs of cystic fibrosis transmembrane conductance regulator protein mutant mice. Am J Physiol Lung Cell Mol Physiol 281(1):L31-8. 100. Boxer, L. A., J. M. Oliver, S. P. Spielberg, J. M. Allen, and J. D. Schulman. 1979. Protection of granulocytes by vitamin E in glutathione synthetase deficiency. N Engl J Med 301(17):901-5. 101. Hayes, J. D., and L. I. McLellan. 1999. Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress. Free Radic Res 31(4):273-300. 102. Feher, J., G. Lengyel, and A. Blazovics. 1998. Oxidative stress in the liver and biliary tract diseases. Scand J Gastroenterol Suppl 228:38-46. 32  103. Roum, J. H., R. Buhl, N. G. McElvaney, Z. Borok, and R. G. Crystal. 1993. Systemic deficiency of glutathione in cystic fibrosis. J Appl Physiol 75(6):2419-24. 104. Gao, L., J. R. Broughman, T. Iwamoto, J. M. Tomich, C. J. Venglarik, and H. J. Forman. 2001. Synthetic chloride channel restores glutathione secretion in cystic fibrosis airway epithelia. Am J Physiol Lung Cell Mol Physiol 281(1):L24-30. 105. Gao, L., K. J. Kim, J. R. Yankaskas, and H. J. Forman. 1999. Abnormal glutathione transport in cystic fibrosis airway epithelia. Am J Physiol 277(1 Pt 1):L113-8. 106. Hull, J., and A. H. Thomson. 1998. Contribution of genetic factors other than CFTR to disease severity in cystic fibrosis. Thorax 53(12):1018-21. 107. Baranov, V. S., T. Ivaschenko, B. Bakay, M. Aseev, R. Belotserkovskaya, H. Baranova, P. Malet, J. Perriot, P. Mouraire, V. N. Baskakov, G. A. Savitskyi, S. Gorbushin, S. I. Deyneka, E. Michnin, A. Barchuck, V. Vakharlovsky, G. Pavlov, V. I. Shilko, T. Guembitzkaya, and L. Kovaleva. 1996. Proportion of the GSTM1 0/0 genotype in some Slavic populations and its correlation with cystic fibrosis and some multifactorial diseases. Hum Genet 97(4):516-20. 108. Henrion-Caude, A., C. Flamant, M. Roussey, C. Housset, A. Flahault, A. A. Fryer, K. Chadelat, R. C. Strange, and A. Clement. 2002. Liver disease in pediatric patients with cystic fibrosis is associated with glutathione S-transferase P1 polymorphism. Hepatology 36(4 Pt 1):913-7. 109. Arkwright, P. D., V. Pravica, P. J. Geraghty, M. Super, A. K. Webb, M. Schwarz, and I. V. Hutchinson. 2003. End-organ dysfunction in cystic fibrosis: association with angiotensin I converting enzyme and cytokine gene polymorphisms. Am J Respir Crit Care Med 167(3):384-9. 110. Arkwright, P. D., S. Laurie, M. Super, V. Pravica, M. J. Schwarz, A. K. Webb, and I. V. Hutchinson. 2000. TGF-beta(1) genotype and accelerated decline in lung function of patients with cystic fibrosis. Thorax 55(6):459-62. 111. Aron, Y., B. S. Polla, T. Bienvenu, J. Dall'ava, D. Dusser, and D. Hubert. 1999. HLA class II polymorphism in cystic fibrosis. A possible modifier of pulmonary phenotype. Am J Respir Crit Care Med 159(5 Pt 1):1464-8. 112. Wang, D., G. M. Levasseur-Acker, R. Jankowski, G. Kanny, D. A. Moneret-Vautrin, D. Charron, A. Lockhart, and E. Swierczewski. 1997. HLA class II antigens and T lymphocytes in human nasal epithelial cells. Modulation of the HLA class II gene transcripts by gamma interferon. Clin Exp Allergy 27(3):306-14. 33  113. Seregni, E., C. Botti, S. Massaron, C. Lombardo, A. Capobianco, A. Bogni, and E. Bombardieri. 1997. Structure, function and gene expression of epithelial mucins. Tumori 83(3):625-32. 114. Shankar, V., P. Pichan, R. L. Eddy, Jr., V. Tonk, N. Nowak, S. N. Sait, T. B. Shows, R. E. Schultz, G. Gotway, R. C. Elkins, M. S. Gilmore, and G. P. Sachdev. 1997. Chromosomal localization of a human mucin gene (MUC8) and cloning of the cDNA corresponding to the carboxy terminus. Am J Respir Cell Mol Biol 16(3):232-41. 115. Rose, M. C., and J. A. Voynow. 2006. Respiratory tract mucin genes and mucin glycoproteins in health and disease. Physiol Rev 86(1):245-78. 116. Nollet, S., N. Moniaux, J. Maury, D. Petitprez, P. Degand, A. Laine, N. Porchet, and J. P. Aubert. 1998. Human mucin gene MUC4: organization of its 5'-region and polymorphism of its central tandem repeat array. Biochem J 332 ( Pt 3):739-48. 117. Pigny, P., V. Guyonnet-Duperat, A. S. Hill, W. S. Pratt, S. Galiegue-Zouitina, M. C. d'Hooge, A. Laine, I. Van-Seuningen, P. Degand, J. R. Gum, Y. S. Kim, D. M. Swallow, J. P. Aubert, and N. Porchet. 1996. Human mucin genes assigned to 11p15.5: identification and organization of a cluster of genes. Genomics 38(3):340-52. 118. Hovenberg, H. W., J. R. Davies, A. Herrmann, C. J. Linden, and I. Carlstedt. 1996. MUC5AC, but not MUC2, is a prominent mucin in respiratory secretions. Glycoconj J 13(5):839-47. 119. Toribara, N. W., J. R. Gum, Jr., P. J. Culhane, R. E. Lagace, J. W. Hicks, G. M. Petersen, and Y. S. Kim. 1991. MUC-2 human small intestinal mucin gene structure. Repeated arrays and polymorphism. J Clin Invest 88(3):1005-13. 120. Debailleul, V., A. Laine, G. Huet, P. Mathon, M. C. d'Hooghe, J. P. Aubert, and N. Porchet. 1998. Human mucin genes MUC2, MUC3, MUC4, MUC5AC, MUC5B, and MUC6 express stable and extremely large mRNAs and exhibit a variable length polymorphism. An improved method to analyze large mRNAs. J Biol Chem 273(2):881-90. 121. Desseyn, J. L., K. Rousseau, and A. Laine. 1999. Fifty-nine bp repeat polymorphism in the uncommon intron 36 of the human mucin gene MUC5B. Electrophoresis 20(3):493-6.  34  CHAPTER 2: STUDY METHODOLOGY  35  2.0 METHODS AND PROCEDURES The chapter describes clinic and patient study population recruitment, laboratory techniques performed, clinical data (definition of variables, description of outcome and independent variables) collected, hypotheses tested and statistical data analysis techniques. 2.1 SUBJECT AND ELIGIBILITY CRITERIA Patients attending the following Canadian CF clinics were recruited: Adult CF clinic at St. Paul’s Hospital (Vancouver, B.C.). Children’s CF clinic at B.C. Children’s Hospital (Vancouver, B.C.), Adult CF clinic at Victoria General Hospital (Victoria, B.C.), Children’s CF clinic at Victoria Children’s Hospital (Victoria, B.C.). Adult CF clinic at Hamilton Health Sciences (Hamilton, Ontario). Children’s CF clinic at Hamilton Health Sciences (Hamilton, Ontario). Adult CF clinic at Hôtel-Dieu du Chum (Montreal, Quebec). Children’s CF Clinic at Hospital for Sick Children (Toronto, Ontario). Adult CF clinic at St. Michael's Hospital (Toronto, Ontario). Adult CF clinic in Seattle (Seattle, Washington, U.S.A.). An attempt was made to recruit all patients attending the participating clinics. The potential study sample based on clinic patient numbers was 1265. Table 2.1 provides a description of clinic participation for each gene studied. Patients attending the Toronto clinics only participated in the investigation of the alpha-1-antitrypsin (α1-AT) gene as a potential modifier gene and provided 2 years of clinical follow-up data. Patients attending the Seattle clinic only participated in the investigation of the GST and GCLC genes as potential modifier genes and only provided cross-sectional data. The remaining genes were investigated on patients attending the Vancouver, Hamilton, and Montreal clinics. The Victoria clinics did not participate in the investigation of the alpha-1antitrypsin gene and were recruited subsequently to increase sample size and used in our investigation of innate immunity and antioxidant genes. Table 2.1 includes the number of CF patients genotyped for each polymorphism by the center attended.  36  Patients with a diagnosis of CF on the basis of clinical signs, elevated sweat chloride values and/or positive genotyping for mutant CFTR allele(s) were recruited for the study. CF patients who had received a lung transplant were also recruited and pulmonary function data from prior to transplantation were collected for these CF patients. In a separate sub-study we also recruited 31 consecutive patients from the St. Paul’s Hospital adult CF clinic (mean age (±SEM) 27.5(1.1) years) who developed an acute pulmonary exacerbation, to measure serum α1-AT levels during the acute phase and 2-3 months later during a stable phase. Details of the experimental procedures, the risks and benefits involved were explained to subjects before obtaining written consent, which was approved by the Ethics Committees of the institutions participating in this study (Appendix A). 2.2 LABORATORY TECHNIQUES 2.2.1 Extraction of genomic DNA from human whole blood Ten milliliters of whole blood was collected from consenting CF patients. Samples for DNA were collected in one 10ml or two 7ml EDTA lavender topped tubes and stored in a 4°C fridge for between 0-4 days and then either immediately extracted or stored in a -20°C freezer prior to extraction. Thawed whole blood for each sample was transferred into a 50ml sterile centrifuge Falcon tube and the tube was rinsed with 10ml of 0.1% IGEPAL CA-630 (octylphenyl-polyethylene glycol) in sterile water and added to the Falcon tube. Additional 0.1% IGEPAL was added to the Falcon tube to a final volume of 40ml. The tube with the sample mixture was mixed and allowed to sit for 20 minutes for complete lysis of the red blood cells. The tube was then spun for 20 minutes (in a Beckman Model TJ-6 Centrifuge at 2300RPM (1100g)) to pellet the white cells. The supernatant was poured out and the cell pellet washed with 40ml of 0.1% IGEPAL and mixed on vortex for 5 -10 seconds. The sample tube was spun again (at 2300RPM for 20 minutes), the supernatant (which contained the remaining lysed red blood cells) was poured off. The cell pellet was suspended in 10ml of digestion buffer containing 37  50μl of 20mg/ml proteinase K. The sample was incubated overnight at 50°C with agitation at 250 RPM. On the second day an equal volume (10ml) of phenol/chloroform/isoamyl alcohol (25:24:1) and 1ml of 2M NaCl were added to each sample and mixed gently using a vortex. The tubes were spun in the centrifuge (at 2300RPM for 20 minutes), then the upper aqueous layer was transferred to a new 50ml Falcon tube. To precipitate the DNA, 2 volumes of 95% ethanol were added and the solution gently mixed. A 1.5ml sterile microfuge tube with 1ml 70% ethanol was prepared and using a sterile pipette tip the DNA was spooled out of the falcon tube and added to the microfuge tube. The DNA-ethanol mixture was centrifuged for 5 minutes at 14,000 RPM and the ethanol poured off and the pellet air-dried under Kimwipes. The DNA pellets were finally dissolved in 500-800μl of TE buffer (pH=8.0) at room temperature. 2.2.2 DNA Quantification The optical density (OD) of each sample was measured at a wavelength of 260 nm and 280 nm using an ultraviolet spectrophotometer. We measured a 100-fold dilution of the original DNA solution. One OD260 unit corresponds to a double strand DNA concentration of 50 μg/ml. Proteins or RNA also absorb UV light at 260nm and 280nm. The purity of DNA samples were estimated by the ratio A260/A280. A ratio of 1.7 to 1.9 was indicative of a pure DNA sample. 2.3 GENOTYPING METHODS 2.3.1 Amplification of DNA using the Polymerase Chain Reaction The purpose of the polymerase chain reaction (PCR) is to artificially replicate a short sequence of DNA, in order to make millions of copies of this DNA sequence. Two oligonucleotides are used as primers for a series of synthetic reactions that are catalyzed by a thermostable DNA polymerase (e.g., Taq DNA polymerase). The oligonucleotides are complementary to sequences that lie on opposite strands of the template DNA and flank the segment we want to amplify. Genomic DNA is first heated (approximately 94°C) to denature the double stranded DNA molecules in a solution containing an excess of the two oligonucleotide primers and the four deoxynucleoside triphosphates (dNTPs). The reaction mixture is then cooled (45-65°C) to allow the primers to anneal to their target sequences and 38  then heated to an intermediate temperature (i.e., 70-75°C) which allows extension of the annealed primers from their 3’ termini by Taq DNA polymerase. The cycle of denaturation, annealing and primer extension is repeated 35-40 times using a thermal cycler.  The  products of each round of amplification serve as the template for the next, effectively doubling the amount of synthesized DNA with each cycle. Typically, this exponential amplification is terminated when the amount of active Taq DNA polymerase is exhausted at which time the selected DNA sequence has been amplified 106 times. 2.3.2 Genotyping using Restriction Fragment Length Polymorphisms Restriction enzymes (restriction endonucleases) bind specifically to, and cleave double stranded DNA at specific sites that are within or adjacent to particular sequences of nucleotides. Restriction enzymes are purified from bacteria; many have been isolated and they recognize specific sequences of double stranded DNA which are 4-8 nucleotides long, although there are a few enzymes which recognize longer sequences. These enzymes cut DNA at specific sequences and produce DNA fragments of specific sizes which are called restriction fragments. Single nucleotide polymorphisms can create or destroy restriction enzyme recognition sites. Therefore the presence or the absence of a recognition site can be used to detect a polymorphism. The restriction enzyme is added to the PCR mixture and digested in a specific restriction enzyme buffer and then analyzed on a gel. When more than one PCR is performed in one reaction, this is defined as a multiplex PCR. 2.3.3 Amplification of DNA using Site Directed Mutagenesis Polymerase Chain Reaction followed by Restriction Digestion In this technique, a primer is used which contains a one base mismatch and gives rise to a new restriction enzyme site following the PCR. Then an appropriate restriction enzyme is used to cut the DNA fragment at this new site. This method is used when the polymorphism does not lie within a naturally occurring restriction enzyme recognition site; therefore we create a site which can be recognized by a known restriction enzyme.  39  2.3.4 Amplification of DNA using allele-specific oligonucleotide PCR or Sequence Specific Priming Polymerase Chain Reaction In allele-specific oligonucleotide (ASP) PCR or sequence-specific priming (SSP) PCR, the primer contains the specific polymorphic site. We used two allele specific primers that were complementary to the respective two alleles. If the allele contains the specific polymorphism then we obtain a PCR product. This is visualized on a gel where we obtain a band when the allele is present. If the allele is not present no band is seen. In order to ensure that the PCR has worked an internal control is used. 2.4 GENE POLYMORPHISMS INVESTIGATED AND GENOTYPING METHODS 2.4.1 Gene Polymorphisms investigated and genotyping techniques Patients were genotyped for polymorphisms in the following genes: α1-AT gene: We studied the S and Z mutations in exons 3 and 5, respectively and the TaqI polymorphism in the 3’ untranslated region. Innate immunity genes: MBL gene: promoter polymorphism (Y or X [G→C] at -221) and the 3 single base structural polymorphisms at codons 52 (Arg→Cys, D allele), 54 (Gly→Asp, B allele), and 57 (Gly→Glu, C allele). SPA-1: we studied a polymorphism in exon 4 (C655T; Arg219Trp). SPA-2: we studied two polymorphisms. One polymorphism in exon 2 (A26C;Thr9Asn) and one in exon 4 (A667C; Lys223 Gln). SPD: we studied a polymorphism in exon 1 (C32T; Thr11Met). Other genes: MUC2 gene: we studied the imperfectly conserved repeats (mRNA sequence accession number: NM_002457 (1)). MUC5b gene: we studied the variable number tandem repeat region in intron 36 (genomic DNA accession number: Y09788 (2, 3)). GSTP1 gene: we studied the polymorphism in exon 5 (A313G; Ile105Val). GSTM1 and GSTT1 genes: we studied deletions of these two genes. GCLC gene: we studied the trinucleotide repeat (GAC) upstream from the translation start codon(4). 40  Table 2.2 shows the gene polymorphisms studied and genotyping methods used. The PHASE program (5, 6)was used to identify haplotypes for the MBL gene and the surfactant genes. 2.4.2 Alpha-1-antitrypsin gene The Z and S polymorphisms were examined by multiplex site directed mutagenesis PCR / restriction fragment length polymorphisms (RFLP) assays using primers which amplified regions of exon 5 and 3, respectively(7). Two mixtures were prepared; the first mixture consisted of 0.5μl of 10μM EX55 upstream primer (5’-TAA GGC TGT GCT GAC CAT CGT C-3’) and 0.5μl of 10μM BYZ downstream primer (5’-CAA AGG GTT TGT TGA ACT TGA CC-3’) for the Z mutation, and 1μl of 10μM EX35 upstream primer with a mismatch to produce a TaqI site (5’-GAG GGG AAA CTA CAG CAC CTC G-3’) and 1μl of 10μM EX33 downstream primer (5’-ACC CTC AGG TTG GGG AAT CAC C-3’) for the S mutation, 1μl (0.1μg) DNA and water (6 μl ddH20) to a final volume of 10μl was heated (7 minute denaturation step at 94°C followed by 10 minutes at 80°C). Mixture 2 of the PCR reaction contained 2μl 10xPCR buffer (by Invitrogen containing 500mM KCl, 200mM Tris-Cl, pH 8.4), 2μl 2mM deoxynucleoside triphosphates, 0.6μl 50mM MgCl2, 0.1μl Taq DNA polymerase (0.5U), with water (5.3 μl ddH20) added to a final volume of 10μl. Mixture 2 was added to mixture 1 followed by thermal cycling (35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 58°C, 30 second extension at 72°C and a final 5 minute extension at 72°C). Digestion of PCR products was performed with restriction enzyme TaqI (0.5μl (10U)), 3μl 10xTaqI buffer, 0.3μl 100X BSA and 6.2μl ddH20) at 65°C for five hours. At the Z mutation site, the restriction enzyme cuts the wild-type allele into 2 fragments (123bp and 21bp) and leaves the Z allele uncut (144bp). At the S mutation site, the restriction enzyme cuts the wild-type allele into 2 fragments (78bp and 20bp) and leaves the S allele uncut (98bp). The PCR product was electrophoresed on a 2.5% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.1). The 3 prime TaqI polymorphism (G1237→A) was examined by PCR/RFLP using primers which amplified a 373bp region. Each PCR reaction containing 1μl of 10μM P2 forward 41  primer (5’-CTC TCA GGT CTG GTG TCA TCC C-3’) and 1μl of 10μM P4 reverse primer (5’-GAC ACA GCA GCC AGG AAG TCC-3’), 1μl (0.1μg) DNA and 7 μl ddH20 to a final volume of 10μl was heated (7 minute denaturation step at 94°C followed by 10 minutes at 80°C) (Mixture 1). Mixture 2 of the PCR reaction contained 2μl 10xPCR buffer (by Invitrogen containing 500mM KCl, 200mM Tris-Cl, pH 8.4), 2μl 2mM deoxynucleoside triphosphates, 0.6μl 50mM MgCl2, 0.1μl Taq DNA polymerase (0.5U), with 5.3 μl ddH20 added to a final volume of 10μl. Mixture 2 was added to mixture 1 followed by thermal cycling (40 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C). Digestion of PCR products was performed with the restriction enzyme TaqI (0.5μl (10U) TaqI, 3μl 10x TaqI buffer, 0.3μl 100xBSA and 6.2μl ddH20) at 65°C for five hours. The restriction enzyme cuts the wild-type G allele into 2 fragments (191bp and 182bp) and leaves the A allele uncut. The PCR product was electrophoresed on a 2.0% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.2). 2.4.3 Mannose-Binding Lectin gene MBL2-B, C and D allele genotyping was performed by PCR-RFLP as described by Madsen et al. (8) with some modification. The B and C alleles were detected respectively by BanI and MboII restriction enzyme digestion of the 329bp product that had been amplified by the MBP1left upstream primer (5'-GTA GGA CAG AGG GCA TGC TC-3') and the MBP1right downstream primer (5'-CAG GCA GTT TCC TCT GGA AGG-3'), followed by a 2% agarose gel electrophoresis in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light. BanI cleaves the A allele into two fragments (245bp and 84bp) and leaves the B allele undigested. MboII cleaves the C allele into two fragments (266bp and 63bp) and leaves the A allele undigested (Figure 2.3 A and B). Specifically for MBL2-B, the PCR reaction contained 2μl 10xPCR buffer (by Invitrogen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7), 0.6μl 50mM MgCl2, 2μl 2mM deoxynucleoside triphosphates (dNTPs), 1μl of 10μM primer MBP1left and 1μl of 10μM primer MBP1right, 1μl (0.25μg) DNA, 0.1μl Taq DNA polymerase (by Invitrogen), with 12.3 μl ddH20 added to a final volume of 20μl. Following a 2 minute denaturation step at 42  94°C, 35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products was performed with 0.25μl (10U/μl) of the restriction enzyme 0.25μl BanI (10U/μl), 2.5μl 10x NE buffer 4, and 2.25μl ddH20 to a final volume of 5μl at 37°C overnight (Figure 2.3 A). Specifically for MBL-C, the PCR reaction contained 2μl 10xPCR buffer (by Invitrogen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7), 0.6μl 50mM MgCl2, 2μl 2mM dNTPs, 1μl of 10μM primer MBP1left and 1μl of 10μM primer MBP1right, 1μl (0.25μg) DNA, 0.1μl Taq DNA polymerase (by Invitrogen), with 12.3 μl ddH20 added to a final volume of 20μl. Following a 2 minute denaturation step at 94°C, 35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products was performed with 0.5μl (5U/μl) of the restriction enzyme 0.5μl MboII (5U/μl), 2.5μl 10x NE buffer 2, and 2μl ddH20 to a final volume of 5μl at 37°C overnight (Figure 2.3B). A MluI restriction enzyme site was introduced into the amplification product of the D allele by site-directed mutagenesis (SDM)-PCR. The PCR reaction contained 2μl 10xPCR buffer (by Invitrogen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7), 0.6μl 50mM MgCl2, 2μl 2mM dNTPs, 1μl of 10μM primer MBP1left (5'-CAA CGG CTT CCC AGG CAA AGA CGC G3') and 1μl of 10μM primer MBP1right (5'-ATC CCC AGG CAG TTT CCT CTG GAA GG-3'), 1μl (0.25μg) DNA, 0.1μl Taq DNA polymerase (by Invitrogen), with 12.7 μl ddH20 added to a final volume of 20μl. Following a 2 minute denaturation step at 94°C, 35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products was performed with 0.5μl (10U/μl) of the restriction enzyme 0.5μl MluI (10U/μl), 2.5μl 10x NE buffer 3, and 2μl ddH20 to a final volume of 5μl at 37°C overnight. The 121bp product was visualized with ethidium bromide staining under ultraviolet light on a 3% agarose gel following electrophoresis in a 0.5xTBE buffer. MluI cleaves the D allele into two bands (21 bp and 100 bp) and leaves the A allele uncut (Figure 2.4A).  43  MBL2-X/Y genotyping was performed by PCR using sequence-specific priming (SSP) PCR as described by Madsen et al. (9) with some modification. Two upstream primers contain alleles X and Y respectively (MBP-X: 5'-CAT TTG TTC TCA CTG CCA CC-3', MBP-Y: 5'-CAT TTG TTC TCA CTG CCA CG-3') and one downstream primer matches the sequence of MBL2 (5'-ACA TTC CTT GTG ACA CTG CG-3'). A PCR of the β-globulin gene was used as a positive control (Figure 4B). The PCR was performed in a total volume of 20 μl, containing 2μl 10xPCR buffer (by Invitrogen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7), 2μl 2mM dNTPs, 1 μl (0.25μg) genomic DNA, 0.5μl β-globulin-F (sense: 5’-CAA CTT CAT CCA CGT TCA CC-3), 0.5μl β-globulin-R (antisense: 5’-GAA GAG CCA AGG ACA GGT AC-3) and 0.75μl 10μM of the specific primers described above in the presence of 0.6μl 1.5 mM MgCl2 and 1μl HotStarTaqDNA Polymerase (by Qiagen) with 12.4 μl ddH20 added to a final volume of 20μl. PCRs were initiated by a 15 min polymerase activation step at 950C and completed by a final 5 min extension step at 720C. The temperature cycles for the PCRs were as follows: 35 cycles of 20 seconds at 940C, 50 seconds at 650C, and 20 sec at 720C. The 154bp product was visualized with ethidium bromide staining under ultraviolet light on a 1.5% agarose gel following electrophoresis in a 0.5xTBE buffer (Figure 2.4 B). 2.4.4 Pulmonary surfactant protein-A1 gene The SPA-1 polymorphism in exon 4 (C655T; Arg219Trp) was examined by site directed mutagenesis PCR/RFLP using primers which amplified a 159bp region (10). Each PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15mM MgCl2), 2μl 2mM deoxynucleoside triphosphates, 1μl of 10μM sense (5’GCC ATT GCA AGC TTC GTG AA-3’) with a mismatch to produce a TaqI site for the C allele and 10μM antisense (5’-CAC ACA CTG CTC TTT TCC TC-3’) primers, 1μl (0.25μg) DNA, 0.1 hot start Taq DNA polymerase (Qiagen), with 12.9 μl ddH20 added to a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 35 cycles were performed consisting of a 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products with restriction enzyme Taq1 (0.5μl (10U), 2.5μl 10xTaq1 buffer, 0.25μl 100xBSA and 1.75μl ddH20) was at 65°C for three hours. The restriction 44  enzyme cuts the C allele into 2 fragments (138bp and 21bp) and leaves the T allele uncut (159bp). The digested products were electrophoresed on a 3% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.5). 2.4.5 Pulmonary surfactant protein-A2 gene Two polymorphisms were examined in the SPA-2 gene. The polymorphism in exon 4 (A667C; Lys223Gln) was examined by site directed mutagenesis PCR/RFLP using primers which amplified a 235bp region (11). Each PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15mM MgCl2,), 2μl 2mM deoxynucleoside triphosphates, 1μl of 10μM sense (5’-GAG CCT GCA GGT CGG GGA AAA GC-3’) with a mismatch to produce a HhaI site for the C allele and 10μM antisense (5’-CCT CCA GCT CTA ATA GCC ACA AGT-3’) primers, 1μl (0.25μg) DNA, 0.1μl (0.5U) hot start Taq DNA polymerase (5U/μl , QIAGEN), with added water (12.9 μl ddH20) to a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products with restriction enzyme HhaI (0.25μl (5U)), 2.5μl 10xNEB4, 0.25μl 100xBSA and 2μl ddH20 at 37°C overnight. The restriction enzyme cuts the C allele into 2 fragments (24bp and 211bp) and leaves the A allele uncut (235bp). The digested products were separated on a 3% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.6A). The second SPA-2 polymorphism investigated in exon 2 (A26C; Thr9Asn) was examined by site directed mutagenesis PCR/RFLP using primers which amplified a 150bp region (11). Each PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15 mM MgCl2), 2μl 2mM deoxynucleoside triphosphates, 1μl of 10μM sense (5’-GCT GTG CCC TCT GGC CCT TA-3’) with mismatch to produce a MseI site for the A allele and 10μM antisense (5’-TCC TTT GAC ACC ATC TC-3’) primers, 1μl (0.1μg) genomic DNA, 0.1μl hot start Taq DNA polymerase (5U/μl, Qiagen), with 12.9 μl ddH20 added to give a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 35  45  cycles were performed consisting of 30 second denaturation at 95°C, 30 second annealing at 56°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products with 5 μl restriction enzyme Mse1 mixture (0.5μl (10U/μl) Mse1, 2.5μl 10xNEB2 buffer, 0.25μl 100xBSA and 1.75μl ddH20) was performed at 37°C overnight. The restriction enzyme cuts the A allele into 2 fragments (18bp and 132bp) and leaves the C allele uncut (150bp). The digested products were electrophoresed on a 3% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.6 B). 2.4.6 Pulmonary surfactant protein D gene The SPD polymorphism in exon 1 (32C/T, 11Thr/Met) was examined by site directed mutagenesis PCR/RFLP using primers which amplified a 101bp region. Sequence was obtained from Genbank (AH005286). Each PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15mM MgCl2), 2μl 2mM deoxynucleoside triphosphates, 1μl of 10μM sense (5’-CTC CTC TCT GCA CTG GTC CT3’) with mismatch to produce a FspI site for the T allele and 10μM antisense (5’-ACC AGG GTG CAA GCA CTG CG-3’) primers, 1μl (0.1μg) genomic DNA, 0.1μl (0.5U) hot start Taq DNA polymerase (by Qiagen), with added 12.9 μl ddH20 added to a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 35 cycles were performed consisting of 30 second denaturation at 94°C, 30 second annealing at 60°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products with restriction enzyme FspI (0.5μl (2.5U) FspI, 2.5μl 10xNEB4 buffer, 0.25μl 100xBSA and 2μl ddH20) was performed at 37°C overnight. The restriction enzyme cuts the T allele into 2 fragments (82bp and 19bp) and leaves the C allele uncut (101bp). The digested products were electrophoresed on a 3% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.7).  46  2.4.7 Other genes 2.4.7.1 MUC2 gene The MUC2 polymorphism was examined by PCR using primers which amplified the repetitive threonine/serine/proline-rich subdomain (Genbank NM 002457). There is a region of imperfectly conserved repeats (nucleotides 4228-5268 in GenBank accession # NM 002457) in this subdomain. Each PCR reaction contained 2μl 10xPCR buffer (500mM KCl, 200mM Tris-HCl, (pH 8.4)), 0.56μl 50mM MgCl2 (final concentration 1.4mM), 3μl 2mM dNTPs, 2μl of 10μM sense (5’-GTGTCAATTGTTGCTGGCCC-3’ nucleotides 4181-4200 in Genbank NM 002457) and 10μM antisense (5’-CCAGCCAGTCCAATGCAGA-3’ nucleotides 5381-5400 in Genbank NM 002457) primers, 2μl (0.2μg) DNA, 0.1μl (0.5U) Taq DNA polymerase (Invitrogen), and 3μl 50% glycerol, 2μl 0.1% Triton x-100, with 3.34 μl ddH20 added to give a final volume of 20μl. Following a 3 minute denaturation step at 94°C, 40 cycles were performed consisting of 30 second denaturation at 95°C, 30 second annealing at 64°C, 2 minute extension at 70°C and a final 6 minute extension at 72°C. The PCR product was electrophoresed on a 1% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.8). 2.4.7.2 MUC5B gene The variable number tandem repeat (VNTR) in intron 36 of the MUC5B gene (Genbank Y09788) was examined by PCR using primers which amplified the entire VNTR region (nt 4000-4451 - again these numbers are meaningless unless we know which sequence they refer to). Each PCR reaction contained 2μl 10xPCR buffer (500mM KCl, 200mM Tris-HCl, (pH 8.4)), 2μl 2mM deoxynucleoside triphosphates, 2μl of 10μM sense (5’-AGT GTG CAG TGA CTG GCG AG-3’ nt 3967-3986) and 10μM antisense (5’-CTA GAG TTG CAG GTG GCA GG-3’ nt 4655-4674) primers, 0.56μl 50mM MgCl2 (final concentration 1.4mM), 2μl (200ng) DNA, 0.1 (0.5U) Taq DNA polymerase (Invitrogen), 2μl 0.1% Triton x-100, and 3μl 50% glycerol, with 4.34 μl ddH20 added to a final volume of 20μl. Following a 3 minute denaturation step at 94°C, 30 cycles were performed consisting of 30 second denaturation at 95°C, 30 second annealing at 64°C, 2 minute extension at 70°C and a final 6 minute extension at 72°C. The PCR product was electrophoresed on a 1.5% agarose gel in a  47  0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.9). 2.4.7.3 GSTP1 gene The GSTP1 polymorphism in exon 5 (A313G; Ile105Val) was examined by PCR/RFLP using primers which amplified a 433bp region. Each PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15mM MgCl2), 2μl 2mM deoxynucleoside triphosphates, 1μl of 10μM sense (5’-GTA GTT TGC CCA AGG TCA AG-3’) and 10μM antisense (5’-AGC CAC CTG AGG GGT AAG-3’) primers, 1μl (0.1μg) DNA, 0.1μg (0.5U) hot start Taq DNA polymerase (Qiagen), with added ddH20 to a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 15 cycles were performed consisting of 30 second denaturation at 95°C, 30 second annealing at 60°C, 60 second extension at 72°C and another 20 cycles were performed consisting of 30 second denaturation at 95°C, 30 second annealing at 57°C, 30 second extension at 72°C and a final 5 minute extension at 72°C. Digestion of PCR products with restriction enzyme BsmAI (1.5μl BsmAI (7.5U), 2.5μl 10xNEB buffer3, and 1.0μl ddH20) at room temperature overnight. The restriction enzyme cuts the A allele into 2 fragments (104bp and 329bp) and the G allele at into three fragments (104bp, 222bp and 107bp). The digested products were electrophoresed on a 1.5% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light (Figure 2.10 A). 2.4.7.4 GSTM1 and GSTT1 genes The GSTM1 and GSTT1 gene deletions were examined by multiplex PCR as described by Yim and associates (12)using the β-globulin gene as an internal control. The primer pairs for each gene were as follows: GSTM1F (sense):5’-GAA CTC CCT GAA AAG CTA AAG C-3’ GSTM1R (antisense): 5’-GTT GGG CTC AAA TAT ACG GTG G-3’ GSTT1F (sense): 5’-TTC CTT ACT GGT CCT CAC ATC TC-3’ GSTT1R (antisense): 5’-TCA CCG GAT CAT GGC CAG CA-3’ β-Globulin F (sense): 5’-CAA CTT CAT CCA CGT TCA CC-3’ 48  β-Globulin R (antisense): 5’-GAA GAG CCA AGG ACA GGT AC-3’ The PCR reaction contained 2μl 10xPCR buffer (by Qiagen containing KCl, (NH4)2SO4, Tris-Cl, pH 8.7, and 15mM MgCl2), 2μl 2mM deoxynucleoside triphosphates, 0.5μl βglobulin F and 0.5μl β-globulin R, 0.75μl GSTM1-F and 0.75μl GSTM1-R, 0.5μl GSTT1-F and 0.5μl GSTT1-R primers, 1μl (0.1μg) DNA, 0.1μl (0.5U) hot start Taq DNA polymerase (Qiagen), with 11.4 μl ddH20 added to a final volume of 20μl. Following a 15 minute denaturation step at 95°C, 35 cycles were performed consisting of 45 second denaturation at 94°C, 45 second annealing at 60°C, 45 second extension at 72°C and a final 5 minute extension at 72°C. The PCR product was electrophoresed on a 2% agarose gel in a 0.5xTBE buffer and visualized with ethidium bromide staining under ultraviolet light. Visualized on the gel were: a 215bp fragment for GSTM1, a 480bp fragment for GSTT1, and a 268bp fragment for the β-globulin gene (Figure 2.10 B). 2.4.7.5 GCLC gene The trinucleotide repeat guanine-adenine-guanine (GAG) was examined by microsatellite PCR as described by McKone and associates (4) for the Canadian cohort and the Seattle cohort. 2.5. PHENOTYPIC DATA COLLECTION 2.5.1 Pulmonary function The subjects performed post-bronchodilator spirometry (forced expiratory volume in one second (FEV1) and forced vital capacity (FVC)) in accordance with ATS criteria (13). A standard protocol was used by all centers. Values were expressed as a percent of the predicted normal values based on age, gender and height (14). The best-recorded postbronchodilator measurements were used for this study. Predicted values were calculated from equations derived for adults (14) and children (15). For patients who had received lung transplantation, lung function data prior to transplantation were collected and used for statistical analyses.  49  2.5.2 Cross-sectional and longitudinal data sheets Cross-sectional and longitudinal data were collected for the study. A list of variables and the format of the variables are presented in Table 2.3 and 2.4 for cross-sectional data and Table 2.5 for longitudinal data. Data collected for the cross-sectional data sheet included dates to calculate age of CF diagnosis, age of first infection with Pseudomonas aeruginosa and when possible age of chronic infection with P. aeruginosa, age at infection with, and age of chronic colonization with Burkholderia cepacia complex (BCC), age of transplantation and death. We also created categorical variables for the presence of chronic infection with P. aeruginosa to increase the study sample size as in some cases it was not possible to obtain the age of first or chronic infection with this pathogen, but it was possible to categorize them as chronically infected or not.  Additional variables that were collected were CFTR  genotype, sweat chloride levels, pancreatic sufficiency status, and presence/absence of liver disease, gastrointestinal reflux and diabetes mellitus. The subjects’ most current stable clinical status data were entered in the cross-sectional data sheet as well (included lung function, height, weight and date of clinic encounter). Stable clinical status is defined below. Our objective was to collect over 2 years of longitudinal data for each subject and ideally up to 10 years of continuous follow-up data. In some cases, for all the Canadian clinics, we have more than 10 years of clinical data collected and this is for CF patients who became transiently or chronically infected with BCC and data was collected for 2 intervals pre- (2 years) and post-acquisition of BCC (to most current encounter). Ultimately we have collected a minimum of seven to ten years of continuous follow-up data for the Vancouver and Victoria centers. Less than seven years of data were available for 11 (range 2.5-6.7 years) out of 107 patients from the Vancouver adult clinic and 5 out of 11 patients 1.0-3.6 years of data) from the Victoria children’s clinics due to their young age or transfer of a patient from another center in Canada for whom we could not readily obtain the missing data (2 out of 18 from the Adult clinic in Victoria). Ten years of data were collected from the Vancouver B.C. Children’s hospital for 89 out of 122 patients and for the remainder of the study sample due to their young age we were able to collect 2-5 years of clinical data for 25 patients and less than 2 years (1.2-1.9 years) of data for 8 patients. For the Montreal clinic we were able to collect 3-10 years of clinical data. We collected 5-10 years of data on 100 out of the 146 CF patients who volunteered for the study. We collected over 5 years and up to 10 years of clinical data on 61 out of the 92 CF patients who volunteered from the two 50  Hamilton clinics. The Toronto centers only participated in the study of the alpha-1antitrypsin gene and longitudinal data (> 2 years) was not collected for the alpha-1antitrypsin gene analysis. The following data were obtained for the longitudinal data file: pulmonary function (i.e., FEV1 and FVC; absolute and percent predicted values when available), height, weight, bacteriology (the primary pathogens Pseudomonas aeruginosa and BCC, as well as for Staphylococcus aureus, Hemophilus influenza and Xanthomonas maltophilia) and reason for visit. The date the clinical data were collected was recorded as the date of the encounter. These data were obtained from the patients’ medical charts, as was information on frequency and duration of hospitalizations for each year of data collection. In the longitudinal data file, columns were available for entry of the date of admission and discharge from hospital. Hospitalizations were coded based on whether they were for pulmonary infections requiring intravenous (IV) antibiotic therapy, or for other nonpulmonary reasons. In the case of hospitalizations for pulmonary infections, the duration of therapy (start and end date for IV therapy) rather than the duration of hospitalization was collected for the longitudinal data file. Therapy with home IV for pulmonary infections has become the standard mode of treatment in the past seven years, in some cases patients are admitted for 2-5 days and then discharged to continue with IV therapy at home. In other cases patients are started immediately on home IV therapy without a short hospitalization. Clinical data in this instance (i.e., pulmonary function, height and weight) were collected close to the end of IV therapy (1-3 days prior to IV therapy completion) or 2-10 days following completion of IV therapy to represent stable clinical status. Exam date for this encounter is the date of clinical data collection. Bacteriology represents pathogens grown at time of diagnosis with a pulmonary exacerbation. Outpatient encounters were coded as clinic stable or clinic ill (from a respiratory perspective). Since our main phenotype of study was pulmonary status, clinic visits were coded as stable or ill based on a respiratory perspective. Sometimes CF patients may be seen by the clinic during the interval of IV therapy and these encounters are also coded as such in order to identify them and exclude them as pulmonary function is expected to be adversely affected. 51  Longitudinal data collected from Hamilton for visits coded as clinic ill did not distinguish between ill from pulmonary causes versus other (such as weight loss, hemoptysis, liver disease, and diabetes). In cases where all the visits within a 12 month period were coded as clinic ill, the clinic visit with the highest pulmonary function data, which was also similar (higher or equal to) to pulmonary function data from the previous and/or subsequent 12 month interval which was coded as clinic stable was used as the encounter characterizing the patient for this year and accepted as the stable value for this year. 2.5.3 Definition of stable clinical status and pulmonary exacerbation in CF Stable clinical status was defined as the absence of pulmonary exacerbation over the previous 4 weeks, absence of a current mild exacerbation requiring oral antibiotics and the absence of two or more clinical symptoms (increased cough, sputum volume and purulence, increased dyspnea, reduced weight and a fall in FEV1>10%). Pulmonary exacerbation was defined as a pulmonary infection that required the administration of IV antibiotics based on clinical signs assessed by the CF physician. 2.5.4 Identification and typing of B. cepacia Complex Organisms suspected of being BCC were sent to the Canadian B. cepacia complex Research and Referral Repository where BCC confirmation was performed as described in Henry and associates (16). Isolates were checked for purity, and then screened for growth on BCSA agar (17).  Organisms that were BSCA positive were set up to the API Rapid NFT strip  (Biomerieux Vitek Inc., Hazelwood MO), supplemented by glucose, maltose, lactose, adonitol, xylose and sucrose oxidation / fermentation sugars and lysine and ornithine decarboxylases. Organisms that matched the phenotypic identification criteria of Henry and associates (16) were confirmed by species specific molecular probes. Epidemiologic typing of BCC was performed by randomly amplified polymorphic DNA (RAPD) fingerprinting as described in Mahenthiralingam and coworkers (18). Sequential isolates of BCC recovered from individual patients were typed whenever possible over a minimum period of one year to ensure consistency of colonization with a single strain type.  52  2.5.4 Sub-study data collection: Measurement of alpha-1-antitrypsin levels during a pulmonary exacerbation episode The CF patients were recruited exclusively from the St. Paul’s Hospital adult CF clinic. In total 31 consecutive patients from the CF clinic (mean age (±SEM) 27.5(1.1) years), who developed an acute pulmonary exacerbation were recruited to participate in the study. We measured α1-AT levels during the acute phase and 2-3 months later during a stable phase. Details of the experimental procedures, risks and benefits involved were explained to subjects before obtaining written consent, which was approved by the Ethics Committee of the institution in this study (Appendix A). CF patients with a history of liver disease, or liver/lung transplanted were excluded from the α-AT levels study. Patients from the St. Paul’s Hospital CF clinic were recruited at the time of admission for an acute pulmonary exacerbation. These patients were characterized into two groups based on %predFEV1 during clinical stability: mild/moderate pulmonary impairment  (%predFEV1>50%  predicted),  and  severe  pulmonary  impairment  (%predFEV1<50% predicted). For this group we calculated Schwachman-Kulczycki (S-K) (19) and Brasfield (20) scores during clinical stability. S-K scores were calculated for clinical status at the time of testing, which represented stable (non-pulmonary exacerbation) clinical status. Radiographic data were obtained from the patients’ medical charts, as was information on activity patterns, pulmonary and nutritional status for S-K scoring. From a maximal score of 100 (maximum 25 per category), points were deducted for the level or degree of: ƒ  Inactivity, fatigue and non-participation in daily living.  ƒ  Pulmonary symptoms and finger clubbing.  ƒ  Growth and nutritional deficiencies.  ƒ  Chest radiographic abnormalities  S-K scoring was performed by me and the results were reviewed with Dr. Wilcox. The guidelines for scoring were followed as described by Schwachman and Kulczycki (19). Chest radiographs were scored utilizing the Brasfield clinical scoring system (20) by Dr.  53  P.G. Wilcox, who was blinded to the identity of the patient. The severity of lung disease was determined from a maximal score of 25, with points deducted for: ƒ  Degree of hyperinflation.  ƒ  Peribronchial thickening.  ƒ  Nodular cystic structures (bronchiectasis).  ƒ  Areas of atelectasis or pneumonia.  ƒ  Assessment of overall severity.  We also obtained from the patients’ medical charts the number of days treated for pulmonary infections over a 2-year period (number of hospitalization days). Blood samples for measurement of α-AT were obtained within the first two days of hospital admission and repeated at days 4, 7, 10 and 13 of a 14-day therapeutic intervention for most patients. Stable α-AT levels were measured 2-3 months post-exacerbation during clinical stability. These patients were also included in the larger study and were genotyped for the α-AT polymorphisms. 2.6 GENETIC ANALYSES 2.6.1 Haplotype construction To calculated haplotypes we used PHASE version 1.0.1. PHASE is a Linux based program for reconstruction of haplotypes (5). The mean error is half that obtained by the expectationmaximization (EM) algorithm. We used this program to infer haplotypes for the 4 MBL gene polymorphisms and for the pulmonary surfactant polymorphisms investigated. 2.6.2 Calculation of linkage disequilibrium and Hardy Weinberg equilibrium Linkage disequilibrium and Hardy Weinberg equilibrium were calculated using Arlequin version 2.0 software package (21). We used the program PHASE version 2 (5, 6) to infer haplotypes probabilities. We used the chi-square statistic to test for linkage disequilibrium between the MBL and pulmonary surfactant genes.  54  2.7 STATISTICAL ANALYSIS Specific statistical analyses used to investigate the proposed modifier genes are explained in detail in section 2.5 for each modifier gene, or family of genes investigated. We used univariate analysis of variance (ANOVA) for group comparisons. Cox Regression was used for survival analysis. Data analysis was performed using SPSS statistical software (SPSS® statistical software (Chicago, IL)) and Splus (Insightful Corporation).General statistical methods are described in this section. 2.7.1 Description of outcome variables The primary outcome variables used to characterize: ƒ  Pulmonary disease severity was %predFEV1.  ƒ  Pulmonary disease progression was %predFEV1 over time; with time being the longitudinal interval (prospective and retrospective) collected on the study subjects for a maximum 10 year interval (or more for BCC infected CF patients).  ƒ  Survival was the end point of age of death or lung transplantation.  ƒ  First and chronic infection with P. aeruginosa was age of first and chronic infection with the pathogen P. aeruginosa.  ƒ  Frequency of pulmonary infections was the frequency of encounters defined as pulmonary infections that were treated with intravenous antibiotics in the longitudinal data set for set time intervals investigated for statistical analyses (2 and 5 years) as described in section 2.7.6.  Forced expiratory volume in one second (FEV1) was performed in accordance with American Thoracic Society criteria {Anonymous., 1995 #365}. FEV1 was expressed as percent predicted values based on height, age and gender and calculated from equations derived for adults {Crapo, 1981 #363} and children {Hibbert, 1989 #344}. Pulmonary disease progression was defined as %predFEV1 over time; with time being the longitudinal interval (prospective and retrospective) collected on the study subjects. Pulmonary function data from clinical encounters coded as stable were used for statistical analyses. A maximum 10 year interval of clinical data was collected for controls used for BCC infection study (N=196) and 5-10 years of clinical data was collected for controls used in the MBL2 study (N=428). For the BCC infected cohort, clinical data was collected retrospectively and prospectively. Specifically for the BCC patients, FEV1 was retrospectively collected for 2-5 55  years pre-colonization with BCC and data post-acquisition of BCC were collected to July 2004 (or to time of death/lung transplantation). There were 90 (transient / chronic infection N=20/70) and 39 (chronic infection) CF patients infected with BCC that comprised the study group for the BCC infection and MBL2 study, respectively. 2.7.2 Description of independent variables CFTR grouping.We used three different groupings for CFTR mutations. In Chapter 3, the CFTR grouping for the α-AT gene polymorphisms was fairly simplistic (and related to the commonly occurring CFTR mutation deltaF508) as was the case with studies being published at the time. Chapter 3 presents the α-AT gene polymorphism study as published in the journal American Journal of Respiratory and Critical Care Medicine in 1998 (22). During the duration of the thesis study further research in CFTR mutations lead to the categorization of CFTR in the literature into grouping based on the functional affect of the mutations and our groupings for CFTR have incorporated this knowledge. There is still no recognized or established grouping for presenting CFTR mutation classes in research studies. CFTR genotype was coded as follows for the investigation of the α-AT gene: ƒ  Homozygous ΔF508,  ƒ  Heterozygous ΔF508 (i.e., ΔF508/ other)  ƒ  Other.  Other included all other CFTR mutations; known and unknown. For subsequent genes investigated in this study, the CFTR gene was coded based on the CFTR class of the mutation. The following categories were developed based on the severity of the CFTR mutation: ƒ  CF class homozygous severe: Class 1, 2, or 3 mutations on both chromosomes (homozygous for class 1, 2 or 3 mutations).  ƒ  CF class mild: Class 4 or 5 on one chromosome and either: o Class 4 or 5 on the second chromosome. o Class 1, 2, and 3 on the second chromosome. o Unknown or unclassified mutation on the second chromosome.  56  ƒ  CF class heterozygous severe: Class 1, 2, or 3 on one chromosome and unknown or unclassified mutation on the second chromosome.  ƒ  CF class unknown/unclassified: Unknown or unclassified mutations on both chromosomes.  When we analyzed the GST and GCLC gene polymorphisms (chapter 5) the grouping CFTR was collapsed into 2 groups in some cases. This secondary CFTR grouping was used where there was a small number of subjects when grouped by GCLC grouping and mild CFTR which otherwise caused estimation problems with the statistical models. In this grouping we excluded subjects who were classified as CF class unknown/unclassified for two reasons: a) small sample size and b) the unknown nature of the mutation(s) prohibited us from including this group with either the severe or mild CFTR groups described below. The following two CFTR groups were created: ƒ  CF class severe: Included CF subjects classified above as CF class homozygous and heterozygous severe.  ƒ  CF class mild: Same grouping as defined above (see 2).  Pancreatic sufficiency status.In some cases we were unable to use CFTR class grouping in our statistical models as was the case in chapter 4 for BCC related analyses. In this particular case we found that all BCC infected CF patients carried severe CFTR mutations. In this case we also used the variable pancreatic sufficiency status as a measure of disease severity. CF patients were categorized as pancreatic: ƒ  Insufficient: The exocrine pancreas was affected and CF patients were required to take fat metabolizing enzymes orally,  ƒ  Sufficient. The exocrine pancreas was functioning properly.  P. aeruginosa and BCC infection.Categorical variables described the infection status of CF patients for P. aeruginosa and BCC. Specifically CF patients were categorized independently for these two pathogens initially into the following four categories: ƒ  Not infected  ƒ  Infection with the pathogen only on one encounter.  ƒ  Transient (Short-term/sporadic) infection: the CF patient showed infection with the pathogen on more than one occasion but was limited to infection with the pathogen for less than 6 months and thereafter did not show growth of the pathogen.  ƒ  Chronic infection with the pathogen. 57  The grouping was also collapsed for use of the variables in our models into three categories: ƒ  Not infected  ƒ  Transient  ƒ  Chronic  The age (day-month-year) of first and chronic infection with the pathogens were also collected. In the case of BCC infection the pathogen was further characterized by genomovar grouping and RAPD type (see Table 2.3). 2.7.3 Analysis of pulmonary disease progression Mixed effects linear regression models were used to model the effect of the candidate modifier gene genotype or haplotype on our primary outcome variable which we defined as pulmonary disease progression, which is %predFEV1 over the longitudinal interval. Pulmonary disease severity was defined as the mean %predFEV1 over the entire longitudinal interval. Parameters also used as independent variables in our equations were current age, sex (categorized as male (0), female (1)), age of CF diagnosis, CFTR genotype or pancreatic sufficiency status (categorized as insufficient (0) and sufficient (1)), center code (categorized as Vancouver (1), Montreal (2), Hamilton (3), Toronto (4), Victoria (5), and Seattle, Washington, U.S.A (6)), body mass index (BMI) and infection with Pseudomonas aeruginosa. The latter was used as a categorical variable (Pseudomonas aeruginosa (PA) status as infected (1) and not infected (0)) in the mixed effects regression models. The models investigated for each gene polymorphism are presented in Tables 2.6 -2.12. α-AT gene. We investigated 3 polymorphisms in two independent analyses for the α-AT gene. The S and Z polymorphisms were analyzed in one model with α-AT deficiency defined by the genotypes MS, SS and MZ. The base group was MM. The 3 prime mutation in the untranslated region was investigated in the second model and the grouping was GG versus GA and AA. We investigated whether there was a difference in the rate of decline in %predFEV1 (2-year interval). The models and variables used in the analyses are presented in Table 2.6. 58  MBL2 gene. We investigated four polymorphisms in the MBL2 gene. Subjects were grouped for MBL polymorphisms based on the functional effect of the polymorphisms as described in Garred and associates (23) into two groups; deficient and wild-type. We investigated whether there was a difference in the rate of decline in %predFEV1 with P. aeruginosa infection controlling for CFTR genotype, P. aeruginosa infection status and gender. We next investigated whether there was a difference in the rate of decline in %predFEV1 with BCC infection; whether MBL deficiency and chronic infection with BCC were associated with worse pulmonary disease progression in CF. The pre and postacquisition interval for BCC infection were included in the models as fixed and random effects. Patients chronically infected with BCC regardless of genomovar group were grouped together, based on our findings that the rate of decline in %predFEV1 with BCC infection was similar regardless of BCC genomovar. The models and variables used in the analyses are presented in Table 2.7. SPA-1 gene. We investigated the SPA-1 (C655T) polymorphism. The TT genotype was not represented in our study cohort. We investigated whether having the less common CT versus CC genotype was associated with a different rate decline in %predFEV1. The models and variables used in the analyses are presented in Table 2.8. SPA-2 gene. Inferred haplotypes generated by the PHASE program for the two polymorphisms (A26C and A667C) studied in this gene were used to group the study cohort into three groups based on whether there were zero (SPA-2CA0), one (SPA-2CA1), or two (SPA-2CA2) copies of the inferred haplotype CA and investigate whether there was a difference in the rate of decline in %predFEV1. The base group in our models was the most common diplotype (i.e., two copies of CA: SPA-2CA2). The models and variables used in the analyses are presented in Table 2.9. SPD gene. The grouping for the gene polymorphism was based on genotype into three groups: CC, TT and CT. We first examined the rate of decline in %predFEV1 based on our SPD genotype grouping. In a reduced model we combined the CT and CC group, that is having one or two copies of the polymorphism versus being homozygous for the common 59  allele TT, and examined the rate of decline in %predFEV1. The models and variables used in the analyses are presented in Table 2.10. GST genes and GCLC gene. We investigated whether having the gene deletion or not for GSTM1 and T1 were associated with a different rate of decline in %predFEV1. The grouping for the gene deletion was whether subjects had one or two gene deletions for GSTMI and T1. In the case of the GSTP1 analyses our grouping was whether having one or two G-alleles for the GSTP1 (Ile105Val) polymorphism were associated with different rates of decline in %predFEV1. For the GCLC gene we elucidated the number of GAC repeats in our study population and then investigated the commonly occurring ones. Our grouping for GCLC was: ƒ  Homozygosity for 7 GAC repeats (i.e., GCLC7/7),  ƒ  7/8 GAC repeats (i.e., GCLC7/8),  ƒ  7/9 GAC repeats (i.e., GCLC7/9), and  ƒ  Greater than 7 GAC repeats on both chromosomes (i.e., GCLCgr7; includes 8/8, 8/9 and 9/9 GAC repeats).  Genotypes 6/9 and 7/10 GAC repeats which were also observed in our study cohort were rare and were not used in our analyses. We were unable to investigate P. aeruginosa infection status and the GST and GCLC polymorphisms by CFTR genotype severity in all cases due to the small number of patients who were carriers of mild CFTR mutations and who were also not infected chronically with P. aeruginosa, which caused estimation problems with the models. The models and variables used in the analyses are presented in Table 2.11. MUC2 and MUC5B genes. CF patients were categorized into 2 groups for the MUC2 polymorphism: a) heterozygous / homozygous for 2 repeats and b) the common genotype in the cohort (homozygous for 1 repeat). CF patients were grouped based on the common genotype for the cohort for the MUC5B polymorphism into two groups: a) common genotype, which was homozygosity for the 7 repeat allele and b) all others. Lastly, we examined the two genes together and categorized our cohort into two groups; the common diplotype observed (MUC2-MUC5B/ MUC2-MUC5B=1-7/1-7, N=162) and all others (MUC2-MUC5B/ MUC2-MUC5B=1-3/1-7, 1-5/1-7, 1-7/1-8, 2-7/1-7, N=110). CFTR genotype was not included in any of our statistical models as it was not well represented in 60  the mucin gene groupings. We first examined the rate of decline in %predFEV1 based on our individual gene groupings and then for the diplotype grouping. The models and variables used in the analyses are presented in Table 2.12. 2.7.4 Survival Cox proportional hazards regression was used to investigate survival. Models used for each gene are described in the respective chapter and tables. In our survival models: Dependent variables were: The time to event (death or lung transplantation (coded 0/1=alive/deceased or lung transplanted) and current age or age of event. Current age was used for CF patients who were still alive and age of event was used for deceased and lung transplanted CF patients. Main effects were: Modifier gene SNP(s), CFTR class. P. aeruginosa (PA) infection status was also used in select models. Other variables also included in select models were: CF diagnosis age, current %predFEV1, BMI and age. Interactions were: Modifier gene* CFTR class, Modifier gene* PA infection status, Modifier gene* CF diagnosis age. Other variables used as main effect and for interactions terms were: PSS instead of CFTR class BCC infection status and BCC related variables: genomovar group, RAPD group type, and BCC and P. aeruginosa co-infection. 2.7.5 Age of first infection and chronic infection with P. aeruginosa Cox proportional hazards regression was used to investigate age of first infection and chronic infection with P. aeruginosa. chronic infection) with  In our models investigating age of first infection (and  P. aeruginosa):  Dependent variables were: Age of 1st infection (age of chronic infection), categorical variable PA infection status (0/1=not infected/chronically infected).  61  Main effects: Sex + Modifier gene SNP(s) + CFTR class + CF diagnosis age + current %predFEV1 + current BMI + current age. Interactions: Modifier gene* CF diagnosis age, Gene * current %predFEV1. 2.7.6 Pulmonary infections requiring therapy with intravenous antibiotics The Poisson regression model was used to investigate differences in the frequency of pulmonary infections requiring IV therapy and our candidate modifier genes. The response variable was the number of pulmonary infections requiring IV therapy over time. We used a fixed time point from our longitudinal data collection file of January 1, 2000. To ensure that each patient included in the analysis data set was alive or that we had knowledge of the patient’s clinical status (dead or lost to follow-up) we determined that there were clinical data available on the patient after the fixed time point of January 1, 2000. We investigated two longitudinal time intervals: ƒ  Two years length of follow-up (January 1, 1998-January 1, 2000).  ƒ  Up to 5 years length of follow-up (January 1, 1995-January 1, 2000).  The total number of subjects who had the full 2 year length of follow-up was 257. For the 5year interval, cases who met the inclusion criterion for interval A were used. Cases who met inclusion criteria for A were included for analysis of interval B. In our analysis we adjusted the length of follow-up for cases that had less than 5 years of data to the length of follow-up to reflect the shorter interval available. 2.8 HYPOTHESES TESTED α-AT gene Pulmonary disease severity and progression 1.  We hypothesized that heterozygosity for the Z and S alleles, or homozygosity for the S allele of α1-AT would be associated with more severe pulmonary disease severity and progression in CF.  2.  We hypothesized that heterozygosity or homozygosity for the A allele for the 3’ G1237→A  polymorphism of α1-AT would be associated with more severe  pulmonary disease severity and progression in CF. Frequency of pulmonary infections 62  3.  We hypothesized that MZ, MS, SS genotype for α1-AT would be associated with a higher frequency of pulmonary infections requiring IV antibiotics over the 2year interval followed in the CF cohort.  4.  We hypothesized that heterozygosity or homozygosity for the A allele for the 3’ G1237→A polymorphism of α1-AT would be associated with a higher frequency of pulmonary infections requiring IV antibiotics over the 2-year interval followed in the CF cohort  MBL2 gene Pulmonary disease severity and progression 1. We hypothesized that CF patients who have an MBL2 deficient genotype would show more severe pulmonary disease severity and progression (steeper decline) than CF patients with a MBL2 wild-type genotype and this association would be further exaggerated when also controlling for P. aeruginosa infection and CFTR genotype. That is the rate of decline in %predFEV1 would be: MBL2 deficient > MBL2 wild-type 2. We hypothesized that the changes in pulmonary disease severity and progression would occur and be more evident in statistical analyses when distinguishing the time of BCC acquisition and specifically we hypothesized that: a. There would be similar pulmonary disease progression (i.e., rate of decline in pulmonary function) during the pre-acquisition with BCC interval for CF patients who later become chronically infected with BCC compared with non-BCC infected CF patients. That is for the 2-year pre-acquisition with BCC interval we hypothesized that the rate of decline in pulmonary function would be: MBL2 deficient ≠ MBL2 wild-type BCCMBL2deficient = No BCCMBL2deficient BCCMBL2wild-type = No BCCMBL2wild-type  63  b. Once infected with BCC that CF patients with a deficient MBL genotype would show more severe pulmonary disease severity and progression in the post-acquisition interval than CF patients with a: a) MBL2 wild-type genotype who are chronically infected with BCC and b) MBL2 deficient genotype not infected with BCC. That is the rate of decline in %predFEV1 would be: BCCMBL2deficient > BCCMBL2wild-type and BCCMBL2deficient > No BCCMBL2deficient Frequency of pulmonary infections 3. We hypothesized that having an MBL2 deficient genotype compared with MBL2 wild-type would be associated with a different frequency of pulmonary infections requiring IV antibiotics over the 2-year and 5-year interval followed in the CF cohort. Survival (death or lung transplantation) 4. We hypothesized that CF patients with an MBL2 deficient genotype would show worse outcome than MBL wild-type CF patients. 5. We hypothesized that CF patients with an MBL2 deficient genotype who were also chronically infected BCC would show worse outcome. 6. We also hypothesized that CF patients infected chronically with both pathogens (BCC and P. aeruginosa) would be more likely to experience an event than CF patients infected with only one of the pathogens or neither of the pathogens. Susceptibility to BCC and P. aeruginosa infection 7. We hypothesized that MBL genotype is not directly associated with susceptibility to BCC chronic infection. 8. We hypothesized that MBL2 genotype is not associated with a different age of first infection and chronic infection with P. aeruginosa.  64  BCC infection, Genomovar and RAPD type grouping Pulmonary disease severity and progression 1  We hypothesized that infection with BCC would be associated with more severe pulmonary disease severity and steeper pulmonary disease progression compared to CF patients not infected with BCC and specifically that the decline in %predFEV1 over time would be: Chronic BCC ≠ Controls (i.e., non BCC infected) Transient BCC ≠ Controls Chronic BCC ≠ Transient BCC  2  We hypothesized that chronic infection with BCC and P. aeruginosa would be associated with more severe pulmonary disease severity and steeper pulmonary disease progression compared to chronic infection with only one pathogen or infection with neither pathogen. Specifically: Chronic BCC > Controls (i.e., non BCC and P. aeruginosa infected) Chronic P. aeruginosa > Controls (not infected with P. aeruginosa)  Chronic BCC and P. aeruginosa > Chronic BCC or Chronic P. aeruginosa ≠ Controls 3  We hypothesized that infection with BCC genomovar 2 versus BCC genomovar 4 would be associated with different pulmonary disease severity and pulmonary disease progression and specifically: BCC genomovar 2 ≠ Non-BCC infected BCC genomovar 4 ≠ Non-BCC infected BCC genomovar 2 ≠ BCC genomovar 4  4  We hypothesized that infection with BCC genomovar 4 RAPD-type 2 versus BCC genomovar 4 RAPD-type 1,4,6 would be associated with different pulmonary disease severity and pulmonary disease progression. Survival (death or lung transplantation) 65  5  We hypothesized that there would be a different outcome (i.e., rapid deterioration to the event that is death or requiring lung transplantation) associated with CF patients who were either transiently or chronically infected with BCC or not infected with the pathogen.  6  We hypothesized that there would be a different outcome in CF patients associated with chronic BCC infection when also considering co-infection in these patients with P. aeruginosa (PA). Specifically: Chronic BCC and PA ≠ Chronic BCC and No PA ≠ No BCC or PA chronic infection  7  We hypothesized that being chronically infected with BCC genomovar 2 versus BCC genomovar 4 would show a different outcome.  8  We hypothesized that infection with BCC genomovar 4 RAPD-type 2 versus BCC genomovar 4 RAPD-type 1, 4, and 6 would show a different outcome.  SPA-1 gene Pulmonary disease severity and progression 1. We hypothesized that having the less common CT versus CC genotype for the SPA1 polymorphism would be associated with different pulmonary disease severity and pulmonary disease progression in CF. 2. We hypothesized that having the less common CT versus CC genotype for the SPA1 polymorphism and chronic infection with P. aeruginosa would be associated with different pulmonary disease severity and pulmonary disease progression in CF. Susceptibility to P. aeruginosa infection 3. We hypothesized that having the less common CT versus CC genotype for the SPA1 polymorphism would be associated with a different age of first infection and chronic infection with P. aeruginosa. Frequency of pulmonary infections 4. We hypothesized that having the less common CT versus CC genotype for the SPA1 polymorphism would be associated with a different frequency of pulmonary infections requiring IV antibiotics over the 2-year and 5-year interval followed in the CF cohort. 66  Survival (death or lung transplantation) 5. We hypothesized that having the less common CT versus CC genotype for the SPA1 polymorphism and chronic infection with P. aeruginosa would show a different outcome (i.e., rapid deterioration to the event that is death or requiring lung transplantation). SPA-2 gene Pulmonary disease severity and progression 1. We hypothesized that zero versus one versus two copies of the inferred haplotype CA for the SPA-2 gene would be associated with different pulmonary disease severity and pulmonary disease progression in CF. 2. We hypothesized that zero versus one versus two copies of the inferred haplotype CA for the SPA-2 gene and chronic infection with P. aeruginosa would be associated with different pulmonary disease severity and pulmonary disease progression in CF. Susceptibility to P. aeruginosa infection 3. We hypothesized that zero versus one versus two copies of the inferred haplotype CA for the SPA-2 gene would be associated with a different age of first infection and chronic infection with P. aeruginosa. Frequency of pulmonary infections 4. We hypothesized that zero versus one versus two copies of the inferred haplotype CA for the SPA-2 gene would be associated with a different frequency of pulmonary infections requiring IV antibiotics over the 2-year and 5-year interval followed in the CF cohort. Survival (death or lung transplantation) 5. We hypothesized that zero versus one versus two copies of the inferred haplotype CA for the SPA-2 gene and chronic infection with P. aeruginosa would show different outcome (i.e., rapid deterioration to the event that is death or requiring lung transplantation).  67  SPD gene Pulmonary disease severity and progression 1. We hypothesized that the three genotypes for the SPD polymorphism would be associated with different pulmonary disease severity and pulmonary disease progression in CF. 2. We hypothesized that the three genotypes for the SPD polymorphism and chronic infection with P. aeruginosa would be associated with a different pulmonary disease severity and pulmonary disease progression in CF. Susceptibility to P. aeruginosa infection 3. We hypothesized that the three genotypes for the SPD polymorphism would show different susceptibility to first infection and chronic infection with P. aeruginosa. Frequency of pulmonary infections 4. We hypothesized that the three genotypes for the SPD polymorphism would be associated with a different frequency of pulmonary infections requiring IV antibiotics over the 2-year and 5-year interval followed in the CF cohort. Survival (death or lung transplantation) 5. We hypothesized that the three genotypes for the SPD polymorphism and chronic infection with P. aeruginosa would show a different outcome (i.e., rapid deterioration to the event that is death or requiring lung transplantation). GST genes and GCLC gene Pulmonary disease severity and progression 1. We hypothesized that CF patients with some CFTR function (i.e. homozygous or heterozygous for CFTR class IV or V mutations) and who do not have the deletion polymorphism for GSTM1 and GSTT1 will have better pulmonary function and a lower rate of decline in pulmonary function over time compared to CF patients who have the deletion polymorphisms for GSTM1 and GSTT1. 2. We hypothesized those CF patients who are homozygous or heterozygous for isoleucine for the GSTP1 (Ile105Val) polymorphism and have some CFTR function (i.e., mild CFTR genotype) will show increased pulmonary disease severity and a steeper decline in pulmonary function over time compared to CF patients with a mild CFTR genotype who are homozygous for the valine polymorphism for the GSTP1 gene. 68  3. We hypothesized that GSTT1, GSTM1 and GSTP1 polymorphisms resulting in decreased levels of these enzymes, regardless of CFTR genotype will be associated with more severe pulmonary disease severity and progression. 4. We hypothesize that CF patients who have a lower number of GCLC GAC repeats will show worse pulmonary disease severity and a steeper decline in pulmonary function over time. 5. We hypothesized that these associations will be exaggerated in CF patients who have a severe CFTR genotype. Susceptibility to pathogen (BCC or P. aeruginosa) infection 6. We hypothesized that GSTT1, GSTM1 and GSTP1 polymorphisms resulting in decreased levels of these enzymes, regardless of CFTR genotype will be associated with increased susceptibility to chronic infection with common CF respiratory pathogens P. aeruginosa and BCC. 7. We hypothesized that those CF patients who have a lower number of GCLC GAC repeats, regardless of CFTR genotype, will be associated with increased susceptibility to chronic infection with common CF respiratory pathogens. Susceptibility to liver disease 8. The GSTT1 and GSTM1 deletion polymorphisms and GSTP1 (Ile105Val) polymorphism will be associated with susceptibility to CF liver disease. MUC2 and MUC5B genes Pulmonary disease severity and progression 1. We hypothesized that variation in the length of the tandem repeats of MUC2 (i.e., higher number of repeats) would be associated with more severe pulmonary disease severity and progression in CF. 2. We hypothesized that a higher number of repeats of the VNTR of the MUC5B gene would be associated with more severe pulmonary disease severity and progression in CF. Susceptibility to pathogen (BCC or P. aeruginosa) infection 3. We hypothesized that the prevalence of chronic infection with P. aeruginosa or BCC is affected by MUC5B or MUC2 genotype. Specifically we hypothesized that: 4. Polymorphisms which may increase the viscosity of mucus such as higher number of repeats of the VNTR of the MUC5B polymorphism may contribute to earlier repeat 69  infection with common CF pathogens followed by earlier chronic colonization with P. aeruginosa or BCC. 5. A longer length of the tandem repeats of the MUC2 polymorphism may contribute to earlier repeat infection with common CF pathogens followed by earlier chronic colonization with P. aeruginosa or BCC.  70  Table 2.1. A description of participating clinics and contribution to sample size for modifier gene analysis in the study. Center Vancouver Adult Vancouver Children’s Hamilton Health Sciences Adult clinic Hamilton Health Sciences Children’s clinic Montreal Adult Victoria Adult Victoria Children Toronto Adult Toronto Children Seattle Adult  AAT 97  MBL2 Surfactant 107 107  GSTM1/T1/P1 GCLC 108/106/106 106  MUC2/5B 90/80  97  122  122  122/55/55  69  64/10  46  44  45  45  45  43/45  45  47  47  45  45  43/44  146  145  146  146/143/142  146  139/145  0  11  11  11  18  0  0  9  9  9  11  0  187  0  0  0  0  0  106  0  0  0  0  0  0  0  0  101  101  0  71  Table 2.2. Summary of genes and polymorphisms studied and genotyping method utilized. Gene  Polymorphisms  Genotyping method  AAT  S and Z  multiplex SDM / RFLP  AAT  3 prime (G1237→A)  PCR/RFLP  MBL2  B and C allele  PCR-RFLP  MBL2  D allele  SDM PCR  MBL2-X/Y  Promoter  ASP  SPA-1  Arg219Trp (C655T)  SDM PCR/RFLP  SPA-2  Thr9Asn (A26C)  SDM PCR/RFLP  SPA-2  Lys223Gln (A667C)  SDM PCR/RFLP  SPD  11Thr/Met (32C/T)  SDM PCR/RFLP  GSTP1  Ile105Val (A313G)  PCR/RFLP  GSTM1  Gene deletion  multiplex PCR  GSTT1  Gene deletion  multiplex PCR  GCLC  GAC repeat  PCR  MUC2  Imperfectly  conserved PCR  repeats (Thr/Ser/Pro rich subdomain) upstream from 69 bp VNTR MUC5B  59 bp/repeat in intron 36  72  PCR  Table 2.3. Demographic spreadsheet variables (Part 1): Clinical parameters. Name of variable ID code  Number of columns 1  Center code Date of birth Date of CF diagnosis CFTR genotype Sweat chloride level Pancreatic function  1 3 3 2 1 1  Sex  1  Data values  Data type Continuous  Diabetes mellitus Liver disease GI reflux Meconium ileus  1 1 1 1  Transplantation  1  Date of transplant Liver disease enzymes Aspartate aminotransferase (AST) Alanine aminotransferase (ALT) Gamma glutamyltransferase (GGT) Alkaline phosphatase (AP) Home postal code  3  Same value as for longitudinal spreadsheet To be entered by Vancouver center Day-month-year Day-month-year One column for each allele Value in mmol/L 0=insufficient 1=sufficient 0=male 1=female 0=no disease 1=disease present 2=disease not looked for See row above See shaded row above See shaded row above 0=no 1=yes 0=no transplant 1=lung transplant 2=heart-lung transplant 3=liver transplant Day-month-year  1  In μ/L  Continuous  1  In μ/L  Continuous  1  In μ/L  Continuous  1  In μ/L  Continuous  Other complicating conditions  1  Continuous Continuous Continuous String Continuous Categorical Categorical  Categorical Categorical Categorical Categorical Categorical  Continuous  String  73  Table 2.4. Demographic spreadsheet variables (Part 2): Pathogen infection parameters and current status (deceased/alive). Name of variable  Number of columns Date of 1st P. aeruginosa 3 infection Date of P. aeruginosa 3 chronic infection* P. aeruginosa infection 1 status P. aeruginosa infection 1 status Burkholderia cepacia 1 complex (BCC) infection status BCC genomovar group 1  BCC RAPD* group  1  Date of BCC 3 colonization Date of death 3 Primary cause of death 1  Data values  Data type  Day-month-year  Continuous  Day-month-year  Continuous  0=no growth 1=one time growth 2=sporadic growth 3=chronic growth 0=not chronically infected 1=chronically infected 0=no growth 1=one time or short-term growth 2=chronic growth 0=not infected 1=Genomovar I 2=Genomovar II (B. multivorans) 3.1= Genomovar IIIa 3.2= Genomovar IIIb 4= Genomovar IV 5= Genomovar V (B. vietnamesis) 0=not infected RAPD type entered as word Day-month-year  Categorical  Categorical Categorical Categorical  String Continuous  Day-month-year Continuous 1= death due to end stage respiratory Continuous disease (hemoptysis, right heart failure) 2= death following transplantation 3= death due to BCC colonization 4= death due to non-CF cause (e.g. motor vehicle accident) 5= death due to CF but not respiratory related (liver disease, pancreatitis) ** These dates may be the same as for 1st infection with the pathogen in many cases. More recently there has been more aggressive treatment to eradicate this pathogen when patients first grow it and therefore there are cases where the pathogen is successfully eradicated for up to a few years. An attempt was made to elucidate the date of first infection and date of chronic colonization. *RAPD = Random Amplification of Polymorphic DNA 74  Table 2.5. Longitudinal data spreadsheet. Name of variable ID code Name or initials Status at encounter  Date of encounter Height (cm) Weight (kg) Pathogen infection at encounter Staphylococcus aureus P. aeruginosa BCC H. influenza X. maltophilia Pulmonary function FEV1 % predicted FEV1 (%) FVC in liters % predicted FVC (%) Hospital admission date Hospital discharge date Reason for encounter If coding is different than the one presented in column 3 then please supply us with your codes.  Number of columns 1 1 1  3 1 1 1 1 1 1 1 1 1 1 1 3 3 1  Data values  1=alive, 2=transplant, 3=deceased at this encounter, 4=death post-transplant, 5=BCC acquisition Day-month-year In cm In kg 0=not present in sputum 1=present in sputum 0/1 0/1 0/1 0/1 0/1 In liters In percent In liters In percent Day-month-year Day-month-year 1= clinic visit, 2=hospitalization or home IV therapy for pulmonary infection, 3=clinic visit but ill 5=clinic visit, but on Home IV antibiotics, 6= hospitalization for other non-pulmonary complication, 9=specifics of visit / hospitalization not available.  75  Data type Continuous String Categorical  Continuous Continuous Continuous Categorical Categorical Categorical Categorical Categorical Continuous Continuous Continuous Continuous Continuous Continuous Categorical  Table 2.6. Models for statistical analysis of pulmonary disease severity and progression of α-AT polymorphisms for Chapter 3. Gene  Base  polymorphism  model  S and Z  MM  group  in Model α-AT coding: 1=MS, MZ, SS, or for the promoter polymorphism 1=GA and AA  TaqI promoter  GG  polymorphism  %predFEV1= Time + α-AT genotype + Sex  in 3’ region  (0/1=male/female) + CF diagnosis age + CFTR  (G1237→A)  genotype + BMI Alternatively using PSS instead of CFTR genotype %predFEV1= Time + α-AT genotype + Sex (0/1=male/female) + CF diagnosis age + Current Age + Pancreatic sufficiency status + BMI Including PA infection status* %predFEV1= Time + α-AT genotype + Sex + PA infection status + CF diagnosis age + Current Age +CFTR grouping + BMI %predFEV1= Time + α-AT genotype + Sex + Current age + PA infection status + CF diagnosis age + BMI  * +  PA status- P. aeruginosa infection status CFTR genotype is an abbreviation for the CFTR grouping into 3 groups (as described for  α-AT analyses in section 2.4.1).  76  Table 2.7. Models for statistical analysis of pulmonary disease severity and progression of MBL2 gene polymorphisms and the effects of chronic BCC infection (on pre and post BCC acquisition) on pulmonary disease progression for chapter 4. Gene Base group in Model polymorphisms model MBL2 MBL deficient Model 4.2.3-A %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + PA infection status† (0/1=not infected/chronically infected) + CFTR genotype+ + MBL2 deficiency * Time Including 2 and 3 way interactions of covariates. Model 4.2.3-B %predFEV1 = Time + Sex + MBL2 deficiency + PA infection status + BCC infection status (0/1=not infected/chronically infected) + MBL2 deficiency * Time + PA infection status * Time + BCC infection status * Time + MBL2 deficiency * PA infection status * Time + MBL2 deficiency * BCC infection status * Time Model 4.2.3-C (pre-acquisition of BCC) %predFEV1 = Time + Sex + MBL2 deficiency + PA infection status + BCC infection status + MBL2 deficiency * Time + PA infection status * Time + BCC infection status * Time Model 4.2.3-D %predFEV1 = Time + Sex + MBL2 deficiency + BCC infection status + AgePP (0/(>0)=precolonization/days post-colonization) + MBL2 deficiency * AgePP). † PA status- P. aeruginosa infection status +  CFTR genotype is an abbreviation for the CFTR class groupings into 4 groups (see 2.4.1).  77  Table 2.8. Models for statistical analysis of pulmonary disease severity and progression of SPA-1 gene polymorphism for chapter 4. Gene  Base group in Model  polymorphism model SPA-1  CC genotype  Model 4.2.3-E %predFEV1= Time + Sex (0/1=male/female) + SPA-1 genotype (0/1=CC/CT) + SPA-1 genotype * Time). Model 4.2.3-F %predFEV1= Time + Sex + SPA-1 genotype + PA infection status† (0/1=not infected/chronically infected) + SPA-1 genotype * PA infection status + SPA-1 genotype * Time + PA infection status * Time + SPA-1 genotype * PA infection status * Time Model 4.2.3-G %predFEV1= Time + Sex + SPA-1 genotype + PA infection status + CFTR genotype+ + SPA-1 genotype * PA infection status + CFTR genotype * SPA-1 genotype + SPA-1 genotype * Time + PA infection status * Time + CFTR genotype+ * Time + SPA-1 genotype * PA infection status * Time + CFTR genotype * SPA-1 genotype * Time + CFTR genotype * SPA-1 genotype * PA infection status * Time  † +  PA status- P. aeruginosa infection status CFTR genotype is an abbreviation for the CFTR class groupings into 4 groups (see 2.4.1).  78  Table 2.9. Models for statistical analysis of pulmonary disease severity and progression of SPA-2 gene polymorphism for chapter 4. Gene  Base group in model  Model  SPA-2  2 copies of CA.  Model 4.2.3-H %predFEV1 = Time + Sex (0/1=male/female) + SPA-2CA0 (0/1=1 or 2 copies of CA/0 copies of CA) + SPA-2CA1 (0/1=0 or 2 copies of CA/1 copy of CA) + SPA-2CA0 * Time + SPA-2CA1 * Time. Model 4.2.3-I %predFEV1 = Time + Sex (0/1=male/female) + SPA-2CA0 (0/1=1 or 2 copies of CA/0 copies of CA) + SPA-2CA1 (0/1=0 or 2 copies of CA/1 copy of CA) + PA infection status + SPA-2CA0 * PA infection status + SPA-2CA1 * PA infection status + SPA-2CA0 * Time + SPA2CA1 * Time + PA infection status * Time + SPA-2CA0 * PA infection status * Time+ SPA2CA1 * PA infection status * Time.  79  Table 2.10. Models for statistical analysis of pulmonary disease severity and progression of SPD gene polymorphism for chapter 4. Gene  Base group in model  Model  SPD  TT genotype  Model 4.2.3-J %predFEV1 = Time + Sex (0/1=male/female) + SPD-CC (0/1=CT or TT/CC) + SPD-CT (0/1=CC or TT/CT) + SPD-CC * Time + SPD-CT * Time. Model 4.2.3-K %predFEV1 = Time + Sex + SPD-CC (0/1=CT or TT/CC) + SPD-CT (0/1=CC or TT/CT) + PA infection status + SPD-CC * PA infection status + SPD-CT * PA infection status + SPD-CC * Time + SPD-CT * Time + PA infection status * Time + SPD-CC * PA infection status * Time+ SPD-CT * PA infection status * Time. Reduced model: CT and CC genotypes combined Model 4.2.3-L %predFEV1 = Time + Sex (0/1=male/female) + SPD-CC/CT (0=TT and 1=CC/CT) + SPDCC/CT * Time.  80  Table 2.11. Models for statistical analysis of pulmonary disease severity and progression of GSTs and GCLC polymorphisms for chapter 5. Gene GSTM1 and T1  Base group in Model model Zero null in Longitudinal data set GSTM1 or T1 Model 5.2.3-A %predFEV1 = Time + Sex (0/1=male/female) + 1 null in GSTM1 or T1 (0/1=0 or 2 null copies/1 null copy) + 2 null in GSTM1 and T1 (0/1=0 or 1 null copies/2 null copies) + 1 null in GSTM1 or T1 * Time + 2 null in GSTM1 or T1 * Time Cross-sectional data set Model 5.2.4-A  GSTP1  GCLC  AA genotype  Current %predFEV1 = Current age (yrs) + Sex + CFTR genotype (0/1=severe/mild) + BMI + PA infection status (0/1=not infected /chronically infected) + 1 null in GSTM1 or T1 + 2 null in GSTM1 and T1 Model 5.2.3-B  7/7 GAC repeats  %predFEV1 = Time + Sex (0/1=male/female) + GG genotype for GSTP1 (0/1=0 or 1 copy of G-allele/ GG) + AG genotype for GSTP1 (0/1=0 or 2 copies of Gallele/ AG) + GG genotype for GSTP1 * Time + AG genotype for GSTP1 * Time Model 5.2.3-C %predFEV1 = Time + Sex (0/1=male/female) + GCLC7/8 + GCLC7/9 + GCLCgr7 (0/1 1=genotypes with >7GAC repeats on both chromosomes) + GCLC7/8 * Time + GCLC7/9 * Time + GCLCgr7 * Time  81  Table 2.12. Models for statistical analysis of pulmonary disease severity and progression for MUC2 and MUC5B polymorphisms for chapter 6. Gene MUC2  MUC5B  Base group in Model model Heterozygous / Model 6.2.3-A homozygous for 2 repeats %predFEV1 = Time + MUC2 (0/1= homozygous for one repeat) + Sex (0/1=male/female) + MUC2 * Time  other than 7-7 VNTR genotype  Model 6.2.3-B %predFEV1 = Time + MUC2 + MUC2 * Time + PA infection status (0/1=not infected /chronically infected) + MUC2 * Time + PA infection status * MUC2 + PA infection status * Time + PA infection status * MUC2 * Time Model 6.2.3-C %predFEV1 = Time + MUC5B (0/1= 7-7 VNTR genotype) + Sex (0/1=male/female) + MUC5B * Time Model 6.2.3-D %predFEV1 = Time + Sex + MUC5B + PA infection status (0/1=not infected /chronically infected) + MUC5B * Time + PA infection status * MUC5B + PA infection status * Time + PA infection status * MUC5B * Time  MUC2/ MUC5B diplotype  1-7/1-7 repeats  Model 6.2.3-E %predFEV1 = Time + Sex (0/1=male/female) + MUC2/5B diplotype (0/1=1-7/1-7) + MUC2/5B diplotype * Time Model 6.2.3-F %predFEV1 = Time + Sex (0/1=male/female) + MUC2/5B diplotype + PA infection status (0/1=not infected /chronically infected) + MUC2/5B * Time + PA infection status * MUC2/5B + PA infection status * Time + PA infection status * MUC2/5B * Time  82  Figure 2.1.α1-AT S and Z alleles; TaqI digestion and visualization of PCR product on an agarose gel. On Gel A, positive controls are in columns 1-3 and column 4 is the negative control (i.e., no template DNA). Subjects in columns 4, 6 and 7 are homozygous wild-type for the Z and S alleles. Subjects in columns 4 and 5 are heterozygous for the S allele (i.e., MS MM). On Gel B, the subject in column 2 is heterozygous for the Z allele (MM MZ), the subject in column 9 is homozygous for the S allele (SS MM), and all other subjects are wildtype for both the S and Z polymorphisms. A. 1  Z 144bp M 123bp S 98bp M 78bp  H20  2  3  4  5  6  7  ˆ ˆ ˆ ˆ  B. 1  Z 144bp M 123bp S 98bp M 78bp  2  3  4  5  6  ˆ ˆ ˆ ˆ  83  7  8  9  10  7  Figure 2.2. α1-AT 3 prime polymorphism; TaqI digestion and visualization of the PCR product on an agarose gel. The TaqI enzyme cuts the wild-type G allele into two fragments (of size 191 and 182bp) and the A allele is left uncut (373 bp fragment). In the gel shown, subjects 1-4, 6, 7, 9 are homozygous for the G allele, subject 5 is homozygous for the A allele and subject 8 is heterozygous for the G allele.  1  2  3  4  5  6  7  8  9  ← A allele (373 bp)  ← G allele: 191 bp 182 bp  84  Figure 2.3. MBL2 gene B and C alleles; the B (codon 54) and C (codon 57) alleles for MBL2 gene were detected by restriction enzymes BanI and MboII, respectively. A. BanI cleaves the A allele into two fragments (245 bp and 84 bp) and leaves the B allele undigested. Subjects 1, 2, 4-6 and 8 are homozygous wild-type (AA), subject 3 is heterozygous wild-type (AB) and subject 7 is homozygous for the B allele. B. MboII cleaves the C allele into two fragments (266bp and 63bp) and leaves the A allele undigested. Subjects 1 and 3-9 are homozygous wild-type (AA), subjects 2, 10 are heterozygous wildtype (AC). A. 100bp 1  2  3  4  5  6  7  8  329bp 245bp 84bp  B. 100bp 1  2  3 4  5  6  7  8  9  10  329bp 266bp  85  Figure 2.4. The D allele and the XY promoter polymorphisms for MBL2 gene. A. The MluI restriction enzyme site was introduced into the amplification product of the D allele (codon 52) by site-directed mutagenesis (SDM)-PCR; MluI cleaves D allele into two bands (21 bp and 100 bp) and leaves A allele uncut. Subject 1 is homozygous for the D allele, subjects 25, 7-11 are homozygous wild type (AA), subjects 6,12 are heterozygous wild-type (AD) and. B. The MBL promoter polymorphism was investigated by PCR using sequence-specific priming. β-globulin was used as a positive control. The presence of a band on the gel at 154bp indicates the X or Y allele was present. In this gel, subjects 2, 5, 7, 9, 10 have at least one X allele, whereas subjects 1, 3, 4, 6, and 8, who do not show a band at 154bp, are homozygous for the Y allele. A. 100bp 1  2  3  4  5  6  7  8  9  10 11 12  121bp 100bp  B. 1  2  3  4  5  6  7  8  9 10  N  β-globulin 154bp  86  Figure 2.5. Site directed mutagenesis PCR/RFLP for the SPA-1 polymorphism. A 159bp region of the gene was amplified and the restriction enzyme Taq1 cuts the C allele into 2 fragments (138bp and 21bp) and leaves the T allele uncut (159bp). Subjects 1, 4, 6, 7 are heterozygous (CT) and subjects 2, 3, 5, 8-10 are homozygous for the C allele.  1  2  3  4  5  6  7  8  9  10  N  159 bp 138 bp  21bp  87  Figure 2.6. Exon 2 and 4 polymorphisms investigated in the SPA-2 gene. A. Site directed mutagenesis PCR/RFLP for SPA-2 (A667C) polymorphism. A 235bp region of exon 4 of the gene was amplified and the restriction enzyme HhaI cuts the C allele into 2 fragments (24bp and 211bp) and leaves the A allele uncut (235bp). Subjects 1,3, 4, 8-10 are homozygous for the C allele, subjects 2,6,7 are heterozygous (AC) and subject 5 is homozygous for the A allele. B. A 150bp region of the gene on exon 2 was amplified and the restriction enzyme Mse1 cuts the A allele into 2 fragments (18bp and 132bp) and leaves the C allele uncut (150bp). Subjects 1, 5 are homozygous for the A allele, subject 2, 4, and 6 are heterozygous (AC) and subject 3 is homozygous for the C allele. A. 1  2  3  4  5  6  7  8  9  10  235 bp 211 bp  B. 1  2  3  4  5  6  150bp 132bp  88  Figure 2.7. Site directed mutagenesis PCR/RFLP of SPD polymorphism in exon 4. A 235bp region of exon 4 of the gene was amplified and the restriction enzyme FspI cuts the T allele into 2 fragments (82bp and 19bp) and leaves the C allele uncut (101bp). Subjects 2, 6, and 7 are homozygous for the C allele, subjects 1 and 3-5 are heterozygous (CT) and subjects 8 and 9 are homozygous for the T allele.  1  2  3  4  5  6  7  8  9  101 bp 82 bp  89  Figure 2.8. MUC2 polymorphism was examined by PCR using primers which amplified the repetitive threonine/serine/proline-rich subdomain. Subjects differ in the number of imperfectly conserved repeats which visualized on an agarose gel as differences in the size of the amplified region (subjects 1 and 3 are heterozygous for 2 long repeats, subjects 2 and 5 are homozygous for 2 long repeats and subject 4 is homozygous for medium repeat). N = negative control. SL  1  2  3  4  5  N  100bp  1500bp 1200bp 600bp  100bp  90  Figure 2.9. MUC5B polymorphism was examined by PCR using primers which amplified the repetitive threonine/serine/proline-rich subdomain. Subjects differ in the number of imperfectly conserved repeats which is visualized on an agarose gel as differences in the size of the amplified region. Subject 1 is homozygous for 7 repeats, subject 2 has 7/5 repeats, subject 3 has 3 repeats, subject 4 has 7/8 repeats, subject 5 has 5 repeats, subject 6 has 8 repeats, subject 7 has 7/6 repeats, subject 8 has 2/3 repeats, subject 9 has 7/4 repeats. N = negative control.  100bp  1  2  3  4  5  6  7  8  9  N  100bp  700bp 600bp 500bp 400bp 300bp 200bp 100bp  91  Figure 2.10. Polymorphisms investigated in GST genes P1, T1 and M1. A. PCR/RFLP of the GSTP1 polymorphism in exon 5 (A313G). A 433bp region of exon 5 of the gene was amplified and the restriction enzyme BsmAI cut the A allele into 2 fragments (104bp and 329bp) and the G allele into three fragments (104bp, 222bp and 107bp). Subjects 1, 3, 6, 8 are homozygous for the A allele, subjects 2, 4-5, 7, 9 are heterozygous (AG) and subjects 10 and 11 are homozygous for the G allele. B.  Multiplex PCR of the GSTM1 and GSTT1  deletions were examined by multiplex PCR using primers which amplified regions of exon 5 and 3. Genotype is shown in table beside the gel. Subjects 1 and 3 are null for GSTM1 and subject 5 is null for GSTT1. A. 1  2  3  4  5  6  7  8  9 10 11  329 bp 222 bp 107 bp 104 bp  B. 100bp  1  2  3  4  5  6  GST-T1 GST-M1 GST-T1 480bp β-Globulin 268bp GST-M1 215bp  92  1 2 3 4 5 6  + + + + +  + + + +  2.9 BIBLIOGRAPHY 1.Gum, J. R. J. 1992. Mucin genes and the proteins they encode: structure, diversity, and regulation. Am J Respir Cell Mol Biol. 7(6):557-64. 2.Pigny, P., V. Guyonnet-Duperat, A. S. Hill, W. S. Pratt, S. Galiegue-Zouitina, M. C. d'Hooge, A. Laine, I. Van-Seuningen, P. Degand, J. R. Gum, Y. S. Kim, D. M. Swallow, J. P. Aubert, and N. Porchet. 1996. Human mucin genes assigned to 11p15.5: identification and organization of a cluster of genes. Genomics 38(3):340-52. 3.Pigny, P., I. Van Seuningen, J. L. Desseyn, S. Nollet, N. Porchet, A. Laine, and J. P. Aubert. 1996. Identification of a 42-kDa nuclear factor (NF1-MUC5B) from HT-29 MTX cells that binds to the 3' region of human mucin gene MUC5B. Biochemical & Biophysical Research Communications 220(1):186-91. 4.McKone, E. F., J. Shao, D. D. Frangolias, C. L. Keener, C. A. Shephard, F. M. Farin, M. R. Tonelli, P. D. Pare, A. J. Sandford, M. L. Aitken, and T. J. Kavanagh. 2006. Variants in the glutamate-cysteine-ligase gene are associated with cystic fibrosis lung disease. Am J Respir Crit Care Med 174(4):415-9. 5.Stephens, M., N. J. Smith, and P. Donnelly. 2001. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 68(4):978-89. 6.Stephens, M., and P. Donnelly. 2003. A comparison of bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet 73(5):1162-9. 7.Sandford, A. J., T. Chagani, J. J. Spinelli, and P. D. Pare. 1999. alpha1-antitrypsin genotypes and the acute-phase response to open heart surgery. Am J Respir Crit Care Med 159(5 Pt 1):1624-8. 8.Madsen, H. O., P. Garred, J. A. Kurtzhals, L. U. Lamm, L. P. Ryder, S. Thiel, and A. Svejgaard. 1994. A new frequent allele is the missing link in the structural polymorphism of the human mannan-binding protein. Immunogenetics 40(1):37-44. 9.Madsen, H. O., P. Garred, S. Thiel, J. A. Kurtzhals, L. U. Lamm, L. P. Ryder, and A. Svejgaard. 1995. Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein. J Immunol 155(6):3013-20. 10. White, R. T., D. Damm, J. Miller, K. Spratt, J. Schilling, S. Hawgood, B. Benson, and B. Cordell. 1985. Isolation and characterization of the human pulmonary surfactant apoprotein gene. Nature 317(6035):361-3. 11. Katyal, S. L., G. Singh, and J. Locker. 1992. Characterization of a second human pulmonary surfactant-associated protein SP-A gene. Am J Respir Cell Mol Biol 6(4):446-52. 12. Yim, J. J., G. Y. Park, C. T. Lee, Y. W. Kim, S. K. Han, Y. S. Shim, and C. G. Yoo. 2000. Genetic susceptibility to chronic obstructive pulmonary disease in Koreans: combined analysis of polymorphic genotypes for microsomal epoxide hydrolase and glutathione Stransferase M1 and T1. Thorax 55(2):121-5. 13. Anonymous. 1995. Standardization of Spirometry, 1994 update. American Thoracic Society. American Journal of Respiratory and Critical Care Medicine 152(3):1107-36. 14. Crapo, R. O., A. H. Morris, and R. M. Gardner. 1981. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 123(6):65964. 15. Hibbert, M. E., A. Lannigan, L. I. Landau, and P. D. Phelan. 1989. Lung function values from a longitudinal study of healthy children and adolescents. Pediatric Pulmonology 7(2):101-9.  93  16. Henry, D. A., E. Mahenthiralingam, P. Vandamme, T. Coenye, and D. P. Speert. 2001. Phenotypic methods for determining genomovar status of the Burkholderia cepacia complex. J Clin Microbiol 39(3):1073-8. 17. Henry, D. A., M. E. Campbell, J. J. LiPuma, and D. P. Speert. 1997. Identification of Burkholderia cepacia isolates from patients with cystic fibrosis and use of a simple new selective medium. J Clin Microbiol 35(3):614-9. 18. Mahenthiralingam, E., M. E. Campbell, D. A. Henry, and D. P. Speert. 1996. Epidemiology of Burkholderia cepacia infection in patients with cystic fibrosis: analysis by randomly amplified polymorphic DNA fingerprinting. J Clin Microbiol 34(12):2914-20. 19. Shwachman, H., and L. Kulczycki. 1958. Long-term study of one hundred five patients with cystic fibrosis. American Journal of Diseases of Children 96:6-15. 20. Brasfield, D., G. Hicks, S. Soong, and R. E. Tiller. 1979. The chest roentgenogram in cystic fibrosis: a new scoring system. Pediatrics 63(1):24-9. 21. Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin: A software for population genetics data analysis., 2.000 ed. Genetics and Biometry Lab, Dept. of Anthropology, University of Geneva., Geneva. 22. Frangolias, D. D., J. Ruan, P. J. Wilcox, A. G. Davidson, L. T. Wong, Y. Berthiaume, R. Hennessey, A. Freitag, L. Pedder, M. Corey, N. Sweezey, J. Zielenski, E. Tullis, and A. J. Sandford. 2003. Alpha 1-antitrypsin deficiency alleles in cystic fibrosis lung disease. Am J Respir Cell Mol Biol 29(3 Pt 1):390-6. 23. Garred, P., T. Pressler, H. O. Madsen, B. Frederiksen, A. Svejgaard, N. Hoiby, M. Schwartz, and C. Koch. 1999. Association of mannose-binding lectin gene heterogeneity with severity of lung disease and survival in cystic fibrosis. J Clin Invest 104(4):431-7.  94  CHAPTER 3: ALPHA-1-ANTITRYPSIN DEFICIENCY ALLELES IN CYSTIC FIBROSIS LUNG DISEASE  A version of this chapter has been published and the citation is: ‘Frangolias D.D., Ruan J., Wilcox P.G., Berthiaume Y., Davidson G., Hennessey R., Corey M., Tullis E., Zielenski J., Wilson W.M., Freitag A., Sandford A.. Alpha-1-antitrypsin deficiency alleles in cystic fibrosis lung disease. American Journal of Respiratory Cell and Molecular Biology. 29:390-396, 2003.’  95  3.0. INTRODUCTION Polymorphisms in the serine protease inhibitor gene, alpha-1-antitrypsin (α1-AT) were investigated. Polymorphisms in the gene which have been shown to affect either the levels of the protein produced or regulate expression of the gene were investigated. In this study we investigated the associations of these polymorphisms on pulmonary disease severity and progression, survival and succeptibility to infection with common CF pathogens. 3.1. RATIONALE AND MAIN HYPOTHESIS The primary pathophysiological processes responsible for premature death and disability in patients with cystic fibrosis (CF) are chronic pulmonary infection and inflammation. The inflammatory process in response to pulmonary infection in CF airways is characterized by a massive influx of neutrophils (1). Neutrophils represent less than 5% of the cells recovered in bronchoalveolar lavage fluid (BALF) in normal individuals but in adults and children (1-5 years of age) with CF, neutrophils may comprise up to 95% of the cell population (2). Neutrophils contain a number of proteolytic enzymes one of which, neutrophil elastase (NE), has been implicated in the excessive pulmonary damage observed in cigarette smokers and in patients with CF. Elevated levels of NE have been reported in the sputum of patients who have CF (3, 4). NE is capable of causing direct lung damage by hydrolyzing all the major connective tissue proteins that make up the lung and airway matrix. In additition, excess NE adversely affects the airways in CF, by enhancing mucous secretion (5, 6), and by interfering with the opsonization and elimination of bacterial pathogens, particularly P. aeruginosa (5, 7). In normal hosts, the actions of NE are prevented primarily by alpha-1antitrypsin (α1-AT), a serine protease inhibitor that binds to NE and inhibits the breakdown of elastic tissue in the lung. Normal to elevated levels of α1-AT have been reported in airway secretions (2) and plasma (2, 8) in patients who have CF. Elevated levels of α1-AT have been reported during pulmonary infections in this patient population (9). The extremely high levels of NE in the airways of CF patients clearly indicate that there is an imbalance between α1-AT and elastase in the airways of patients with CF. Since great variation in disease severity and progression exists among CF patients possessing the same 96  cystic fibrosis transmembrane conductance regulator (CFTR) genotypes (10, 11) it is possible that genes other than the CFTR may contribute to pulmonary disease progression. If this were the case, individuals who have lower than normal levels of α1-AT may be at increased risk for lung damage. Several mutations of the α1-AT gene result in a deficiency of this antiprotease. There is also evidence that α1-AT genotype influences the acute phase response (12). Findings to date are inconclusive concerning the role that α1-AT may play in pulmonary disease progression in CF (13-16). The main limitation of these studies is their small sample sizes and therefore the high possibility of type 2 error (false negative). The purpose of this study was twofold. First, to investigate whether the α1-AT gene (the Z, S deficiency alleles and the 3’ G1237→A mutation) is a modifier of pulmonary disease progression in a large cohort of CF patients who were characterized by a heterogeneous severity of pulmonary disease? Second, we wanted to measure α-AT levels during a stable clinical phase and the acute increases during pulmonary exacerbations in an adult group of CF patients who had varying degrees of pulmonary dysfunction. The rationale was to characterize the acute phase response to pulmonary infection in this patient group to help tailor the possible future administration of anti-proteolytic therapeutic agents. Our main hypothesis was that heterozygosity for the Z, S and/or the 3’ G1237→A alleles of α1-AT would result in earlier onset of pulmonary disease, more rapid deterioration in pulmonary function and consequently more severe pulmonary dysfunction after controlling for other known predictors of pulmonary function decline.  97  3.2 RESULTS Tables 3.1 and 3.2 show demographic and clinical measures for our study cohort stratified by α1-AT genotype. Prevalences of the S and Z alleles were similar to expected frequencies (Z = 1-3% and S = 2-4%) in the normal Caucasian population(17-21). Heterozygosity for the 3’ G1237→A mutation has been documented from smaller studies to be 5-15% (22, 23). Figures 3.1 and 3.2 show %predFEV1 and sample size by α1-AT genotype. Z and S alleles of the α1-AT gene did not predict pulmonary disease severity and were not included in the final regression equation (R2adj=0.23, p=0.0001) Model 3.2-A %predFEV1=70.1 + 0.76*CF diagnosis age – 1.29*Exam Age +7.43*Pancreatic sufficiency status + 0.71*BMI Base group for α1-AT was: MM genotype The effect of the variable Center was included as a random effect; however, center to center variability did not reach statistical significance (was less than 1% of the total variability). To determine whether inclusion of patients < 18 years of age (who had less severe pulmonary disease) could reduce the power to detect a significant association with a genetic modifier, the analysis was repeated in the subset of patients over 18 years of age. This analysis confirmed the lack of association of the Z and S alleles with disease severity seen in the entire group (α1-AT genotype p=0.96). To determine whether there was a specific interaction of α1-AT genotype with ∆F508, which may have been obscured by the presence of other CFTR alleles, we also ran this model selecting only those who were homozygous ∆F508, however α1-AT genotype was not a significant predictor. In the second model we included the categorical variable P. aeruginosa infection status, available for 555 of the 714 study subjects. This model accounted for 25% of the variability in cross-sectional pulmonary disease severity (R2adj=0.25, p=0.0001). Model 3.2-B  98  %predFEV1=79.21 – 1.23*Exam Age – 9.56* P. aeruginosa infection status + 0.66*CF diagnosis age + 0.50*BMI Burkholderia cepacia complex (BBC) colonization status (i.e., not infected (0) versus chronic infection with the pathogen (1)) was available for 558 patients (from the Vancouver, Hamilton and Toronto centers) and was evaluated as a predictor of pulmonary disease severity. This parameter was not a significant predictor of pulmonary disease in our cohort. The Toronto cohort did show higher prevalence of chronic BCC infection (26% of their sample) compared to the Hamilton and Vancouver cohorts (9.5% and 10.2% of the study samples, respectively). Similar analyses were used to investigate the 3’ G1237→A mutation in the α1-AT gene and results showed that this mutation also was not predictive of pulmonary disease severity. The base group used for this polymorphism in the mixed effects regression analysis was GG genotype. We did not show increased prevalence of death or lung transplantation in CF patients who were homozygous or heterozygous for the Z, S or the 3’ G1237→A alleles versus wild type (Tables 3.1 and 3.2). CF patients who had died or who were lung transplant recipients were significantly older (mean (±SEM) = 28.6(±1.3) years) than CF patients who were still alive (21.3(±0.4) years; p=0.0001), although both groups had been diagnosed with CF at similar mean (±SEM) ages (3.7(±0.8) and 4.5(±0.3) years respectively, p=0.44). We showed similar frequency of pulmonary exacerbations and duration of IV therapy (over 12 months) in the α1-AT genotypic groups (Tables 3.1 and 3.2). BCC colonization was not more prevalent in those CF patients who were carriers of the Z, S or 3’ alleles (data not shown). Liver disease status was available for 553 of the study subjects. We did not show increased prevalence of the Z or S alleles in patients identified with liver disease (7 out of 76) compared to wild type patients (45 out of 477; p=0.58). Specifically, only 3 of the 16 patients heterozygous for the Z allele were identified with liver disease. Table 3.3 shows the anthropometric, clinical and lung function data as well as stable and acute phase increases in α-AT levels in our sub-study group investigating α-AT levels. As expected, there were significant differences between groups in measures of disease severity (i.e., %predFVC, S-K and Brasfield scores). The stable status levels of α-AT were within 99  the normal range for the mild/moderate group and outside the upper limit of the normal range in our laboratory for the severe group (Figure 3.3). The peak values for α-AT during pulmonary exacerbation were significantly elevated above normal in both groups, and declined at a similar rate throughout the 14-day intervention. Although the levels in patients with severe pulmonary disease were lower at most time points neither group returned to normal in the 14-day period (Figure 3.3). The percent change in α-AT levels from peak to stable was greatest in patients with mild/moderate pulmonary disease, whose mean value was within the upper limit of the normal range (0.95-1.77 g/L) at the time of stable status, whereas that of patients with severe pulmonary disease was not. There were no significant associations between spirometric and clinical measures and stable or percent changes in α-AT levels. The percent change in α-AT levels was positively skewed and was log base 10 transformed prior to regression analysis. Stepwise regression identified BMI as a predictor of percent change in α-AT (%change in α-AT =50.5+3.5*BMI; adjusted R2=0.08, p=0.05). Significant correlation coefficients were shown for BMI and stable α-AT (r=-0.34, p=0.04) and percent change in α-AT (r=0.32, p=0.05). When we ran the above analysis selecting out the five S/Z subjects who were all in the severe group, there was no change in the model.  100  3.3 DISCUSSION The results of this study indicate that the α1-AT gene is not a modifier gene in CF. In the large study group Z and S polymorphisms were not associated with increased pulmonary disease severity as defined by pulmonary function, frequency and duration of pulmonary exacerbations, death or lung transplantation. Similarly, the A allele of the 3 prime mutation in the α1-AT gene was not protective. We showed a blunted acute phase increase in α1-AT to pulmonary infection in those CF patients in a malnourished state. Since we recruited consecutive patients for the α-AT levels study as they were admitted for a pulmonary exacerbation it is conceivable that our cohort is biased towards more severe CF patients who require more frequent IV antibiotic therapy. However patients with mild pulmonary disease infrequently require admissions for pulmonary exacerbations and therefore such a bias is difficult to avoid. The α1-AT MZ genotype has been shown to be a risk factor for COPD (24). In CF cohorts, the association of α1-AT genotype and pulmonary disease severity is unclear. Doring and colleagues (13) found no association between α1-AT S and Z alleles and pulmonary disease severity but they did show an earlier age of onset of P. aeruginosa infection in individuals with these deficiency alleles (6 out of a total sample of 215). Mahadeva and associates questioned this association and in fact showed that patients who were heterozygous for the S and Z alleles (19 out of a total sample of 147) had higher levels of pulmonary function than the wild type individuals (15). In another study, the same authors showed that the α1-AT Z and S deficiency alleles were not more prevalent in those CF patients with severe pulmonary disease (dead or lung transplanted CF patients)(14). We also showed no difference in outcome (death or transplant) for our cohort, but our cohort was followed prospectively for a relatively short time interval (5.5 years). Kalsheker and associates (22) showed that the 3’ mutation was associated with COPD (22). Morgan and colleagues provided in vitro evidence that the association with COPD may be due to deficiency in the α1-AT acute phase response (12).  However, Sandford and  associates (23) did not show that the 3’ mutation attenuated the acute phase rise in α1-AT in patients undergoing open-heart surgery. Similarly, Madadeva et al. showed that the 3' 101  mutation had no effect on α1-AT levels in CF patients (15). In a recent study by Henry and associates (16) they showed less severe pulmonary disease and fewer infective pulmonary exacerbations over 2 years in CF patients who were heterozygous for the A allele. These data suggest that heterozygotes may have a slower disease progression. The results of our study (which also included 7 homozygous individuals for the A allele) do not support the findings of Henry and associates (16). None of our measures of disease severity showed the A allele to be associated with less severe pulmonary disease. In genetic association studies such as this, population stratification based on ethnicity can be a confounding factor. However, as expected the vast majority of our study sample was Caucasian, and therefore it is unlikely that the lack of association in this study represents a type 2 error due to stratification. In a multicenter study such as this there may have been differences in ethnic diversity between centers. To address this, and other potential confounders, we created a categorical variable called CENTER. However, center was not a significant predictor of lung function and center to center variability was small (less than 1% of the total variability, which was less than the variability within centers). This result suggests that there were no large ethnic (or other) differences between centers that could have affected the association of genotype with measures of lung function. Other possible confounders in our study include social class and environment (i.e., smoking / passive smoking, increased exposure to air pollutants and infectious agents) and differences in center care. Center care is not likely a confounder in Canada as the care received by CF patients across Canadian CF clinics is standardized and therefore it comes as no surprise that the variable Center was not a significant predictor of lung function. Interestingly, infusions of α1-AT have been shown to reduce NE to undetectable levels in CF BALF (25). We showed CF patients with severe pulmonary disease (i.e., %predFEV1<50%) had a blunted acute phase rise in α1-AT (Figure 3.3). Also noted is a much lower percent change in α-AT levels from peak to stable in CF patients with severe disease. Possible reasons for this difference are the higher baseline values in the severe group and also the possibility that milder exacerbations in the severe group led to hospitalization or initiation of home IV therapy compared to the mild/moderate group. CF 102  patients with severe disease also had poor nutritional status, which has been shown in nonCF cohorts to affect α1-AT levels. The mean 20% increase in α1-AT during an acute infection reflects a relatively small acute phase response.  Voulgari and associates reported a 78% increase in α1-AT levels in  response to bacterial infection in a non-CF population (26). Kueppers measured α1-AT levels in response to a typhoid vaccine injection in otherwise healthy males who were homozygous wild type and heterozygous for the α1-AT deficiency alleles Z and S (27). While there was a lower baseline α1-AT levels in heterozygotes, a similar percent rise in levels was seen across the groups. In the study by Kueppers (27), α1-AT levels were monitored over 15 days; both the homozygous wild type and heterozygotes for the deficiency alleles showed a gradual return to normal values unlike the pattern seen in our group who had severe pulmonary disease. It may be that the high baseline level and blunted acute phase increase in α1-AT levels in our study is related to the chronic pulmonary bacterial infection in CF. Alternatively, poor dietary intake and malabsorption leading to malnutrition and cachexia could play a role. Morlese and associates measured the acute α1AT phase response to infection in nine 10-year old children who were also diagnosed with severe malnutrition (28). These children showed a blunted acute phase increase in α1-AT to bacterial infection. CF patients who showed heterozygosity for the S and Z polymorphisms also showed BMI values indicative of malnutrition as well as a significantly lower percent increase in α1-AT levels during pulmonary exacerbation. In our analyses BMI was an important predictor of α1-AT levels even when the CF patients who were heterozygous for the S and Z polymorphisms were selected out, suggesting that BMI irrespective of α1-AT genotype is a predictor of percent change in α1-AT levels during pulmonary exacerbation. While we did not show significant differences for BMI when our cohort was grouped by pulmonary disease severity, the mean BMI for the severe group was below 20 kg/m2 (Table 3.3) indicative of poor nutritional status and malnutrition. All the Canadian clinics promote patients to maintain normal body weight (i.e. adult BMI values between 20-25 kg/m2 and preferably between 22-25 kg/m2) in order to allow for weight loss that usually occurs during periods of pulmonary exacerbation or other CF-related illnesses. It is likely that a combination of poor nutritional status and chronic pulmonary bacterial 103  infection, which is common in CF patients with severe pulmonary disease, contributes to the pattern shown in stable and acute rise in α1-AT levels observed. An exogenous source of α1AT may be of potential benefit in malnourished CF patients with more severe pulmonary disease during pulmonary exacerbation, while measures are taken to normalize their body weight. In this study we used the common lung function parameter %predFEV1 as a measure of disease severity. This parameter is universally used as a measure of pulmonary disease severity and also correlates with measures of nutritional status (i.e., BMI in adults and percent of ideal weight in children) and pancreatic sufficiency status. Clinical scoring using Schwachman-Kulczycki (S-K) and Brasfield chest radiographic scores are also used, but have not been commonly used in large-scale studies. A benefit of using S-K scores is that this score takes into account pulmonary, nutritional, chest radiographic status and a measure of activity or mobility. Exercise capacity (29) (30), quality of life (31-33) and sputum volume (33), are alternative measures of pulmonary disease but these measures are not commonly utilized in CF patients in clinical assessment. As our study showed that nonpulmonary measures (i.e., BMI, PSS) affect pulmonary disease severity, this suggests that a measure for multi-system CF disease severity such as BMI and PSS is essential.  104  3.4 CONCLUSIONS The results of our study show that α1-AT genotype is not a major contributor to the variability of pulmonary disease severity in CF. Specifically α1-AT genotype did not correlate with %predFEV1, pulmonary infections and death or lung transplantation. Our study shows, however, that the levels of α1-AT during pulmonary infections may be affected by poor nutritional status independent of α1-AT genotype.  105  Table 3.1. Clinical characteristics of study subjects stratified by α1-AT S and Z genotypes. Values are shown as mean (SEM)*. Total MM MS, SS or MZ p-value Sex (M/F) 378/338 317/299 61/39 0.08 Age (years) 22.8(0.4) 23.0(0.5) 21.8(1.1) 0.31 CFTR genotype 391/295/72 342/254/59 49/41/13 0.44 (∆F508/∆F508; ∆F508/other; other/other) Age of CF diagnosis (years) 4.7(0.3) 4.6(0.5) 5.5(0.9) 0.32 %predFEV1 64.8(0.9) 64.6(1.0) 65.7(2.7) 0.68 BMI (kg/m2) 20.3(0.2) 20.4(0.2) 19.7(0.3) 0.22 Pancreatic sufficiency status 119/648 101/561 18/87 0.35 (sufficient/insufficient) # of P. aeruginosa positive/not 478/85 421/67 57/18 0.04 colonized Age of 1st infection with P. 11.1(0.5) 11.4(0.5) 9.1(1.1) 0.10 aeruginosa Age of chronic P. aeruginosa 14.4(0.8) 14.9(0.9) 11.3(1.5) 0.16 infection Frequency of IV Treatment/year 0.9(0.1) 0.9(0.1) 0.7(0.2) 0.22 Days of IV Treatment/year 13.7(1.2) 14.1(1.3) 11.5(3.1) 0.41 Dead or lung transplanted / alive 63/644 54/555 9/89 0.85 *The study sample varied between 713-716 for univariate analyses. For analysis of age of first infection with P. aeruginosa and P. aeruginosa status the study sample size was 461 and 555, respectively. For analysis of age of chronic infection with P. aeruginosa the study sample size was 159.  106  Table 3.2. Clinical characteristics of study subjects stratified by α1-AT 3’ G1237→A genotype. Values are shown as mean (SEM)*. Total  G1237→G  G1237→A A1237→A 48/54 23.5(1.2) 48/52/12  or p-value  Sex (M/F) 379/337 331/283 Age (years) 22.8(0.4) 22.7(0.4) CFTR genotype 391/296/72 343/244/60 (∆F508/∆F508; ∆F508/other; other/other) Age of CF diagnosis (years) 4.7(0.3) 4.6(0.3) 5.4(0.9) %predFEV1 64.8(0.9) 64.9(1.0) 63.7(2.3) 2 BMI (kg/m ) 20.3(0.2) 20.4(0.2) 20.0(0.4) Pancreatic sufficiency status 119/649 95/561 24/88 (sufficient/insufficient) # of P. aeruginosa positive / not 478/85 416/70 62/15 colonized Age of 1st infection with P. 11.1(0.5) 11.3(0.5) 10.0(1.1) aeruginosa Age of chronic P. Aeruginosa 14.4(0.8) 14.5(0.9) 13.6(1.6) infection Frequency of IV Treatment/year 0.9(0.1) 0.9(0.1) 0.9(0.2) Days of IV Treatment/year 13.7(1.2) 13.5(1.6) 15.4(4.5) Dead or lung transplanted / 63/644 53/551 10/92 alive *The study sample varied between 713-716 for univariate analyses. For analysis  0.20 0.49 0.14 0.38 0.64 0.49 0.05 0.23 0.36 0.71 0.86 0.60 0.71 of age of  first infection with P. aeruginosa and P. aeruginosa status the study sample size was 461 and 555, respectively. For analysis of age of chronic infection with P. aeruginosa the study sample was 159.  107  Table 3.3. Characteristics of the sub-group of patients in the acute phase α1-AT level study.  Gender (male/female) Age (yrs) CF diagnosis age (yrs) Infection with P. aeruginosa*** Age of chronic infection with P. aeruginosa BMI %predFEV1 (%) %predFVC (%) S-K score Brasfield score # of hospitalization days Stable α1AT (g/L) Acute α1AT levels (g/L) % change in α1AT *- Chi-square analysis.  Total 24/15 27.5 (1.1)** 3.1 (1.0) 2/34/3 14.6 (1.4)  Mild/Moderate 7/7 28.3 (2.2) 2.6 (1.4) 0/12/2 17.8 (2.8)  Severe 17/8 27.1 (1.3) 3.4 (1.4) 2/22/1 12.8 (1.5)  P-value 0.22* 0.60 0.73 0.43* 0.09  20.5 (0.5) 43.7 (3.5) 64.0 (3.8) 59.5 (3.0) 13.7 (0.7) 38.7 (5.1) 1.84 (0.06) 2.18 (0.10) 20.5 (5.4)  21.5 (0.7) 67.8 (8.0) 88.3 (4.2) 75.0 (2.7) 16.9 (0.6) 26.8 (3.5) 1.73 (0.07) 2.38 (0.23) 36.7 (10.4)  19.9 (0.7) 30.2 (2.1) 50.5 (2.9) 50.8 (3.3) 11.9 (0.8) 45.4 (7.4) 1.90 (0.09) 2.07 (0.09) 11.4 (5.4)  0.11 0.0001 0.0001 0.0001 0.0001 0.08 0.19 0.13 0.02  **- Values are mean (SEM). ***- Values are number (#) of subjects not infected with P. aeruginosa / # of subjects chronically infected with P. aeruginosa / # of subjects infected with P. aeruginosa but not chronically. One of the patients in the severe group not infected with P. aeruginosa was chronically infected with Burkholderia cepacia complex.  108  Figure 3.1. Comparison of pulmonary disease severity and Z and S alleles of the α1-AT gene. Values are presented as mean+SEM.  76 74 %predFEV 1 (%)  72 70 68 66 64 62  64.5  65.6  65.6  66.5  MZ (N=18)  MS (N=69)  SS (N=13)  60 58 MM (N=616)  109  Figure 3.2.Comparison of pulmonary disease severity and the 3’ G1237→A mutation of the α1-AT gene. Values are presented as mean+SEM.  68  %predFEV1 (%)  66 64 62 60  64.9  64.1  58 58.7  56 54 GG (N=614)  GA (N=95)  110  AA (N=7)  Figure 3.3. Alpha-1-antitrypsin levels during a pulmonary exacerbation and postexacerbation levels during stable clinical status. Levels are shown during the intravenous antibiotic intervention period (14 day) and at a post-exacerbation (stable) time point by pulmonary disease severity. Dotted lines show the range of normal levels for α1-AT.  3 Mild/Moderate (N-14)  Severe (N=25)  2.5  a1AT (g/L)  2  1.5  Normal range: 0.95-1.77 g/L 1  0.5 Acute  4  7  10  Therapeutic intervention interval (days)  111  14  Stable  3.5 BIBLIOGRAPHY 1.  Cantin, A. M., S. Lafrenaye, and R. O. Begin. 1991. Antineutrophil elastase activity in cystic fibrosis serum. Pediatric Pulmonology 11(3):249-53.  2.  Birrer, P., N. G. McElvaney, A. Rudeberg, C. W. Sommer, S. Liechti-Gallati, R. Kraemer, R. Hubbard, and R. G. Crystal. 1994. Protease-antiprotease imbalance in the lungs of children with cystic fibrosis. American Journal of Respiratory & Critical Care Medicine 150(1):207-13.  3.  O'Connor, C. M., K. Gaffney, J. Keane, A. Southey, N. Byrne, S. O'Mahoney, and M. X. Fitzgerald. 1993. alpha 1-Proteinase inhibitor, elastase activity, and lung disease severity in cystic fibrosis. Am Rev Respir Dis 148(6 Pt 1):1665-70.  4.  Goldstein, W., and G. Doring. 1986. Lysosomal enzymes from polymorphonuclear leukocytes and proteinase inhibitors in patients with cystic fibrosis. American Review of Respiratory Disease 134(1):49-56.  5.  Plotkowski, M. C., G. Beck, J. M. Tournier, M. Bernardo-Filho, E. A. Marques, and E. Puchelle. 1989. Adherence of Pseudomonas aeruginosa to respiratory epithelium and the effect of leucocyte elastase. J Med Microbiol 30(4):285-93.  6.  Plotkowski, M. C., A. M. Saliba, S. H. Pereira, M. P. Cervante, and O. BajoletLaudinat. 1994. Pseudomonas aeruginosa selective adherence to and entry into human endothelial cells. Infect Immun 62(12):5456-63.  7.  Hilman, B. C. 1997. Genetic and immunologic aspects of cystic fibrosis. Ann Allergy Asthma Immunol 79(5):379-90; quiz 391-4.  8.  Brown, M. A., W. J. Morgan, P. R. Finley, and P. Scuderi. 1991. Circulating levels of tumor necrosis factor and interleukin-1 in cystic fibrosis. Pediatric Pulmonology 10(2):86-91.  9.  Suter, S. 1986. Imbalance between polymorphonuclear leukocyte proteases and antiproteases in chronic pyogenic infections and its relation to the proteolysis of complement component C3. Complement 3(1):1-24.  10.  Santis, G., L. Osborne, R. Knight, M. E. Hodson, and M. Ramsay. 1990. Genetic influences on pulmonary severity in cystic fibrosis. Lancet 335(8684):294.  11.  Santis, G., L. Osborne, R. Knight, M. Ramsay, R. Williamson, and M. Hodson. 1990. Cystic fibrosis haplotype association and the delta F508 mutation in adult British CF patients. Hum Genet 85(4):424-5. 112  12.  Morgan, K., G. Scobie, P. Marsters, and N. A. Kalsheker. 1997. Mutation in an alpha1-antitrypsin enhancer results in an interleukin-6 deficient acute-phase response due to loss of cooperativity between transcription factors. Biochim Biophys Acta 1362(1):67-76.  13.  Doring, G., H. Krogh-Johansen, S. Weidinger, and N. Hoiby. 1994. Allotypes of alpha 1-antitrypsin in patients with cystic fibrosis, homozygous and heterozygous for deltaF508. Pediatric Pulmonology 18(1):3-7.  14.  Mahadeva, R., S. Stewart, D. Bilton, and D. A. Lomas. 1998. Alpha-1 antitrypsin deficiency alleles and severe cystic fibrosis lung disease. Thorax 53(12):1022-4.  15.  Mahadeva, R., R. C. Westerbeek, D. J. Perry, J. U. Lovegrove, D. B. Whitehouse, N. R. Carroll, R. I. Ross-Russell, A. K. Webb, D. Bilton, and D. A. Lomas. 1998. Alpha1-antitrypsin deficiency alleles and the Taq-I G-->A allele in cystic fibrosis lung disease. Eur Respir J 11(4):873-9.  16.  Henry, M. T., S. Cave, J. Rendall, C. M. O'Connor, K. Morgan, M. X. FitzGerald, and N. Kalsheker. 2001. An alpha(1)-antitrypsin enhancer polymorphism is a genetic modifier of pulmonary outcome in cystic fibrosis. Eur J Hum Genet 9(4):273-8.  17.  Crystal, R. G. 1989. The alpha 1-antitrypsin gene and its deficiency states. Trends Genet 5(12):411-7.  18.  Cox, D. W., S. L. Woo, and T. Mansfield. 1985. DNA restriction fragments associated with alpha 1-antitrypsin indicate a single origin for deficiency allele PI Z. Nature 316(6023):79-81.  19.  Curiel, D. T., A. Chytil, M. Courtney, and R. G. Crystal. 1989. Serum alpha 1antitrypsin deficiency associated with the common S-type (Glu264----Val) mutation results from intracellular degradation of alpha 1-antitrypsin prior to secretion. J Biol Chem 264(18):10477-86.  20.  Carrell, R. W. 1986. alpha 1-Antitrypsin: molecular pathology, leukocytes, and tissue damage. J Clin Invest 78(6):1427-31.  21.  Carrell, R. W., and D. A. Lomas. 2002. Alpha1-antitrypsin deficiency--a model for conformational diseases. N Engl J Med 346(1):45-53.  22.  Kalsheker, N. A., G. L. Watkins, S. Hill, K. Morgan, R. A. Stockley, and R. B. Fick. 1990. Independent mutations in the flanking sequence of the alpha-1- antitrypsin gene are associated with chronic obstructive airways disease. Dis Markers 8(3):1517. 113  23.  Sandford, A. J., T. Chagani, J. J. Spinelli, and P. D. Pare. 1999. alpha1-antitrypsin genotypes and the acute-phase response to open heart surgery. Am J Respir Crit Care Med 159(5 Pt 1):1624-8.  24.  Seersholm, N., J. T. Wilcke, A. Kok-Jensen, and A. Dirksen. 2000. Risk of hospital admission for obstructive pulmonary disease in alpha(1)-antitrypsin heterozygotes of phenotype PiMZ. Am J Respir Crit Care Med 161(1):81-4.  25.  McElvaney, N. G., R. C. Hubbard, P. Birrer, M. S. Chernick, D. B. Caplan, M. M. Frank, and R. G. Crystal. 1991. Aerosol alpha 1-antitrypsin treatment for cystic fibrosis. Lancet 337(8738):392-4.  26.  Voulgari, F., P. Cummins, T. I. Gardecki, N. J. Beeching, P. C. Stone, and J. Stuart. 1982. Serum levels of acute phase and cardiac proteins after myocardial infarction, surgery, and infection. Br Heart J 48(4):352-6.  27.  Kueppers, F. 1968. Genetically determined differences in the response of alphaantitrypsin levels in human serum to typhoid vaccine. Humangenetik 6(3):207-14.  28.  Morlese, J. F., T. Forrester, and F. Jahoor. 1998. Acute-phase protein response to infection in severe malnutrition. Am J Physiol 275(1 Pt 1):E112-7.  29.  Frangolias, D. D., C. L. Holloway, S. Vedal, and P. G. Wilcox. 2003. Role of exercise and lung function in predicting work status in cystic fibrosis. Am J Respir Crit Care Med 167(2):150-7.  30.  Frangolias, D. D., and P. G. Wilcox. 2001. Predictability of oxygen desaturation during sleep in patients with cystic fibrosis : clinical, spirometric, and exercise parameters. Chest 119(2):434-41.  31.  Quittner, A. L., and A. Buu. 2002. Effects of tobramycin solution for inhalation on global ratings of quality of life in patients with cystic fibrosis and Pseudomonas aeruginosa infection. Pediatr Pulmonol 33(4):269-76.  32.  Goldbeck, L., and T. G. Schmitz. 2001. Comparison of three generic questionnaires measuring quality of life in adolescents and adults with cystic fibrosis: the 36-item short form health survey, the quality of life profile for chronic diseases, and the questions on life satisfaction. Qual Life Res 10(1):23-36.  33.  Bradley, J., O. McAlister, and S. Elborn. 2001. Pulmonary function, inflammation, exercise capacity and quality of life in cystic fibrosis. Eur Respir J 17(4):712-5.  114  CHAPTER 4: INNATE IMMUNITY GENES AS POTENTIAL MODIFIER LOCI IN CYSTIC FIBROSIS  A version of this chapter may be published in the future.  115  4.0 INTRODUCTION Polymorphisms in genes involved in innate immunity were investigated. Polymorphisms in the coding region of the mannose binding lectin (MBL2) gene have been shown to affect the levels of the protein produced and polymorphisms in the promoter region have been shown to affect regulation. Numerous polymorphisms have been identified in the pulmonary surfactant genes A1, A2 and D. It is not known at this time whether the polymorphisms identified affect protein function and levels. In this study we investigated the association of amino acid changing polymorphisms in the pulmonary surfactant genes and MBL2 gene polymorphisms on pulmonary disease severity and progression, survival and succeptibility to infection with common CF pathogens. 4.1 RATIONALE FOR THE INVESTIGATION OF INNATE IMMUNITY GENES AS POTENTIAL MODIFIERS IN CF Chronic colonization with Pseudomonas aeruginosa is a common phenotypic characteristic in the CF population and an earlier age of chronic infection is associated with more rapid pulmonary disease progression (1-4). Commonly, repeated infections with Staphylococcus aureus and Haemophilus influenza describe the clinical course prior to colonization with P. aeruginosa. A less common pathogen in the CF population is Burkholderia cepacia complex (BCC), and colonization with this pathogen is associated with a more severe clinical outcome. First line innate defenses against pathogens invading the airways are therefore important in preventing colonization and possibly modulating the inflammatory response. Mannose binding lectin (MBL2) and pulmonary surfactant A (SPA) proteins have been shown to interact with Staphylococcus aureus, a pathogen known to cause recurrent early infections in the CF population and also recurrent infections and chronic colonization following chronic colonization with the pathogen P. aeruginosa in this population. In addition to binding to Staphylococcus aureus (5), MBL2 has been shown to bind to P. aeruginosa and BCC (6). Davies and associates, using BCC isolates obtained from chronically infected CF patients, showed that MBL2 binds to BCC and activates complement, as measured by C4 deposition (6). SPA and SPD bind to pathogens, which are known to be responsible for respiratory infections and include common CF respiratory pathogens Staphylococcus aureus, P. aeruginosa and Haemophilus influenza (7). Although 116  it has been suggested that direct binding of SPA and SPD to the pathogen is not always necessary to trigger microbicidal activity, it is believed that the presence of these proteins may somehow mediate host defense (8-11). Studies have shown that SPA (12) and SPD (13) proteins enhance the alveolar macrophage phagocytosis of P. aeruginosa. In this study we limited our investigation of innate immunity genes to polymorphisms in 4 genes: MBL2, SPA-1, SPA-2 and SPD. Other potential modifiers in the innate immunity pathway which could potentially affect pulmonary disease severity, progression and susceptibility to infection with CF pathogens are numerous. Genetic variations that disrupt innate immune sensing of infectious organisms can affect the ability of the host to respond to infection. Therefore variants in proteins which bind to pathogen, receptors on phagocytic cells, proteins in the complement cascade, signaling molecules are all potential candidate modifier genes for CF. Some potential candidates where variants have been identified include: inflammatory mediators IL-1beta, IL-1Ra, IL-8, IL-10, plasma complement proteins of the classical, lectin and alternate pathway of complement activation, constituents of host cells (alpha defensins and acyoxyacyl hydrolase of neutrophils, beta defensins of epithelial cells), innate immune receptors (CD14, toll-like receptors 2, 4 and 6), serine proteases (Mannose binding lectin associated serine protease, C1q, factor P). 4.1.1 Tissue distribution of innate immunity proteins MBL2, SPA and SPD proteins are regulated hormonally and developmentally and are influenced by inflammation. MBL2 is characterized as an acute phase protein and levels of the protein are two thirds of adult levels at birth and increase to adult values within a few weeks following birth (14). SPA and SPD are found in amniotic fluid as early as 16-20 weeks of gestation and increase during pregnancy. Levels of SPA and SPD can be increased prematurely artificially with glucocorticoid treatment (15). MBL2 and its associated serine proteases (MASP-1, 2 and 3 and MAp19) are synthesized by hepatocytes and secreted into the plasma. Alveolar type II cells and unciliated bronchial epithelial cells synthesize SPA and SPD proteins and they are secreted onto the airway surface lining fluid (ASLF) (16). Lin and associates have shown that the SPA genes are also expressed in the small and large intestine (17). SPD is also synthesized and secreted by epithelial cells of the gastrointestinal mucosa and exocrine glands (18-20). 117  4.1.2 Characteristics of innate immunity proteins The human MBL2 and pulmonary surfactant A and D genes are located on the long arm of chromosome 10 with MBL2 mapped to 10q11.2-q21 and SPA and SPD mapped to region 10q22-q23. The order of the genes is (21) (Table 4.1): Centromere _ MBL2 _ SPA-2 _ SPA pseudogene _ SPA-1 _ SPD _ Telomere There are two functional SPA genes and a SPA pseudogene (22). SPA-1 and SPA-2 genes have similar DNA sequences and the organization of each SPA gene consists of four coding exons which span a region less than 5 kb (23, 24). Sequence analysis of the three SPA genes genomic sequences by Hoover and Floros suggests that an ancestral SPA sequence was duplicated giving rise to SPA-1 and SPA-2. A subsequent duplication of SPA-2 produced the SPA pseudogene (21). The orientations of the SPA-1 and SPA-2 genes are ‘head to head’ in opposite transcriptional orientations and may share cis-acting regulatory elements (21). Of the genes listed above, the MBL2 locus is physically the farthest away from SPA and SPD and is considered more distant in evolutionary terms (25). MBL2 is not in linkage disequilibrium (LD) with any of the surfactant genes (21). SPD is closer from an evolutionary perspective and in terms of physical distance to the SPA locus than MBL2 (21, 25). SPA-1 and SPA-2 are in strong LD (21). 4.1.3 Structure and function of innate immunity proteins MBL2, SPA and the SPD proteins bind to carbohydrates commonly found on bacterial and viral surfaces but not found in the host. MBL2, SPA-1, SPA2 and SPD proteins share a similar primary structure. They are composed of four structural domains; polypeptide subunits with N terminal segments which interconnect and form the center of the molecule, a collagen-like domain, an acidic hydrophobic neck domain, and a C-type lectin carbohydrate recognition domain that binds carbohydrates in a calcium dependent manner (26, 27). MBL2, SPA-1 and SPA-2 proteins form a three dimensional structure that resembles a ‘flower or tulip bouquet’, which is similar to the structure of the first component of the classical complement pathway molecule C1q (28, 29). SPD is similar in structure and amino acid composition to MBL2, SPA-1 and SPA-2 proteins, but assembles into an X-like formation. 118  The MBL2 protein is made up of six 96 kDa subunits and each subunit consists of three identical 32 kDa polypeptide chains. MBL2 is shown to opsonize numerous pathogens, which are summarized in Table 4.2. A comprehensive study by Neth and associates investigated MBL2 binding and complement activation of multiple isolates of clinically obtained pathogens (30). They showed that MBL2 binds to specific pathogens and promotes C4 deposition (30). Davis and associates investigated two pathogens commonly infecting CF patients (mucoid and non-mucoid isolates of P. aeruginosa and of BCC obtained from CF patients) and showed that significant amount of MBL2 binds to BCC and that lesser amounts bind to mucoid strains of P. aeruginosa. MBL2 binding to BCC specimens resulted in complement activation as measured by C4 deposition (6). Younger and co-workers also showed intermediate binding of MBL2 to P. aeruginosa, and they found that the alternative pathway accounted for the majority of opsonized P. aeruginosa (31). Younger and coworkers showed that opsonization of P. aeruginosa (strain UI-18) in a murine model was preserved, even when the function of C1q and MBL2 were inhibited, suggesting that the alternative pathway was important in initiating opsonization of this pathogen (31). The MBL2 gene contains four exons and each encodes one of the four regions of the protein. Specifically, exon 1 encodes the N-terminal segment, exon 2 encodes the collagen-like domain, exon 3 encodes the neck region, and exon 4 encodes the carbohydrate recognition domain. Three polymorphisms have been identified in exon 1 that encode a region of the protein that interacts with collectin receptors and the MBL2-associated serine proteases (MASP). These three polymorphisms encode different structural variants of the MBL2 protein (32-35). The polymorphisms result in three different amino acid substitutions (codon 52 (ArgˆCys, D allele), codon 54 (GlyˆAsp, B allele) and 57 (GlyˆGlu, C allele)) that prevent the correct assembly of the MBL2 protein subunits into a trimeric structure making the molecule unstable and decreasing the binding capacity to ligands (36). Interaction between the variant MBL2 and the MASPs still occurs, however, there is lack of complement activation (as measured by C4 deposition) and it is believed that the reason for this is reduced capacity of the variant MBL2 to bind ligand (36). Additionally variant MBL2 is shown to be more susceptible to matrix metalloproteinase proteolysis (37). The wild type allele is denoted as A for all three polymorphisms. Two promoter polymorphisms have been identified at positions -550 (H/L genotype) and -221 (Y/X genotype) which both involve a 119  single nucleotide substitution of G to C (29). The promoter haplotypes HY, LY and LX (HX was not observed) lead to high, intermediate and low MBL2 levels, respectively (29, 38, 39). The LX promoter haplotype is only found with the wild-type coding allele for exon 1 (29). A single nucleotide substitution of G to C has also been documented at position +4 but this polymorphism has not been shown to have an affect on MBL2 levels (38). Haplotype reconstruction and LD analysis performed on a sample consisting of 69 Eskimos, 120 Caucasians and 61 native black Africans in the study by Madsen and associates (29) showed LD between promoter variants except in the black African cohort. They also showed strong LD between the promoter and structural polymorphisms. The X promoter allele was only found in haplotypes with the L promoter allele and only with the A allele for the three coding polymorphisms. The B, C and D coding alleles were only observed with the high expression promoter (i.e., HY) haplotype (29). MBL2 levels increase up to threefold after stimulation. The B, C and D coding alleles are associated with low plasma MBL2 protein concentrations (40-42). Intermediate plasma MBL2 protein concentrations are expressed in those who are heterozygous wild-type. Garred and co-workers (43) measured plasma levels of MBL2 protein in the CF population and showed absent or very low MBL2 levels in CF patients who were homozygous or heterozygous for the coding polymorphisms regardless of promoter haplotype. Low, intermediate and high MBL2 plasma levels were associated with XX, XY and YY promoter genotypes, respectively (in individuals who were wild-type for the exon 1 polymorphisms) (43). Unlike MBL2, SPA and SPD proteins do not associate with MASPs and do not activate complement. While SPD is believed to have strictly antimicrobial functions, SPA proteins are also required for the structure and / or stability of surfactant aggregates. SPA and SPD are synthesized and released following an inflammatory stimulus and their mode of action is: ƒ  Direct binding of SPA and SPD protein via their carbohydrate recognition domains to mannose carbohydrate on the pathogen’s surface (to aggregate and opsonize pathogens directly)(7, 8), or  ƒ  Modulation of macrophage function (potentially by increasing the activity of the macrophage mannose receptor and other receptors on immune cells)(7, 11). 120  It is believed that SPA and SPD proteins bind to a number of different immune cell receptors including the MBL2 macrophage mannose receptor (7). SPA exhibits a similar acute phase response as MBL2 protein after LPS aerosolization (44) and intratracheal instillation (45). SPA binds to the common CF pathogens, Staphylococcus aureus and P. aeruginosa. SPA functions as an opsonin and binds to Staphylococcus aureus and the C1q receptor of the phagocytes and stimulates phagocytosis by alveolar macrophages (8, 46, 47) Macrophage uptake of non-opsonized Staphylococcus aureus, as well as P. aeruginosa, have also been shown to occur in the presence of SPA (48). SPA protein binds to Gram negative bacteria by interacting with the pathogen’s LPS, specifically SPA has been shown to bind to LPS via a lipid A domain (49, 50). It is believed that SPA binding to Gram negative bacteria is dependent on the LPS structure (7, 48). In the case of the pathogen Mycobacterium tuberculosis, SPA acts as a ‘Trojan horse’ for the pathogen. SPA binds to Mycobacterium tuberculosis via its carbohydrate recognition domains and then binds to the macrophage mannose receptor stimulating its uptake (51), and the pathogen can multiply within the macrophage. In the European population, there are at least 153 and 110 polymorphisms in the SPA-1 and SPA-2 genes, respectively (Seattle SNPs, http://pga.gs.washington.edu/). Three (Ala19Val, Leu50Val, Arg219Trp) of the polymorphisms identified in SPA-1 and three (Asn9Thr, Ala91Pro, Gln223Lys) of the polymorphisms identified in SPA-2 are amino acid changing (52). Floros and associates investigated in a cohort of 239 the LD between SPA-1 and SPA2 and found to be strong (22). Many of the SPA-1 and SPA-2 haplotypes previously reported by Flores and associates (22) are rare and only four SPA-1 and SPA-2 gene haplotypes (i.e., 6A2-1A0, 6A3-1A1, 6A3-1A0 and 6A3-1A2) are observed in Caucasians (53). Table 4.3 shows the common single nucleotide polymorphisms that make up the SPA-1 and SPA-2 haplotypes as reported by DiAngelo and associates (52). SPA and SPD bind to a broad range of pathogens with some overlap (Table 4.2) but differ in their mode of interaction with the pathogen (7). For example in the case of Gram-negative bacteria, Van Iwaarden and associates (50) showed that SPA binds to the lipid moiety of LPS, whereas SPD interacts with the core oligosaccharides of the Gram-negative bacteria (7). One receptor that SPD has been proposed to bind to is GP-340 found on alveolar macrophages, although other receptors have been purported to exist, but have not been 121  identified at this time. It is also hypothesized that SPD binds to pathogen and somehow mediates modifications in the presentation of pathogen to host defense cells (54). SPD has been shown to bind to pathogens which infect CF patients: P. aeruginosa (48) and Haemophilus influenza (9, 55, 56). Three polymorphisms in the coding region of the SPD gene have been described; Thr11Met and Thr160Ala (52) and Thr270Ser (52, 57). The clinical implications of SPA-1, SPA-2 and SPD polymorphisms are unclear and there is no consensus on the normal range for these three proteins, or correlations between concentrations of these proteins and genotypes (28). Lower levels of these proteins have been reported in serum and BALF in patients who have cystic fibrosis, adult respiratory distress syndrome, and chronic smokers and in patients characterized as having an increased risk of pneumonia (58, 59). Lower levels of these proteins are used as clinical indicators for lower respiratory tract infections, type II cell hyperplasia and acute lung injury in respiratory diseases (60-64). Mikerov and associates(65) have shown differences in the abilities of SPA1 and SPA-2 proteins to increase phagocytosis of P. aeruginosa by alveolar macrophages and that SPA-2 variants tested (i.e., 1A0 and 1A) were more active than SPA-1 variants (6A4 and 6A2). Increased susceptibility to infection and mortality have been shown in SPA and SPD knock-out mice (66-68) and associations shown between respiratory diseases and known polymorphisms (69-73). 4.1.4 Complement activation pathways The complement system comprises a series of enzymatic and non-enzymatic proteins, which are required for the operation of the innate immune system. The complement pathway is activated in three ways with the latter two involved in first line defense against pathogen infection: ƒ  The classical pathway  ƒ  The alternate pathway  ƒ  The lectin pathway  In the 1940s, researchers proposed that an antibody-independent pathway, the alternative pathway, existed and that complement could be activated by bacterial surfaces. It was later shown that there are pattern recognition receptors expressed by the host that allow the host to recognize pathogen-associated molecular patterns (PAMPs) and thus distinguish pathogens from self antigens (74, 75). The lectin pathway is a recent discovery dated to 122  1987 (76, 77) and a better understanding of this pathway came when the MASPs were identified (14, 78). The lectin and alternative pathways provide a non-adaptive first line of defense against microbial infection since complement activation does not require specific antigen to bind to pathogen surfaces as required in the classical pathway. In addition, the response to pathogen invasion is an immediate activation compared with the classical pathway where there is a five to seven day delay before the production of the required antibody. The three pathways merge at the activation step of C3. Figure 4.1 summarizes the alternate and lectin pathways of complement activation and compares them to the classical pathway of complement activation. The classical pathway is triggered by antibody binding to antigen on the pathogen surface and is part of the adaptive humoral immune response. The binding of immunoglobulin M and G (IgM and IgG) antibodies to the pathogen surface and to C1q activates the complement cascade. The elimination of pathogens is achieved through a series of reactions, which involve plasma proteins that interact with the bound antibodies and are recognized as complement receptors by phagocytes that are stimulated to engulf the pathogen. The alternative pathway is initiated by structures on the pathogen surfaces and in this pathway SPA and SPD proteins coat the surface of pathogens (opsonization) facilitating the uptake by macrophages and neutrophils. The alternative pathway of complement activation does not require the presence of specific antibody. C3 freely circulates in the plasma and can undergo spontaneous cleavage in plasma to generate fragments C3b and C3a. C3 contains an intramolecular thioester bond, which, once exposed to the molecular surface of invading pathogens, forms a covalent bond with it and facilitates the phagocytosis of pathogens through the C3 receptors on phagocytic cells. C3b only binds to pathogen; if it does not bind to pathogen, it becomes inactivated by cell surface proteins on host cells. Pathogen cells lack these proteins and C3b covalently binds to the pathogen surface and also binds to additional proteins (i.e., factor B which is cleaved by factor D forming the C3b,Bb complex which is structurally and functionally homologous to C4b,2b of the classical pathway). C3b,Bb is the C3 convertase of the alternative pathway; the complex is further stabilized on the pathogen surface by binding to factor P. The stabilized C3b,Bb then acts in the same way as the C3 convertase of the classical pathway 123  converting circulating C3 molecules to C3b that coat the pathogen surface and C3a molecules which mediate local inflammation. Some of the C3b molecules bind to existing C3b,Bb complex and form C3b2,Bb. This is the alternative pathway of C5 convertase activation leading to the activation of the terminal complement components through binding and cleavage of C5 by C3b molecules. The C5b that is generated initiates the lytic pathway and C5a is a potent inflammatory mediator. The C-lectin pathway is also initiated in an antibody-independent manner. MBL2 (and the ficolins) bind to carbohydrates (mannose and N-acetylglucosamine) on the surface of pathogens. MBL2 is structurally similar to complement C1 subcomponent, C1q, and the mode of activation for MBL2 is through associated serine proteases called MBL2-associated serine proteases (MASP). MBL2 associates through its collagen-like domain with serine proteases called MASP-1 and MASP-2 (41) and MASP-3 (79). Current research suggests that MASP-2 is the main enzyme responsible for activation of the lectin pathway (80). The MASP proteases are similar in structure, function and activation to C1r and C1s of the classical pathway. Specifically, binding of MBL2 to mannose containing proteins or carbohydrate groups on the surface of pathogens results in conformational changes to MASP-2; MASP-2 activates C4 and C2 and the C3 convertase C4b2a complexes are generated (81). MBL2 is also associated with a truncated form of MASP-2 called small MBL2-associated protein (sMAP). The MBL2-MASP-sMAP complex circulates in the blood and once MBL2 binds to a pathogen, the MASPs are activated acquiring proteolytic activities. Circulating in the complexed form MASPs are an inactive proenzyme (82, 83). The activated lectin pathway ultimately generates C3b and C3a. The killing of the pathogen via the MBL2 lectin pathway is mediated by the host-mediated terminal lytic complex and by clearance after promoting phagocytosis by macrophages. 4.1.5 Review of the literature for MBL2, SPA and SPD: Clinical correlates The deficiency polymorphisms in the MBL2 gene have been investigated as potential risk factors for infections. The coding polymorphisms in exon 1 result in mutant MBL2 molecules, which do not participate in opsonization or complement activation (38). The B polymorphism is unable to bind MASP (82) and both the B and C polymorphisms are unable to activate complement (27, 84). MBL2 deficiency has been associated with increased risk of ear infections (85), respiratory infections (86-88) and other infections (87, 89) in non-CF 124  children . Kielgast and associates showed a potential association between viral infections and low umbilical cord serum MBL2 levels (90). Worse prognosis from viral and bacterial infections has been shown in immuno-compromised adults with low serum MBL2 levels (89, 91-93). Garred and co-workers showed an increased frequency of hospitalizations for infections in children with MBL2 deficient genotypes (91). The MBL2 D allele at codon 52 polymorphism is considered a potential contributor to chronic necrotizing pulmonary aspergillosis (94). Yang and co-workers have shown an association for increased hospitalizations for pulmonary infections in COPD patients with the MBL2 B allele at codon 54 (95). Neth and co-workers showed that children with febrile neutropenia who were undergoing chemotherapy for cancer had on average longer hospitalizations if they had an MBL2 deficient haplotype (5). Peterslund and associates showed a higher frequency of bacteremia and pneumonia in a group of patients undergoing chemotherapy for cancer, who also had low MBL2 levels (96). MBL2 deficiency is common. It has been suggested that there is an evolutionary selection advantage to explain why these deficiency alleles are maintained in the population. In this respect, Garred and associates (97) have shown that MBL2 protein binds to Mycobacterium leprae and patients infected by Mycobacterium leprae are more likely to be MBL2 wild-type (based on MBL2 protein levels) compared with a control group of healthy blood donors residing in the same area. Similarly, an MBL2 haplotype associated with low MBL2 protein levels is protective against infections with Plasmodium falciparum (98), with Cryptosporidium parvum in AIDS patients (99), and Leishmania chagasi (100). Garred and colleagues concluded that during infection with intracellular pathogens the opsonic mechanisms of host defense are used by some pathogens to invade host cells (92, 97, 101). Polymorphisms in the MBL2 gene resulting in reduced levels of this protein have been investigated in multiple CF cohorts to date. In three of these studies, MBL deficiency was associated with more severe pulmonary disease, earlier colonization with Pseudomonas aeruginosa (43, 102) and increased susceptibility to BCC colonization(43, 103). Choi and associates(104) only showed an association for the MBL2 structural deficiency alleles and pulmonary disease severity in CF patients who were homozygous for the delta F508 CFTR mutation. While these studies suggest that MBL2 is a modifier gene in CF, the possibility of type one error due to small sample size (105) cannot be dismissed at present. Also of 125  relevance for the further investigation of these polymorphisms in the CF population are the findings of Davis and colleagues (6), who have shown high levels of MBL2 binding to BCC isolates (obtained from colonized CF patients) and observed that MBL2 protein binding to BCC resulted in complement activation. However, the authors in their recent epidemiological study state that MBL2 may not be important in BCC infection (106), however their study design is limited to investigating susceptibility to BCC infection and not whether infection with BCC affects long term pulmonary disease course. In this study (106)they showed a decline in pulmonary function and oxygen saturation, and a higher frequency of hospital admissions for pulmonary infections requiring IV therapy over the short interval they investigated (one year interval) in CF patients who were MBL2 homozygous for the deficiency alleles. They did not show differences in infection rates with P. aeruginosa and BCC and homozygosity for the MBL2 deficiency alleles (106). Carlsson and co-workers (107) investigated deficiency alleles for MBL2 and MASP-2 in a CF cohort. In their study, they did not show an association between MBL2 or MASP-2 deficiency genotypes and pulmonary disease severity, but did show worse pulmonary disease severity in a subgroup of CF patients (N=27) who were colonized with Staphylococcus aureus. Staphylococcus aureus is a pathogen to which MBL2 protein has been shown to binds to (Table 4.2). Choi and associates showed worse %predFEV1 in CF patients who were carriers of the MBL2 structural deficiency polymorphisms (104). In the study of Trevisiol and associates (102) of a small Italian CF cohort (N=47) showed lower mean %predFEV1 and earlier median age of chronic infection with P. aeruginosa in MBL2 deficient patients. The authors state that the association of worse pulmonary function in MBL2 deficient patients was only shown in those CF patients who were infected with P. aeruginosa were carriers of a severe CFTR genotype(102). Considering the sample size and distribution of CF patients for MBL2 genotype, P. aeruginosa infection status and CFTR genotype their conclusions are based on a very basic descriptive analysis of the data. MBL2 deficiency genotype (0/0, A/0, XA/0, XA/XA) was studied in a larger multicenter study and they did not show an MBL2 deficiency genotype to be associated with severe pulmonary disease even when chronic infection with P. aeruginosa was considered in their models (108). SP-A deficiency and SP-D deficiency are associated with increased risk of infection and death from infection in SPA and SPD knockout mice. LeVine and associates (67) infected mice lacking SPA, or SPD and wild-type mice with group B Streptococcus or H. Influenza 126  and showed deficient uptake of both pathogens by alveolar macrophages in the knock-out mice. There were differences in the bacterial killing, degree of inflammation and oxidant production between SPA knock-out mice and SPD knock-out mice. One study showed that SPD deficient mice developed chronic inflammation, emphysema and lung fibrosis (109). This study and others show that SPA and SPD have related but also potentially distinct functions in response to pathogen infection in the lung (67, 110-112). Sano and associates showed that SPA and SPD bind to LPS but differ in the mechanism by which they modulate the LPS-CD14 interaction (113). The SPD protein has been shown to enhance and suppress inflammatory mediator production and the outcome is dependent on the receptor SPD binds to and the binding orientation of the molecule (114). If the carbohydrate recognition domains of SPD protein are complexed with pathogen carbohydrates or lipid ligands then SPD functions to stimulate phagocytosis and proinflammatory responses, but if SPD is not complexed with ligands then it functions to block proinflammatory mediator production (57, 113). Leth-Larsen and colleagues hypothesize that the different molecular weight forms of SPD related to the coding polymorphisms are linked to the different functions associated with SPD (discussed below) (57). The SP-A loci have been investigated as a site for candidate genes for respiratory distress syndrome (RDS) (69, 115-118). Kala and associates showed a positive association for RDS and the SP-A loci (69). Haataja and co-workers identified an SPA-1-SPA-2 haplotype (6A2, 1A2) that was over-represented in RDS infants and showed that this association was dependent on the degree of prematurity and also homozygosity for threonine at the SPB (IIe131Thr) polymorphism (116). The SPA-1-SPA-2 haplotype 6A3-1A2 was shown to be protective for RDS (115, 116). Homozygosity for the T allele for the SPD (Thr11Met; C32T) gene was shown to be associated with increased susceptibility to severe respiratory syncytial virus infection (70). Choi and associates investigated SPA gene polymorphisms in a CF cohort and showed significantly worse %predFEV1 and S-K clinical scores in CF patients with the SPA-1 6A3 allele and SPA-2 1A1 allele, and this association was also present when investigating the SPA-1-SPA-2 haplotype (6A3/1A1)(104). SPA-1, SPA-2, and SPD gene polymorphisms were investigated as candidate modifier genes for COPD. Guo and associates showed an association between one silent SPA-1 (non amino acid changing) polymorphism at position 60 and increased risk for COPD (72). Lin and colleagues investigated all known polymorphisms in SP-A and SP-D as candidate genes variants in 127  acute respiratory distress syndrome (ARDS) and showed no significant associations between variant alleles and ARDS (119). Leth-Larsen and associates genotyped 206 Danes for the three coding polymorphisms in SPD and measured serum levels of SPD. They showed higher serum levels of SPD in those subjects who were TT versus CC and CT (mean value was 1035 ng/ml for TT versus 849.6 and 744.1 ng/ml for TC and CC) for the Thr11Met polymorphism and no significant differences in serum levels associated with the other two coding polymorphisms. There were two structurally different variant forms of the protein in serum from TT individuals (a high and low molecular weight protein). The low molecular weight form associated with the TT genotype was similar to the protein found in serum from CC individuals. The average distribution of high to low molecular weight SPD protein in serum was 1 to 1.6 for TT individuals and 1 to 5.1 for CC individuals. The high molecular weight variant was shown to preferentially bind to the Influenza A virus and Gram-positive and Gram-negative pathogens, while the low molecular weight variant bound to simpler ligands like LPS (57). The authors (57) also showed that the high molecular weight variants had differential: ƒ  Binding capacities to pathogens,  ƒ  Aggregation,  ƒ  Clearance of pathogen.  Similarly Mikerov and associates(65) have shown differences in the abilities of SPA-1 and SPA-2 proteins to increase phagocytosis of P. aeruginosa by alveolar macrophages and showed that SPA-2 variants tested (i.e., 1A0 and 1A) were more active than SPA-1 variants (6A4 and 6A2). The investigation of polymorphisms in SPA and SPD as potential candidate genes in pulmonary disease pathogenesis is still in its infancy and while variants in these genes have been described, the exact effects these variant alleles have on gene transcription and translation and protein function are still unclear. Variations in the levels of surfactant proteins in patient populations as well as data from mouse knock-out studies of these proteins, and association studies investigating surfactant polymorphisms have shown possible associations with pulmonary diseases. There will be the added difficulty when investigating the SPA gene polymorphisms as this protein is encoded by two genes with several allelic polymorphisms. Current findings show that the surfactant genes are potential targets for investigations into pulmonary disease pathogenesis. 128  4.1.6 Review of the literature for Burkholderia cepacia complex infection in CF In CF, impaired airway epithelial chloride transport leads to a milieu favoring colonization by bacteria, particularly Staphylococcus aureus and P. aeruginosa. Over two decades ago, another microbe, originally named Pseudomonas cepacia and now known as Burkholderia cepacia complex (BCC), came to the foreground as a colonizer of CF patients with a prevalence varying widely between clinics. In the literature this pathogen has been associated with an accelerated decline in clinical status and increased mortality in CF patients colonized with BCC versus P. aeruginosa (120-123), while others have noted a varied response with BCC colonization of CF patients with many reporting no difference in outcome (124-126). In our earlier case-control study, we showed increased mortality and pulmonary exacerbations in BCC infected CF patients, but no differences in pulmonary disease progression (change in %predFEV1 and FVC in pre- and post-acquisition interval). Lewin and associates specifically showed in their retrospective study a higher mortality in the first year post-acquisition in 124 CF patients infected with BCC compared with a similar number of patients colonized with P. aeruginosa but, of interest, not in the second year following acquisition of BCC (124). Taylor and colleagues (127)showed a more rapid deterioration and increased mortality associated with CF patients exhibiting severe pulmonary disease (%predFEV1<40%) at the time of BCC acquisition, while those patients infected with BCC exhibiting mild and moderate disease at the time of acquisition maintained a stable clinical status in that first year of monitoring. aeruginosa  Co-infection with  P.  and BCC has been shown to result in a more severe decline in pulmonary function (123,  128, 129).  A number of more recent studies have reported worse outcomes for CF patients who acquire BCC. Chaparro and associates (130) reviewed the Toronto experience for lung transplantation in CF patients chronically colonized with BCC GEN IIIA, RAPD-type 2 (GEN=genomovar and RAPD= random amplification of polymorphic DNA) and showed reduced 1, 2 and 3 year survival compared to CF patients who received a lung transplant and were chronically infected with P. aeruginosa. De Boeck and colleagues reported on the Belgian experience of poor 5-year survival in CF patients infected with either GEN IIIA or GEN II (131). Aris and associates showned higher mortality in post-lung transplant CF patients who were chronically infected with GEN III compared to CF patients in their study 129  who were chronically infected with either GEN II, P. aeruginosa or other CF respiratory pathogens (132). CF patients infected with GEN IIIA strains in their study were all negative for the cable pilin gene (132), and therefore were not RAPD group 2. Infection with genomovar II is postulated to be associated with a more benign clinical course, based mostly on clinical experience, but this association has not been directly studied. Infection with BCC has a number of implications for patient management in CF. Patient to patient transmission of BCC infection, at least of specific genomovar and RAPD groups, has been clearly shown (133-136). With the overall increase in mortality associated with BCC infection, clinics have instituted rigorous cohorting of CF patients infected with BCC from non-BCC infected CF patients. The possibility of variable pathogenicity of different BCC genomovars provides support for an extension of this cohorting by segregating BCC infected CF patients based on genomovar and RAPD grouping; however, this has yet to be investigated. The heterogeneity in clinical course in CF patients infected with BCC may be related to variability in BCC species. Application of molecular techniques has established the complexity of BCC, with a number of distinct genomovars described. Currently, BCC has been ordered into ten genomovar groups identified and named as distinct species (137-146). Limited information is available to characterize any epidemiological BCC genomovar differences. The possibility of increased virulence of specific genomovars has been extrapolated from the observations of spread of common strains amongst CF patients with environmental contacts. BCC has been shown to adhere to respiratory epithelial cells (147, 148), and BCC secretory products induce the release of proinflammatory cytokines (interleukin 6 and 8) and prostaglandin E2 from these cells which is proposed to contribute to the excessive inflammation that characterizes this infection (149). Sajjan and coworkers(147) have shown that while CF airway mucus still functions in trapping BCC, it has impaired ability to kill the pathogen or prevent BCC related epithelial damage. It is also suggested that CF airway mucus may lack bactericidal agents (147). The GEN IIIA RAPDtype 2 (also known as ET12 lineage) has been found by Sajjan and colleagues (150) in the epithelial cells in the terminal and respiratory bronchioles, in hyperplastic epithelia, within inflamed alveolar septae and in luminal and parenchymal macrophages. BCC has been shown to form biofilms and co-colonize biofilms with P. aeruginosa (151). Others have also 130  shown that BCC persists in macrophages under in vitro conditions and is able to replicate and resist killing by oxidative burst (152, 153). The pattern of distribution of BCC in the CF airways appears different from the common pathogens that colonize CF patients; aeruginosa  P.  is confined to the airway lumen. BCC, particularly in more severe cases, has been  shown to cross the epithelial barrier into the parenchyma and into capillaries (154). With these advances, we can now examine whether variability in clinical course can be explained by genomovar differences. MBL2 binding to BCC should result in complement activation but also in the activation of alveolar macrophages and therefore should modulate the lung defense system through the regulation of inflammatory cytokines. Garred has implied that susceptibility to BCC infection is associated with an MBL2 deficiency status (43). In their experience, they reported that seven of the 10 patients who became chronically infected with BCC during an 8-year data collection interval had an MBL2 deficient genotype. In order to address the effect of modifier genes on BCC susceptibility and infection we first have to address a number of questions, which have not been previously investigated. At this time, there is only one study, a case control study, which addressed pulmonary disease progression pre-and post-colonization with BCC (155). Based on the BCC taxonomy information that is available it is important to elucidate whether the differences observed in pulmonary disease progression and survival are related to genomovar group and/or RAPD type group. At this time we cannot rule out that co-infection with P. aeruginosa may explain most of the heterogeneity observed in pulmonary disease progression and survival. Therefore, before addressing the specific questions related to MBL2 as a potential modifier gene in CF, these questions need to be addressed and will be in this study.  131  4.2 RESULTS 4.2.1 Hardy Weinberg equilibrium and linkage disequilibrium Table 4.4 shows Hardy Weinberg equilibrium (HWE) results and the frequencies of the MBL2 and pulmonary surfactant SPA-1, SPA-2 and SPD gene polymorphisms in the study cohort. We showed deviations from HWE for two of the polymorphisms investigated in the MBL2 gene: MBL2-B and D where we showed a deficiency of heterozygotes (18% versus 27% for B-allele, p = 0.00001 and 12% versus 14% for D-allele, p=0.02; Table 4.4). The prevalence of the MBL2 promoter and structural polymorphisms were within the ranges reported in the literature for CF (43) and non-CF cohorts (29, 86, 87). Turner (81) summarized the literature concerning MBL2 allele frequencies and found the B allele varied between 0-16%, the C allele between 0-29% and the D allele between 0-6%. The frequency for the SPD (Thr11Met) allele was within the reported range for Caucasian populations (52.8-70.8%) (53, 57) with similar distribution of genotypes (TT-35.4%, CT-47.1% and CC17.5%(57)). The SPA-1 polymorphism was also not in HWE and we showed a deficiency in the T-allele, specifically our cohort did not have any CF patients with a TT genotype (Table 4.4). We also showed a deficiency in heterozygotes for the SPA-2 (Thr9Asn) polymorphism (Table 4.4).  Reported frequencies for the SPA genes SNPs are based on commonly  occurring haplotypes based on known polymorphisms in the two genes as described in DiAngelo and associates (52), but we were unable to extrapolate the frequency of the alleles for individual SNPs as the percentages for the haplotypes are given without reference to counts. In order to address the possibility that there may be preferential death in a genotype, we also examined HWE after stratifying the subjects by age (<25 and > 25 years of age). Our results are shown in Tables 4.4 and 4.5. The deviation from HWE for the MBL2-B and MBL2-D alleles occurred, respectively, in the adult and children’s group (Table 4.5 and 4.6). However, the values (frequency of alleles) were within reported ranges in the literature (81). We did not show deviations from HWE for the pulmonary surfactant polymorphisms when stratified by age (Table 4.6). The results for LD are presented in Table 4.7. The MBL2 gene polymorphisms were not all in LD. The MBL2 gene polymorphisms (except for the B-allele) were not in LD with the 132  pulmonary surfactant polymorphisms investigated, and this observation concurs with literature findings (21). In our study, we showed weak LD between the MBL2 B-allele and the SPA-2 (A26C) polymorphism. We showed strong LD between the SPA-1 (C655T) and SPA-2 gene polymorphisms investigated in this study (A26C, A667C). Inferred haplotype probabilities using the software package PHASE (156, 157) are presented in Tables 4.7 and 4.8. The inferred haplotypes for the four MBL2 SNPs inferred using PHASE (156, 157) are shown in Table 4.8. The low expression promoter for MBL2 was shown to always be found with the wild-type allele for the three exon 1 coding polymorphisms (i.e., for haplotype XAAA N=215, 22%) and six inferred haplotypes were shown with the high expression promoter allele and the coding polymorphisms, as previously documented in the literature (158). Haplotypes probabilities were inferred for the four pulmonary surfactant gene polymorphisms investigated (Table 4.9), for SPA-1 with the two SPA-2 polymorphisms (Table 4.10), and for the two SPA-2 polymorphisms (Table 4.11). The CA haplotype was the most common inferred haplotype for the SPA-2 SNPs (frequency=61%). The CAC haplotype was the most frequent inferred haplotype across the three SPA SNPs investigated (frequency=44%). The inferred haplotype results were used for grouping our cohort for the SPA-2 SNPs into three groups based on having zero; one or two copies of the CA haplotype. Because of the number of inferred haplotypes for the three pulmonary surfactant genes and the low LD across SPA-1 and SPA-2 with SPD, we did not use the inferred haplotypes information to group the study cohort, instead we individually investigated the SPD polymorphism. 4.2.2 Descriptive data results and study cohort grouping Table 4.12 presents the data for the cohort used for investigating MBL2 deficiency. In addition, the characteristics of the BCC (transiently and chronically infected) and control groups are presented in Section 4.2.7 and in Table 4.20. Susceptibility to BCC infection in MBL2 deficient versus wild-type CF patients was investigated using logistic regression. We did not show differences in BCC colonization in CF patients who were MBL2 deficient (5.7%) compared to MBL2 wild-type (5.3%, p=0.50). 133  The T allele for the SPA-1 gene polymorphism was rare and only occurred in heterozygous form in the study cohort. The descriptive clinical characteristics of the study cohort presented in Table 4.13 for the SPA-1 polymorphism were based on whether the T allele was present (i.e., CT genotype) or not (CC genotype). Inferred haplotypes generated by the PHASE program for the two polymorphisms (A26C and A667C) studied in the SPA-2 gene were used to group the study cohort based on whether there were zero, one, or two copies of the most common inferred haplotype CA (Table 4.14). SPD genotype was used to group the study cohort into three groups (Table 4.15). 4.2.3 Pulmonary disease progression: mixed effects regression on %predFEV1 MBL2 gene:The subjects were grouped for MBL2 polymorphisms based on the functional effect of the polymorphisms as described in Garred and associates (43). We first investigated whether there was a difference in the rate of decline in %predFEV1 with P. aeruginosa infection. We showed no association with rate of decline in %predFEV1 and MBL2 deficiency status when controlling for CFTR genotype, P. aeruginosa infection status and gender. Model 4.2.3-A  %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + PA infection status (0/1=not infected/chronically infected) + CFTR class mild+ CFTR class heterozygous severe + CFTR unknown/unclassified + MBL2 deficiency * Time Base group was MBL deficient In the reduced model (removing non-significant terms), we showed a similar rate of decline in %predFEV1 irrespective of MBL2 gene grouping. The p-values for the interaction terms were not significant: MBL2 deficiency * Time (p=0.77), MBL2 deficiency * P. aeruginosa infection status (p=0.74), MBL2 deficiency * P. aeruginosa infection status * Time (p=0.66). The only significant interaction term in the model was P. aeruginosa infection status * Time (p=0.001), indicating a steeper decline in %predFEV1 when chronically infected with P. aeruginosa.  134  Model 4.2.3-A (reduced model)  %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + PA infection status (0/1=not infected/chronically infected) + MBL2 deficiency * Time Base group was MBL deficient We next investigated whether there was a difference in the rate of decline in %predFEV1 with BCC infection. Since there is the belief that different genomovars of BCC have a variable effect on pulmonary disease progression and survival, we first investigated whether this was the case and examined if there were differences in pulmonary disease progression post-acquisition of BCC, and whether there were differences based on the BCC genomovar. These analyses are presented at the end of the results section entitled ‘BCC infection and pulmonary disease progression and survival’ (in Section 4.2.7). Based on these results, which showed no differences in pulmonary disease progression in CF patients chronically infected with BCC GENII versus GEN IIIA, we did not stratify by BCC genomovar for the MBL2 analyses. Descriptive data on the study cohort available for statistical analysis are presented in Table 4.22. In our linear mixed effects models we investigated whether MBL2 deficiency and chronic infection with BCC were associated with more rapid pulmonary disease progression in CF. In the first model we used a dummy variable to categorize patients into CF patients not infected with BCC and CF patients chronically infected with BCC (regardless of the genomovar group they were infected with). In this model, we did not distinguish the preand post-acquisition interval with BCC. Our model was similar to the model described above including all two-way, three-way and four-way interactions with the variable Time. Model 4.2.3-B  %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + PA infection status (0/1=not infected/chronically infected) + BCC infection status (0/1=not infected/chronically infected) + MBL2 deficiency * Time + PA infection status * Time + BCC infection status * Time). Base group was MBL deficient 135  The p-values for the interaction terms with MBL deficiency (i.e., MBL2 deficiency * Time, p=0.78; MBL2 deficiency * PA infection status * Time, p=0.46; MBL2 deficiency * BCC infection status * Time, p=0.81 and MBL2 deficiency * BCC infection status * PA infection status * Time, p=0.90) were not significant. The following interaction terms, which did not include MBL deficiency status were significant (PA infection status * Time, p=0.003; BCC infection status * Time, p=0.0006) showing a steeper decline in %predFEV1 over time in CF patients who are chronically infected with either P. aeruginosa or BCC. In the next model, we used the longitudinal data collected during the pre-acquisition with BCC interval for the CF patients chronically infected with BCC and compared to the longitudinal data collected from CF patients not infected with the BCC pathogen (Table 4.16, model 1). Patients chronically infected with BCC regardless of genomovar group were grouped together. We showed no significant differences (p=0.63) in the rate of decline in %predFEV1 over time in MBL2 deficient versus wild-type CF patients. Model 4.2.3-C (pre-acquisition of BCC)  %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + PA infection status (0/1=not infected/chronically infected) + BCC infection status (0/1=not infected/chronically infected) + MBL2 deficiency * Time + PA infection status * Time + BCC infection status * Time). Base group for MBL2 was MBL deficient. Time for the BCC infected patients was BCC pre-acquisition interval In the third model, the pre and post-acquisition intervals for BCC infection were included in the models as fixed and random effects. Patients chronically infected with BCC regardless of genomovar group were grouped together. In this model, we examined if the rates of decline in %predFEV1 were different for MBL2 deficient and wild-type groups following infection with BCC (Table 4.16, model 2). Model 4.2.3-D 136  %predFEV1 = Time + Sex (0/1=male/female) + MBL2 deficiency (0/1=deficient/wild-type) + BCC infection status (0/1=no BCC infection/ BCC colonization) + AgePP (0/(1+∞)=pre-colonization/days post-colonization) + MBL2 deficiency * AgePP). Base group for MBL2 is MBL deficient Our results showed that significant differences existed between MBL deficient and wildtype groups and rate of decline in %predFEV1 in CF patients chronically infected with BCC (p=0.001). Additional comparisons are shown in Table 4.16 model 2 (labeled comparisons). In model 2 comparison 1, we have not distinguished BCC infection status (i.e., BCC infected or not infected) and simply examined pre and post-acquisition. Our results show a significantly different rate of decline in %predFEV1 in CF patients who are MBL deficient versus wildtype (p=0.02). In comparison 2, we examined only controls (i.e., not infected with BCC) and showed no significant difference in the rate of decline in %predFEV1 in CF patients who are MBL deficient versus wild-type (p=0.45). In comparison 3, we examined only CF patients chronically infected with BCC (not distinguishing genomovar group) and showed a significantly steeper rate of decline in %predFEV1 in CF patients who were MBL2 wild-type (p=0.02). In comparison 4, we examined only CF patients who were MBL2 deficient and looked at the rate of decline in %predFEV1 in those CF patients who were chronically infected with BCC versus controls (not infected with BCC) and showed a significantly steeper rate of decline in %predFEV1 in CF patients who were MBL2 deficient and chronically infected with BCC (p=0.01). In our last comparison 5, we examined only CF patients who were MBL2 wild-type and looked at the rate of decline in %predFEV1 in those CF patients who were chronically infected with BCC versus controls (not infected with BCC) and showed a significantly steeper rate of decline in %predFEV1 in CF patients who were MBL2 wild-type and chronically infected with BCC (p<0.0001). SPA-1 gene:In our linear mixed effects models we investigated whether having the less common CT versus CC genotype for the SPA-1 (C655T) polymorphism was associated with a different rate of decline in %predFEV1. We showed a similar rate of decline in %predFEV1 irrespective of SPA-1 genotype 137  Model 4.2.3-E  %predFEV1= Time + Sex (0/1=male/female) + SPA-1 genotype (0/1=CC/CT) + SPA-1 genotype * Time). Base group for SPA-1 was CC The p-value for the interaction term SPA-1 genotype * Time was 0.57. In a second model we included P. aeruginosa infection status (0/1=not infected/chronically infected) including interaction terms (i.e., SPA-1 genotype * P. aeruginosa infection status (p=0.82), P. aeruginosa infection status * Time (p=0.13), SPA-1 genotype * P. aeruginosa infection status * Time (p=0.73)), but did not show significance. Model 4.2.3-F  %predFEV1= Time + Sex (0/1=male/female) + SPA-1 genotype (0/1=CC/CT) + PA infection status (0/1=not infected/chronically infected) + SPA-1 genotype * PA infection status + SPA-1 genotype * Time + PA infection status * Time + SPA-1 genotype * PA infection status * Time Base group for SPA-1 was CC In our final model (i.e., model 4.2.3-G) we included CFTR genotype with P. aeruginosa infection status (0/1=not infected/chronically infected) including the following interaction terms in our final reduced model: SPA-1 genotype * Time (p=0.005), SPA-1 genotype * P. aeruginosa infection status * Time (p<0.0001), and SPA-1 genotype * P. aeruginosa infection status * CF class (base group is homozygous for 2 severe mutations, p=0.02). In this model we showed that ƒ  The slope is steeper if a patient has a CT genotype compared to a CC genotype.  ƒ  The mean %predFEV1 is lower if they have at least one mild CFTR mutation, are chronically infected with P. aeruginosa and have a SPA-1 genotype of CC compared to any other 3-way combination (i.e., for the interaction term: SPA-1 genotype * P. aeruginosa infection status * CF class). 138  Model 4.2.3-G  %predFEV1= Time + Sex (0/1=male/female) + SPA-1 genotype (0/1=CC/CT) + PA infection status (0/1=not infected/chronically infected) + CFTR genotype + SPA-1 genotype * PA infection status + CFTR genotype * SPA-1 genotype + SPA-1 genotype * Time + PA infection status * Time + CFTR genotype * Time + SPA-1 genotype * PA infection status * Time + CFTR genotype * SPA-1 genotype * Time + CFTR genotype * SPA-1 genotype * PA infection status * Time Base group SPA-1 was CC. CFTR genotype is an abbreviation for the CFTR class groupings. SPA-2 gene:Inferred haplotypes generated by the PHASE program for the two polymorphisms (A26C and A667C) studied in the SPA-2 gene were used to group the study cohort based on whether there were zero (SPA-2CA0), one (SPA-2CA1), or two (SPA2CA2) copies of the most common inferred haplotype CA. The clinical characteristics of the cohort based on this grouping are presented in Table 4.14. We showed a similar rate of decline in %predFEV1 based on our SPA-2 grouping, that is between SPA-2CA0 and SPA2CA2 (p=0.37), between SPA-2CA1 and SPA-2CA2 (p=0.21) and between SPA-2CA0 versus SPA-2CA1 (p=1.00). Model 4.2.3-H  %predFEV1 = Time + Sex (0/1=male/female) + SPA-2CA0 (0/1=1 or 2 copies of CA/0 copies of CA) + SPA-2CA1 (0/1=0 or 2 copies of CA/1 copy of CA) + SPA-2CA0 * Time + SPA-2CA1 * Time. Base group for SPA-2 was 2 copies of CA (i.e., SPA-2CA2). In a second model we included P. aeruginosa infection status (0/1=not infected/chronically infected) including interaction terms (i.e., SPA-2CA0 and SPA-2CA1) * P. aeruginosa infection status (p=0.40 and 0.25), P. aeruginosa infection status * Time (p<0.0001), (SPA2CA0 and SPA-2CA1) * P. aeruginosa infection status * Time (p=0.17 and 0.15)), but did not show significance with our gene polymorphism of interest. 139  Model 4.2.3-I  %predFEV1 = Time + Sex (0/1=male/female) + SPA-2CA0 (0/1=1 or 2 copies of CA/0 copies of CA) + SPA-2CA1 (0/1=0 or 2 copies of CA/1 copy of CA) + PA infection status + SPA-2CA0 * PA infection status + SPA-2CA1 * PA infection status + SPA-2CA0 * Time + SPA-2CA1 * Time + PA infection status * Time + SPA-2CA0 * PA infection status * Time+ SPA-2CA1 * PA infection status * Time. Base group for SPA-2 was 2 copies of CA (i.e., SPA-2CA2). The model was reduced but did not show any significant results for the gene polymorphism (p-values for main effects SPA-2CA0 and SPA-2CA1 were 0.34 and 0.32, respectively), except for a significant P. aeruginosa infection status * Time interaction (p<0.0001) meaning a steeper decline in %predFEV1 in those CF patients who were also chronically infected with P. aeruginosa versus not infected with the pathogen. SPD gene:The grouping for the SPD gene polymorphism (C32T) was based on genotype into three groups: CC, TT and CT. We showed a similar rate of decline in %predFEV1 based on our SPD genotype grouping. Model 4.2.3-J  %predFEV1 = Time + Sex (0/1=male/female) + SPD-CC (0/1=CT or TT/CC) + SPD-CT (0/1=CC or TT/CT) + SPD-CC * Time + SPD-CT * Time. Base group was SPD-TT. The p-value for the interaction terms SPD-CC * Time + SPD-CT * Time were 0.93 and 0.80, respectively, showing no difference in the rate of change in %predFEV1 between SPDCC and SPD-TT and between SPD-CT and SPD-TT. There was a similar rate of decline in %predFEV1 for SPD-CC versus SPD-CT (p=0.86).  In a second model we included P.  aeruginosa infection status (0/1=not infected/chronically infected) including interaction terms (i.e., (SPD-CC and SPD-CT) * P. aeruginosa infection status (p=0.95 and 0.20), P. aeruginosa infection status * Time (p<0.0001), (SPD-CC and SPD-CT) * P. aeruginosa 140  infection status * Time (p=0.49 and 0.17)), but did not show significance with our gene polymorphism of interest. Model 4.2.3-K  %predFEV1 = Time + Sex (0/1=male/female) + SPD-CC (0/1=CT or TT/CC) + SPD-CT (0/1=CC or TT/CT) + PA infection status + SPD-CC * PA infection status + SPD-CT * PA infection status + SPD-CC * Time + SPD-CT * Time + PA infection status * Time + SPDCC * PA infection status * Time+ SPD-CT * PA infection status * Time. Base group was SPD-TT. The model was reduced, but we did not show any significant results for the gene polymorphism (p-values for main effects SPD-CC and SPD-CT were 0.59 and 0.49, respectively), except for a significant P. aeruginosa infection status * Time interaction (p<0.0001) meaning a steeper decline in %predFEV1 over time in those CF patients who were chronically infected with P. aeruginosa versus not infected with the pathogen. Lastly, in a reduced model we combined the CT and CC group that is having one or two copies of the polymorphism versus being homozygous for the common allele (i.e., TT). Model 4.2.3-L  %predFEV1 = Time + Sex (0/1=male/female) + SPD-CC/CT (0=TT and 1=CC/CT) + SPDCC/CT * Time. Base group was SPD-TT. In this model we showed similar rate of decline in %predFEV1 for SPD-TT versus SPDCC/CT (p=0.86). Then we included CFTR genotype and P. aeruginosa infection status (0/1=not infected/chronically infected) in the model including interaction terms (i.e., SPDCC/CT * P. aeruginosa infection status (p=0.25), P. aeruginosa infection status * Time (p=0.0001), SPD-CC/CT * P. aeruginosa infection status * Time (p=0.14)). CFTR genotype was not a significant predictor and interaction terms were dropped from the final reduced 141  model above, leaving the variable in the model only as a main effect. We did not show significance with our gene polymorphism of interest. 4.2.4 Survival analysis Our survival models were run using CFTR class grouping and alternatively using PSS. The reason for this was that the number of CF patients with a mild CFTR genotype (i.e., homozygosity or heterozygosity for class 4 or 5 mutation) was small and in some cases mild CFTR genotype was not represented in our gene polymorphism groupings and this caused estimation problems with the statistical models. MBL2 deficiency status was not a significant predictor of survival. The first question we addressed was: Is there a difference in survival (i.e., time to death or lung transplantation) between MBL2 deficient versus MBL2 wild type CF patients? MBL2 deficiency was not a significant predictor of survival (Table 4.17). The second question we addressed was: Is there a difference in survival based on MBL2 genotype and whether the CF patient is chronically infected with BCC? MBL2 deficiency was not significantly associated with age of chronic BCC infection (RR=0.93, CI0.95: (0.5-1.75), p=0.83) or age of transient or chronic infection with BCC (RR=0.98, CI0.95: (0.60-1.62), p=0.95). CF patients used in the investigation of the effect of BCC infection and genomovar species on pulmonary disease progression and survival were not included in entirety in the MBL2 analyses as some of the patients died prior to collection of blood for MBL2 genotyping. To explore the possibility of a potential bias in our cohort who consented to the study and blood was obtained for genotyping (versus BCC infected CF patients for whom we were unable to obtain blood for MBL2 genotyping before their death), we looked at the number of CF patients whom we had genotyped for MBL2 and had died or had received a lung transplant for any major differences in patient numbers based on MBL2 genotype. Of the CF patients chronically infected with BCC who were genotyped for MBL2 there were six deaths 142  in the MBL2 wild-type and six deaths in the MBL2 deficient groups. In the group transiently infected with BCC there were four and three deaths in the MBL2 wild-type and deficient groups, respectively. There were no BCC infected patients who had undergone lung transplantation and were MBL2 deficient. There were three and two patients, respectively, who were lung transplantation recipients and infected either chronically or transiently with BCC, and had a MBL2 wild-type genotype. Of these transplanted patients noted above only one of them chronically infected with BCC was still alive. It would appear that a descriptive examination of the data does not reveal differences in the number of deaths in BCC infected CF patients who were MBL2 wild-type or deficient in our study cohort. However, of the lung-transplanted patients an MBL2 wild-type genotype was found in all cases in our study cohort. It appears that our study cohort was not biased for MBL2 genotype. MBL2 deficiency was not predictive of survival (Table 4.17). SPA-1 genotype, SPA-2 haplotype, and SPD genotype were not predictive of survival. CFTR class, P. aeruginosa infection status, sex and modifier gene * P. aeruginosa infection status were included in the models. The results are presented in Table 4.17. 4.2.5. Effect of modifier genes on P. aeruginosa infection status We investigated whether polymorphisms in the pulmonary surfactant genes and MBL2 deficiency contributed to infection with P. aeruginosa. We separately investigated age of first infection and age of chronic infection with the pathogen and results are presented in Tables 4.17 and 4.18, respectively. CFTR class was included in the models. We did not show a significant effect of MBL2 deficiency and genotype for SPA-1, SPA-2 and SPD on age of first and chronic infection with P. aeruginosa. 4.2.6 Frequency of pulmonary infections requiring intravenous antibiotic therapy We used Poisson regression to investigate the effect of the modifier gene polymorphisms on the mean frequencies of pulmonary infections requiring IV therapy. We investigated two time intervals: ƒ  Two years retrospectively; the interval was January 1, 1998 to January 1, 2000 (N=357), 143  ƒ  Up to 5 years retrospectively; the interval was January 1, 1995 (is the earliest possible start interval) to January 1, 2000 (N=357).  We only analyzed patients who had data within the 2 year interval labeled A. The inclusion criteria were: ƒ  Patients in the data set had to have a clinic visit after January 1, 2000.  ƒ  Patients had to have data for at least the 24-month interval prior to January 2000.  In the case of the 5-year interval, the patients who had more than 2 years but less than 5 years of data were used in this analysis and the analysis was adjusted for the time interval available for the patients. All patients included in the 2-year interval had data to be included as well in the 5-year interval. The response variable was number of pulmonary infections requiring IV therapy with separate models used for intervals A and B. We accounted for gender, P. aeruginosa infection (BCC infection status for MBL2 deficiency only) status and CFTR class. In the case where not all four CFTR classes were represented in the study cohort, we used the variable PSS. We showed a significant difference in the mean frequency of IV therapy for pulmonary infections over 5 years (32% higher) in CF patients who were CT versus CC for the SPA-1 polymorphism (p=0.005, N=357). For the SPA-2 polymorphisms investigated, we used the number of copies of the CA haplotypes as the grouping variable. We showed that CF patients who had zero copies of the CA haplotype had significantly lower mean frequency of pulmonary infections over 5 years by 26.5% than CF patients who had 2 copies of the CA haplotype (p=0.01, N=357). We did not show significant differences in the mean frequency of pulmonary infections in CF patients who had one versus two copies of the CA haplotype (p=0.69). The 2-year analyses were not significant for either the SPA-1 or SPA-2 polymorphisms investigated. For the SPD polymorphism we showed that having a CT genotype was associated with a lower mean frequency of pulmonary infections over the 2-year (21%, p=0.04) and 5-year (21%, p=0.002) intervals compared to CF patients with a TT genotype. There were no significant differences in the mean frequency of pulmonary infections over the 2-year and 5year interval investigated for CF patients who were CC versus TT. Our analyses also showed that CF patients who were chronically infected with P. aeruginosa had significantly 144  higher mean number of pulmonary infections compared to those CF patients not infected with the pathogen for both the 2 (28.0%, p<0.0001) and 5 (33.1%, p<0.0001) year time intervals. MBL2 deficiency was not associated with significant differences in the mean frequency of pulmonary infections over the two (p=0.32, N=356) and 5-year (p=0.89, N=356) interval. To examine MBL2 deficiency including the variable BCC infection status we used the variable PSS as a measure of disease severity instead of CF class since all CF patients who were chronically infected with BCC in our cohort were homozygous for severe CFTR mutations. In our first model we used as covariates: gender, BCC infection status and PSS. We did not show significant differences in the mean frequency of pulmonary infections over two (p=0.14, N=356) and five years (p=0.59, N=356) for MBL2 deficient CF patients, however chronic infection with BCC was associated with significant differences in mean frequency of pulmonary infections over two (p<0.0001, N=356) and five years (p<0.0001, N=356). In the second model, we also included as a covariate P. aeruginosa infection status. We did not show significant differences in the mean frequency of pulmonary infections over two (p=0.43, N=356) and five years (p=0.72, N=356) for MBL2 deficient CF patients. 4.2.7 BCC infection and pulmonary disease progression and survival 4.2.7.1 BCC infection and pulmonary disease progression The characteristics of the BCC (transiently and chronically) infected and control groups are presented in Table 4.20. We identified 93 patients who were transiently or chronically infected with BCC and longitudinal data were available on 90 of these individuals. The majority of CF patients infected with BCC were infected with genomovar II and IIIA. There were 16 (10) and 48 (2) CF patients chronically (transiently) infected with GEN II and GEN IIIA, respectively. There were 1(0), 0(4), 1(1) and 5(3) CF patients chronically (transiently) infected with GEN I, IIIB, IV and V, respectively. The small numbers prevented us from including these latter groups in our models. There were no CF patients who carried known mild CFTR genotypes (homozygous or heterozygous for class 4 or 5 mutations) and were either transiently or chronically infected with BCC (p=0.0001). The remainder of the cohort consisted of: ƒ  Fifty five patients who were homozygous for CFTR class 1, 2, or 3 mutations, 145  ƒ  Eighteen patients who were heterozygous for class 1, 2 or 3 mutation and one mutation classified as unknown or unclassified and  ƒ  Twenty one CF patients who were homozygous for two unknown or unclassified CFTR mutations.  The mixed effects regression models were initially run including CFTR genotype as a fixed effect. Given that BCC infected patients were characterized only with severe CFTR genotypes, we alternatively used pancreatic sufficiency status (PSS) as a measure of CF disease severity. PSS proved not to be a significant predictor in our models and thus both variables (CFTR genotype and PSS) were excluded from the final models (Table 4.21). The effect of the center variable was included as a random effect; however, center-to-center variability was small (less than 1% of the total variability). We have previously shown that CF patients who become chronically infected with BCC do not differ in clinical course from CF controls (without BCC infection) prior to infection with BCC (155). In the present study, we investigated the estimated rate of decline in over time  %predFEV1  for up to 5 years pre-acquisition of BCC using linear mixed effects models to  validate our previous results. Our analysis showed similar rate of decline in %predFEV1 in the pre-acquisition interval for transiently (-2.09 %/year) and chronically (-1.51 %/year) BCC infected groups and for the controls (-1.38 %/year). Table 4.21 summarizes our linear mixed models results for the post-BCC acquisition interval. In model 1, we showed a more rapid decline in %predFEV1 in CF patients chronically infected with BCC compared with CF controls not infected with BCC (p=0.04). Comparisons were also made to the noninfected control group (Table 4.21 model 1). In this model we treated the transiently BCC infected group as a separate group (not shown in Table 4.21). Our results show the transiently BCC infected group has an estimated rate of decline in %predFEV1 over the follow up time which is intermediate between the non-infected controls (p=0.13) and the chronic BCC infected patients (p=0.43). The rate of decline in  %predFEV1  in the transiently BCC  infected group is not statistically significantly different from either the control or the chronically infected with BCC groups. The rate of decline in %predFEV1 was similar when comparing different BCC genomovars and RAPD types in chronically infected CF patients (Table 4.21, model 2). In this model, we further investigated the rate of change in %predFEV1 between CF patients infected with GEN II and GEN IIIA. The rate of decline in %predFEV1 in the  post-acquisition interval with BCC was not statistically different for GEN 146  IIIA RAPD-type groups, these groups were collapsed and the analysis rerun for the truncated GEN IIIA group and the results are shown in Table 4.21 (model 2- labeled collapsed GEN IIIA). 4.2.7.2 BCC and P. aeruginosa infection and pulmonary disease progression We next investigated whether concurrent chronic infection with P. aeruginosa would further contribute to decline in %predFEV1. These results are shown in Table 4.21 model 3. The P. aeruginosa infected CF patients were treated to have a single rate of decline in %predFEV1 prior to their infection with BCC and we are assuming that this single line is a good representation of the patients’ rate of decline in %predFEV1 for the BCC post-acquisition interval. While the majority of co-infected patients first harbored P. aeruginosa, seven acquired the organism after BCC infection. Because of the small numbers, however, we were unable to investigate the order of co-infection in our regression models, although our main interest was whether there was a cumulative effect of infection with both pathogens. Of the CF patients colonized with P. aeruginosa, 12 were transiently infected and 36 were chronically infected with BCC. The mean time from age of first infection with P. aeruginosa to co-infection with BCC was 3.7(1.8) years for transiently infected and 5.2(1.2) years for patients chronically infected with BCC. There were seven patients who were chronically infected with BCC and subsequently became infected with P. aeruginosa (time to co-infection with P. aeruginosa 0.6-13.5 years post-acquisition of BCC). CF patients who were already infected (or were identified within 8 months post-BCC infection to be infected with P. aeruginosa) were included for analyses in model 3 in the co-infection group. In model 3, we investigated the rate of decline in %predFEV1 for CF patients infected chronically with P. aeruginosa or BCC, or both pathogens versus those infected with neither pathogen. Our results show that concurrent chronic infection with BCC (regardless of genomovar) and P. aeruginosa contributed to a greater deterioration in %predFEV1 compared to the decline in %predFEV1 exhibited by BCC or P. aeruginosa infection alone or CF controls infected with neither pathogen. 4.2.7.3 BCC infection and survival We addressed three questions in our survival analysis with Cox Proportional Hazards modeling: 147  Question 1: Are there survival differences between transient and chronic BCC infected patients and patients not infected with BCC? Gender and PSS (but not CFTR class groups) were significant effects and were therefore included in the model. We showed that BCC infection, whether chronic (RR =3.93, CI0.95(2.4, 6.4), p<0.0001) or transient (RR=6.46, CI0.95(2.9, 14.4), p<0.0001) (relative to controls), adversely affected survival. Question 2: Are there survival differences among CF patients chronically or transiently infected with BCC and controls with and without co-infection with P. aeruginosa? In this model, we compared transiently and chronically BCC infected groups against our BCC uninfected cohort but in this case divided the uninfected with BCC group into 2 groups; those who were chronically infected and those who had never been infected with P. aeruginosa. The BCC infected group was also divided for a further analysis in BCC infected CF patients with and without co-infection with P. aeruginosa. We were unable to explore the question of whether there are survival differences with co-infection versus infection with P. aeruginosa or no pathogen including the BCC transiently infected group due to estimation problems fitting the model (small N for BCC transiently infected patients with none of the patients chronically infected with P. aeruginosa). We did not adjust for the order of infection with P. aeruginosa, i.e., whether P. aeruginosa preceded or followed BCC infection, as we were interested in whether there was an additive effect of infection with both pathogens. We showed: ƒ  Chronic infection with P. aeruginosa in the control group was associated with a significantly higher likelihood of death or lung transplant compared with noninfection with both pathogens (RR=4.1, CI0.95(1.4, 12.0), p=0.01).  ƒ  CF patients who were chronically (RR=13.6, CI0.95(4.4, 42.0),  p=0.0001) or  transiently (RR=22.7, CI0.95(6.2, 83.0), p=0.0001) infected with BCC showed a significantly higher likelihood of death or lung transplant relative to non-infected controls. ƒ  CF patients chronically infected with BCC and co-infected with P. aeruginosa showed a significantly higher likelihood of death or lung transplant (RR=6.7, CI0.95(1.7, 26.8), p=0.007) than non-infected controls.  ƒ  There were no significant differences in death or lung transplant between CF patients chronically infected with BCC only and controls infected with (RR=2.0, CI0.95(0.8, 4.7), p=0.12). 148  P. aeruginosa  ƒ  There were no significant differences in death or lung transplant between CF patients chronically infected with BCC only and chronically infected with BCC with coinfection with P. aeruginosa (RR=1.6, CI0.95(0.6, 3.8), p=0.33).  ƒ  CF patients chronically infected with BCC with co-infection with P. aerugin